Anger.

I can get very angry. I don’t want to. I don’t mean to. I have bipolar disorder and I have experienced hypo-mania before. But lately within the last few years, dysphoric mania is making a comeback. I am trying to combat that. I hate this uncontrollable boiling over with anger sensation more than anything. I hate the regret, time wasted, regret, time wasted, loneliness, time wasted, etc.

Bipolar Disorder.

For those that don’t know, bipolar disorder comes with a variety of symptoms. One of those symptoms, you may be aware of, mania. Mania doesn’t always mean fun. In fact, I’d venture to say that all in all, even the fun manias aren’t as fun in hindsight. But my kind of mania lately, is no fun at all. I get angry. What’s happening in my brain is I’m very depressed, but also very hyper at the same time. If I was a chemist, asked to make a formula for uncontrollable anger, I’d look at that. Because that is what it feels like. Blind rage.

I had a nice coffee post written. 907 words on a new blogging site, seemed real easy. It felt cool. Easy on the eyes too. The editor was just what I wanted. I alt-tab and my work is gone. My phone flashes and folks are asking questions. Keri is trying to help me. And I just threw the phone away.

Later.

In hindsight, I think that was the right move. It shouldn’t have even been in that equation. The situation wouldn’t have felt so escalated. So that’s it. No more phone. None. I’m out. I can’t behave like I want to be with it. I am hopelessly addicted to it. I have to remove it from the equation.

Brain Power.

Since I have been off social media I have learned some things. I have more brain power. Let me explain, since I have quit social media, I can read 40 pages in one sitting without issue. It’s been years since I could do that. I have a reading station now.

Stations.

I make stations temporarily around the house depending on projects or what I am into. Right now, I have a demo desk (also easily converted to writing desk), tracking space, lyrics area, guitar jamming space, and a reading space. The reason I have to do all this silly ritualistic stuff is because I have the attention span of a gnat. The reason I have the attention span of a gnat is a ton of unfettered internet time of over 2 decades. Who stands a chance?

No one stands a chance, (really).

No one. Against a fleet of data scientists armed with all our data, no one stands a chance. They’re going to get in the brain. It isn’t the number one industry in the whole world because it vibes. They got the numbers they need, like insurance. They are on the winning team.

Underneath everything is seething anger at the way things are. I try not to be a pain with it. I try to help. I don’t believe in anything radical like acceptance or forgiveness. I don’t do believing much. Never on purpose. I do believe things, but if I find myself believing something, I examine that like hell. I think it is as much compulsion as skepticism. But we’re talking my beliefs, not anyone else’s. So no harm in it. Not trying to crush anyone’s faith. I just don’t really like faith.

Calm.

I feel a lot more calm now. This is a reason I write. I know it will help me feel better. Even if I am only reading or writing my words, I am helping myself. I think that is pretty cool.

How I fight the guilt.

I feel terrible about the dysphoria. The guilt. How awful I must seem to some. I find it hard to move on after an episode. I just want to die. My kids haven’t spoken to me in years. Might be a reason. I get loud, yell, act stupid. Even today. Not often. Not even when anyone can see. But I still do. Shaking with violent anger. I take an Ativan, call it day. I don’t want to be angry. I don’t want to be stupid. I don’t want to be violent. So I got to take myself out of the game for a few sometimes. It’s the only way I have found to combat the guilt. And after all, that’s really what I don’t want.

I evolve, lots of folks do not.

So I am changing, still. If bipolar has done anything “for” me, it’s point me to different ways of processing, being, living, without hurting anyone. That has always been my goal. Still is. I discovered a breakthrough with CBT/DBT group therapy in New York but haven’t been able to find anything like that here in TX. I like CBT/DBT. I think the examining and positive-habits formed helped me as much as anything.

Do something masterful every day.

My counselor there told me “make sure you do something to help you feel masterful everyday, Kelly.” He knew what he was talking about. Years later, after doing masterful things everyday for a while, I’ve learned it’s cultivating good things in me. Much like growing a plant, doing things you like doing grows you. Certain times of the year or after big projects, it dawns on me what I spend all that time being masterful by myself for. It’ s so I can be part of a masterful work. The growth of getting better, slowing the mind a bit, on a regular basis, also combats the guilt, I’ve found. Can’t be all bad if I made this, whatever it is.

No phone but I am fine.

Today no phone. But I have demos. I have a Charles Bukowski book that is a gut-buster it’s so funny. No phone. First time in years. After sitting down to write this, first, I am excited for it.  -Kelly

There is a curious contradiction emerging in Ireland's energy transition.

For generations, energy policy was shaped by a simple assumption: there would never be enough electricity. The challenge was always to generate more, import more, build more capacity and ensure that demand could be met.

Today, a different kind of problem is beginning to emerge. At particular moments, Ireland has more renewable electricity than it can use. This is not a prediction. It is already happening.

According to EnergyCloud's latest assessment, drawing on EirGrid data to 31 July 2026, more than 2,540 GWh of renewable electricity has been dispatched down since the beginning of 2025. In simple terms, electricity that could have been generated could not be accommodated by the system and was therefore unused.

The figures are striking. EnergyCloud estimates the retail value of that energy at approximately €899 million. The numbers are large enough to attract attention, but they are also large enough to become abstract. Most of us struggle to visualise a gigawatt-hour, just as we struggle to picture a million euro beyond seeing a large number accompanied by a currency symbol. To put that figure in perspective, it is roughly comparable to the estimated cost of the proposed Shannon LNG terminal and would represent a substantial contribution towards the cost of a second Turlough Hill-scale storage project.

Yet another set of figures from the same report tells a different story. In the first seven months of 2026, EnergyCloud and its partners used signals from surplus renewable generation to heat more than 40,000 tanks of hot water for households experiencing energy poverty.

Suddenly the scale changes. The discussion is no longer about megawatts, dispatch-down, grid constraints or balancing markets. It is about a family having hot water.

The significance of this may extend beyond the immediate benefits provided by the scheme: for many years, energy policy has largely been framed as a problem of production. How do we generate enough electricity? How do we replace fossil fuels? How do we build renewable capacity quickly enough?

Important questions remain, but the appearance of renewable spillage suggests that another question is emerging. How do we connect available energy with human need? That is a subtly different challenge. The issue is not necessarily that resources are absent. The issue may be that they are disconnected.

A surplus exists in one place within the system. A need exists elsewhere. Between them lie infrastructure, market rules, technologies, institutions and habits that were designed for an earlier era. Seen this way, some forms of waste begin to look less like failures of production and more like failures of connection.

The EnergyCloud project is interesting precisely because it explores one possible connection. It does not solve Ireland's energy challenges. It does not eliminate dispatch-down. It does not remove the need for storage, grid reinforcement or interconnection. What it demonstrates is something simpler. It shows that surplus energy can become a signal. And once that happens, the definition of waste begins to change.

Historically, energy systems flowed in one direction. Electricity was generated at large power stations and delivered to customers. The relationship was straightforward.

Increasingly, however, modern energy systems are becoming networks of relationships. Storage responds to generation. Demand responds to availability. Electric vehicles become temporary energy assets. Hot water cylinders become thermal storage. Digital signals connect places of abundance with places of need. The system begins to behave less like a pipeline and more like an ecosystem.

This may turn out to be one of the most significant but least discussed aspects of the energy transition. The challenge is no longer simply generating renewable electricity. It is learning how to live within a system where energy arrives differently, moves differently and sometimes appears in unexpected abundance.

The EnergyCloud figures therefore point towards a larger question. If Ireland can identify periods of surplus renewable generation and use them to provide benefits for households today, what other connections become possible tomorrow? Community batteries. Flexible demand. Electric vehicle charging. District heating. New forms of social and economic infrastructure that have not yet fully emerged.

The answer may not lie in eliminating every unit of wasted energy. No system will ever be perfectly efficient. The more interesting possibility is that we are beginning to recognise resources that were previously invisible because we did not yet know how to connect them to the people who could use them.

Perhaps the story is not really about wasted electricity at all. Perhaps it is about a society learning to recognise that abundance and scarcity can exist in the same system at the same time, and learning how to connect one to the other. David Marshall Dublin

https://go.marshall.ie/energy-briefing-0-review

Readers interested in the wider context can explore the Energy Briefing Series, including Briefing 0 on Ireland's spilled energy problem. In the coming weeks, BriefMe.ie will bring together layered policy briefings designed to make complex issues accessible at different levels of depth, while NRJ.ie will follow the developments, projects and experiments that are shaping Ireland's energy future in real time.

Just write. That was the decision I made at around 2:35 this morning. Just write. No ambition. No lofty or altruistic goals. Just writing for its own sake. My own selfish sake. My own sane soul.

I took a break from my blogging. Nearly a month, though it felt like three. I stopped because I doubted what I was writing and why. And stopping was the natural response to that doubt. Taking time to reflect and soul-search. To ask myself, “What am I doing here? And where am I going?”

I doubted my reasons. I doubted my fidelity. I doubted my pen.

But what I’ve found is that I’ve been overthinking nearly everything, which I suppose is better than simply not thinking at all. I presume there’s enough empty-headedness devouring the world these days.

Thinking is good, but overthinking is debilitating. Overwhelming in an unhealthy way.

She was right to warn me that my doubts and cynicism could give way to nihilism. And for a split second, they had.

Diana Ross' haunting vocals echoed in my head — Do You Know Where You're Going To. Another musical synchronicity?

I’ve recovered since, however, having come to the conclusion that writing is just my way of understanding myself and making sense of the world, even when the world seems endlessly full of senselessness.

So I’ll stop turning things over and over in my mind, as thoughts only get more and more tangled when the laundry of the mind is set on infinite spin. While the wheels of the world keep churning their machinations of madness.

What will be, will be.

Like Camus’ Sisyphus, I’ll keep pushing the pen across the page, with only that line of black ink as proof any of this was real. And I’ll be as happy as I can be, content in knowing my capabilities and limitations. In knowing her unconditional love despite our current distance.

All else is but the certain mystery of the unknowable.

For the record, the only agents I have representing my work are Booklyn; who focus on placing reasonably-sized art-on-paper at institutional collections, especially within the U.S. They've been instrumental in getting my work into a number of university collections (listed here).

No art-dealer for larger works, no literary agent, and certainly no film/TV agent. Most of my dealings have been conducted as a “free agent” so to speak, which, quite frankly, I would like to change, because I am now spent af.

#journal

There is a curious contradiction emerging in Ireland's energy transition.

For generations, energy policy was shaped by a simple assumption: there would never be enough electricity. The challenge was always to generate more, import more, build more capacity and ensure that demand could be met.

Today, a different kind of problem is beginning to emerge. At particular moments, Ireland has more renewable electricity than it can use. This is not a prediction. It is already happening.

According to EnergyCloud's latest assessment, drawing on EirGrid data to 31 July 2026, more than 2,540 GWh of renewable electricity has been dispatched down since the beginning of 2025. In simple terms, electricity that could have been generated could not be accommodated by the system and was therefore unused.

The figures are striking. EnergyCloud estimates the retail value of that energy at approximately €899 million. The numbers are large enough to attract attention, but they are also large enough to become abstract. Most of us struggle to visualise a gigawatt-hour, just as we struggle to picture a million euro beyond seeing a large number accompanied by a currency symbol. To put that figure in perspective, it is roughly comparable to the estimated cost of the proposed Shannon LNG terminal and would represent a substantial contribution towards the cost of a second Turlough Hill-scale storage project.

Yet another set of figures from the same report tells a different story. In the first seven months of 2026, EnergyCloud and its partners used signals from surplus renewable generation to heat more than 40,000 tanks of hot water for households experiencing energy poverty.

Suddenly the scale changes. The discussion is no longer about megawatts, dispatch-down, grid constraints or balancing markets. It is about a family having hot water.

The significance of this may extend beyond the immediate benefits provided by the scheme: for many years, energy policy has largely been framed as a problem of production. How do we generate enough electricity? How do we replace fossil fuels? How do we build renewable capacity quickly enough?

Important questions remain, but the appearance of renewable spillage suggests that another question is emerging. How do we connect available energy with human need? That is a subtly different challenge. The issue is not necessarily that resources are absent. The issue may be that they are disconnected.

A surplus exists in one place within the system. A need exists elsewhere. Between them lie infrastructure, market rules, technologies, institutions and habits that were designed for an earlier era. Seen this way, some forms of waste begin to look less like failures of production and more like failures of connection.

The EnergyCloud project is interesting precisely because it explores one possible connection. It does not solve Ireland's energy challenges. It does not eliminate dispatch-down. It does not remove the need for storage, grid reinforcement or interconnection. What it demonstrates is something simpler. It shows that surplus energy can become a signal. And once that happens, the definition of waste begins to change.

Historically, energy systems flowed in one direction. Electricity was generated at large power stations and delivered to customers. The relationship was straightforward.

Increasingly, however, modern energy systems are becoming networks of relationships. Storage responds to generation. Demand responds to availability. Electric vehicles become temporary energy assets. Hot water cylinders become thermal storage. Digital signals connect places of abundance with places of need. The system begins to behave less like a pipeline and more like an ecosystem.

This may turn out to be one of the most significant but least discussed aspects of the energy transition. The challenge is no longer simply generating renewable electricity. It is learning how to live within a system where energy arrives differently, moves differently and sometimes appears in unexpected abundance.

The EnergyCloud figures therefore point towards a larger question. If Ireland can identify periods of surplus renewable generation and use them to provide benefits for households today, what other connections become possible tomorrow? Community batteries. Flexible demand. Electric vehicle charging. District heating. New forms of social and economic infrastructure that have not yet fully emerged.

The answer may not lie in eliminating every unit of wasted energy. No system will ever be perfectly efficient. The more interesting possibility is that we are beginning to recognise resources that were previously invisible because we did not yet know how to connect them to the people who could use them.

Perhaps the story is not really about wasted electricity at all. Perhaps it is about a society learning to recognise that abundance and scarcity can exist in the same system at the same time, and learning how to connect one to the other. David Marshall Dublin

https://go.marshall.ie/energy-briefing-0-review

Readers interested in the wider context can explore the Energy Briefing Series, including Briefing 0 on Ireland's spilled energy problem. In the coming weeks, BriefMe.ie will bring together layered policy briefings designed to make complex issues accessible at different levels of depth, while NRJ.ie will follow the developments, projects and experiments that are shaping Ireland's energy future in real time.

Ikke alltid hør på mor og far, sier Petter Stordalen i NRK. Og Petter Stordalen må man høre på. For Petter Stordalen er en vellykket mann.

Følg drømmen, sier Petter Stordalen. Gjør det du vil, som du vil! Og hvis du begynner med noe du ikke liker, er det ikke farlig å begynne på nytt og gjøre noe annet. Det har han gjort. Og han må man høre på. For Petter Stordalen er en vellykket mann.

Når du er konfirmant, har du hele livet foran deg, sier Petter Stordalen. Det samme har alle onkler sagt til alle tider i alle konfirmantselskaper, jada jada, når er det kake? Men Petter Stordalen sier det er viktig. Og Petter Stordalen er en vellykket mann.

Investér i deg selv, sier Petter Stordalen. Ah, det var det jeg ikke gjorde. Jeg kjøpte VHS-spiller til huset for konfirmantpengene, så mamma kunne se på britisk krim og jeg kunne se på NBA-kamper. Og hva slags investering er det? Å kjøpe noe til familien! Jeg mener, NBA-spiller ble jeg jo hvert fall ikke, samme hvor mye jeg drømte om det.

Sett alle konfirmantpengene i fond, sier Christian Ringnes. La dem vokse, urørt, til du er 60 og har opparbeidet deg en pen sum. Og Christian Ringnes må man høre på. For Christian Ringnes er en vellykket mann.

At han er oldebarn av Ellef Ringnes, som startet Ringnes bryggeri i 1877 og finansierte Nansens polekspedisjon og donerte Ibsen og Bjørnson-statuene utenfor Nationaltheatret, har ikke noe å si. At Christian Ringnes vokste opp i et møblert hjem og alltid har hatt økonomisk trygghet nok til å ikke trenge konfirmantpenger til noe annet enn nedsylting til en fremtidig pensjonisttilværelse, er helt uviktig. For Christian Ringnes er en vellykket mann, og vellykkede menn vet best hvordan man skal behandle penger.

Vellykkede menn vokser opp på Oslos vestkant og velger å studere ved handelshøyskolen i Lausanne og ved Harvard mens konfirmantpengene ligger på en konto merket «Hold nedsyltet til du er 60».

Mindre vellykkede menn vokser opp i Halden og velger å studere litteratur på Blindern mens VHS-spilleren støver ned i et kott, det kan jo hende noen klarer å reparere den en dag. Mindre vellykkede menn har en konvolutt et sted merket «MÅ IKKE ÅPNES FØR 2077!!!» — den er kvartfull av mynter fra før 1977, da han ble født, og han tenkte en gang på barneskolen, oj så gamle mynter, i 2077 er den sikkert verdt flere hundre kroner, kanskje oldebarna… For mindre vellykkede menn har ikke vellykkede oldefedre. Og 100 år er uforståelig lenge.

Hadde bare mindre vellykkede menn syltet ned egne konfirmantpenger til seg selv, hadde de hatt nok til pensjonisttilværelsen, de 3000 kronene hadde med renter og rentesrente og rentesrenterente vokst til… 69 370 kroner på 45 år. For 69 370 kroner er en pen sum som sikkert kan gjøre mindre vellykkede menn mer vellykkede.

Men det gjorde de ikke. Istedenfor kjøpte de en VHS-spiller så mamma kunne se på britiske krimserier på kveldene og de selv kunne se på kornete dunker fra forrige ukes NBA Action.

Og nå sitter de der, uten en eneste milliard.

Nå kan de ha det så godt.

(Et svar på denne artikkelen)

The latest newsletter finally went out. This is issue #235: Can't Stop Won't Stop.

As always, sign up lives here.

Recently, I had to turn down working with a novelist I admire. He had a fantastic idea for a graphic novel; an exhilarating historical epic that would need at least 500 pages to unfold and probably another 10 years of my life to make. Mind-blowing and attractive as the concept was, I am sadly way too burnt out after THE SOLAR GRID to plunge into anything that elaborate anytime soon.

#journal

if knowledge is power, then those in the dark are powerless.

four weeks.

iv been grinding for a month. It’s great. Terrific even. But exhausting. I don’t know how, bury client is ecstatic. Thrilled that EIGHTEEN images came out great.

I have my doubts on a few, but they were approved and so shipped.

Maybe it would be good to stay this busy for a while. It seems advertising is picking up. And maybe making better personal choices will result in the kinds of blessing Jah always poured out on us.

I was feeling kind of depressed today, but I ended up going to the gym for three hours and I eventually started to feel a little bit better. I’m glad I went on Hinge and I’m glad that I had some amount of success, because I think to me that is at least showing that effort can pay off in that sense, but I also think that I might be taking a little bit of a step back to reenter myself in just life overall and not worry so much about dating.

I am back in the ferris wheel again

Another more of the same, just like last time,

Except now I am older

So a spiral, then

One of the greatest song texts I ever seen or heard sung I think is ”Oh my Lord” by Nick cave

The ladders of life that we scale merrily Move mysteriously around So that when you think you’re climbing up, man In fact you’re climbing down

And I’m thinking about what a friend said that sometimes it’s very rewarding to do the wrong thing

Yes yes sticking the head in the sand

It’s a very bad thing to do, you know?

Turning a blind eye and then Evil is sneaking round the corner, of course.

I think bravery is the most importance trait, together with kindness of course

And I am thinking (using my own brain) that it’s sometimes hard even knowing what is the right thing to do, I never used to have any doubts about that before.

First knowing what is the right thing to do, and then doing the closest approximation of that in this life

It’s not a ferret wheel, not a ferris wheel either: It has a beginning and an end like a spiral

Or a serpentine

Xx

CHAPTER TWO: BEFORE THE SILENCE

“The most dangerous thing about a routine is how much you trust it to last forever.”

For years, our lives moved in a rhythm that felt permanent.

Morning texts. Random messages in the middle of the day. Phone calls that didn’t always say much, but reminded me he was there.

Sometimes we said, I love you. Sometimes we said, I hate you.

But we said something. Every single day.

It didn’t matter if we were arguing or laughing, exhausted or hopeful, close or frustrated.

There was always a message. Always a call. Always some small reminder that we were still part of each other’s world.

After a while, it stopped feeling like communication.

It became routine, the kind of routine you don’t question because it feels as natural as breathing.

I didn’t realize then how much of my life had quietly grown around that rhythm.

Twenty years is a long time to share a life with someone. It’s long enough for habits to turn into foundations. Long enough for the idea of us to become something you assume will always be there.

We had history. Inside jokes that didn’t make sense to anyone else.

Memories stacked on top of memories until they blurred together into something that felt permanent.

There were good days, days where life felt simple and familiar. And there were hard days, too.

Days when we fought. Days when words were said that couldn’t be taken back.

But even then, the rhythm continued.

Even when we were angry. Even when we were distant. Even when things between us felt complicated.

We still showed up in each other’s lives. That was the part I trusted.

Not perfection. Not some storybook version of love.

Just the quiet understanding that we were still there. Still connected. Still part of each other’s world.

For most of my adult life, that connection was something I never questioned. I didn’t spend time imagining a life without it.

I didn’t prepare myself for this. Because when something has been part of your life for that long, you stop thinking about what it would feel like if it suddenly disappeared.

You assume it will always be there.

That the messages will keep coming. That the conversations will keep happening. That the person will always be somewhere within reach.

I believed that. Completely.

Which is why this didn’t feel like distance. It didn’t feel like the normal ending of a relationship.

It felt like something had been ripped out of the structure of my life without warning. And I was left standing in the quiet that followed.

45th Post

It doesn't seem like this platform has a native way of tracking the number of posts... though I suppose you could look at the number of pages because it does appear to limit to ten posts per page. So that alone can get you within 10 posts of the total by looking at total page count.

It's funny, I had a blog many years ago where much of it was spent counting how much had been posted on the blog. A sort of self-referential numerical/statistical mental masturbation exercise. Anyway, decades later, after traveling the world and finding my way back home, I'm doing it again. This time I'm hoping the results are more content heavy and edifying.

So far, most of the posts have been centered around a few major themes and projects. Part of that was pent up ideas I had that I wanted to get to paper. Over time I think there will be more musing posts and random thoughts, but we shall see. For now, some arbitrary stats:

Longest Post: The Blue Monkey Story (86,670 characters). I wrote this with a friend many years ago, so it is kind of a cheat but it is not available anywhere else online that I know of so I think its fair game.

Series with Most Posts: High Fantasy Epic Sword Fighting Tournament (16 Posts- Completed)

Longest Post Series: Starcraft Co (151,738 characters). This one I'm having the most fun with and am excited to continue. We are 9 missions in out of 26, so there is still a lot in store. I'm going to pace myself with it though so I don't burn out.

Estimated Average Word Count per Post: 2,435 words (this is high because of a few really chunky initial posts, but I would like to keep it above 1000)

Estimated pages if published as a book: Roughly 400

I'm not sure how many people actively use the write.as platform. Since its just text I imagine its cheap to host, so the platform wouldn't need a large number to be successful. Biggest cost is probably labor keeping spam and abuse off the site. I'm not sure how prolific the bloggers on here are. I found one guy who has been blogging since 2019 and has over 2500 posts, but his posts are pretty thin, mostly just a bullet point summary of what he did that day. A random post was 1298 characters, so by that math he's got 3,245,000 characters total in content. We will have him as an aspirational target to beat until I find a more suitable rival.

Also, hyper-aspirational, Wheel of Time is about 25 million characters (roughly 5 million words with the prequel, rounding up to 1 sig fig). Might need some fellow writers to help if I ever want to tackle that, since the main author died and Sanderson came in clutch to finish it. Though I think Stephen King all in is over 10 million, but who knows if ghost writers helped with some of that.

I should note, when I talk about content and length, I'm not just looking to inflate stats. If I was, I could just spam bullshit. I want each post to be unique, edifying and have some sort of purpose.

Looking ahead, we've got Starcraft Co. I've held off on starting the Inbetweeners series, but that will be up soon and should go for quite a while. I have a lot of content I want to share there. Otherwise I'm interested in some more fiction and maybe also some real world analysis.

The hearing in El Centro ran past the hour, and still the line at the microphone did not shorten. One after another, residents of California's Imperial Valley — a place that grows a startling share of the lettuce, spinach, broccoli and carrots that fill American refrigerators between November and March — stepped forward to tell their elected supervisors what they thought of the largest data centre ever proposed in their state. They spoke about asthmatic children. They spoke about the Colorado River, the lifeline that has irrigated this desert for more than a century. They spoke about the speed of the thing: how a nearly one-million-square-foot complex, advertised by its developers as “designed exclusively for advanced artificial intelligence and machine learning operations,” had been ushered toward approval in a matter of months, on land pressed up against fields that have fed the country for generations.

On 16 June 2026, after that more-than-an-hour of testimony, the Imperial County Board of Supervisors did something that the artificial-intelligence build-out has rarely encountered. It stopped. The board declared a 45-day moratorium on data centres, ordered the formation of an advisory commission to rewrite the county's zoning rules, and effectively pulled the brake on a project it had been waving through only weeks earlier. The reporting on the reversal — by CalMatters, by the public-service newsroom inewsource and by KPBS — converged on the same picture: a rural county that had said yes, looked harder, and said wait. Four weeks later the board would extend that wait to a full year.

The wait is the story. Across two countries and at least one ocean of difference in regulatory tradition, the spring and summer of 2026 produced a sudden, legible pattern of agricultural communities pushing back against the machines that the rest of the economy insists they must welcome. The pushback is not abstract. It is about water that a tomato plant and a server rack cannot both drink. It is about land that cannot be both a field and a slab. And underneath all of it sits a question that the technology's velocity has outrun: when AI infrastructure starts competing directly with agriculture for the resources a community depends on to feed itself and others, what right does that community have to be asked first?

The Field Beside the Slab

To understand why a desert county in the southeast corner of California became a flashpoint, you have to understand what the Imperial Valley actually is. It is not marginal land. It is among the most productive agricultural regions on the planet, and its productivity is entirely manufactured — wrung from the desert by the single largest entitlement to Colorado River water held by any user on the river. The Imperial Irrigation District controls roughly 3.1 million acre-feet of that river annually, a volume that dwarfs the allocations of entire states. Around 97 per cent of it flows to farms. Those farms, in turn, supply a remarkable proportion of the United States' winter vegetables; by common reckoning the valley and its neighbours grow something on the order of two-thirds of the leafy greens Americans eat in the cold months, when almost nowhere else in the country can.

That water is not abstract or infinite. The Colorado River is in structural decline, oversubscribed by the compacts that divide it and diminished by two decades of drought and a warming climate. The Imperial Irrigation District holds some of the most senior rights on the river — rights accumulated over more than a century — and it has spent years and hundreds of millions of dollars in conservation deals to keep them while trimming its draw. Every acre-foot in the valley is accounted for, contested, and increasingly precious. Into that finely balanced ledger arrived a proposal on a scale the county had no precedent for: roughly 950,000 square feet of computing hall, rated at up to 330 megawatts, costed in the order of ten billion dollars, on 235 acres purchased so far, targeted to be running by the summer of 2028. Cooling it would take something like 750,000 gallons of water a day.

For most of a year, the developer's answer to the obvious question was reassuring. Since at least December 2025, the environmental case made publicly by Imperial Valley Computer Manufacturing — the entity through which the developer Sebastian Rucci is pursuing the project — rested on a pledge that the cooling water would be recycled wastewater sourced from the cities of Imperial and El Centro, and that the campus would not draw on the Colorado River at all. That pledge answered, in advance, the one objection capable of stopping the project. It allowed the campus to be discussed as a neighbour to agriculture rather than a competitor for its lifeblood.

Then the cheaper alternative failed. The Imperial Irrigation District rejected the company's water application in early May 2026, and the recycled-water negotiations collapsed. In June, days after the company's preferred candidate for the irrigation district's board lost in the primary, IVCM sued the district for 260 million gallons a year of Colorado River water — the very source it had promised never to touch. The company's filings characterise that draw as equivalent to a 160-acre farm, and as roughly 0.03 per cent of the Imperial Valley's supply.

Read that comparison slowly, because it is the whole argument in miniature. Two hundred and sixty million gallons is the water of a 160-acre farm: to cool the servers that would train and run artificial-intelligence models, water currently growing food would be converted into a coolant supply. The field would not merely sit beside the slab. The field would be retired so the slab could drink. And the 0.03 per cent is the developer's own framing rather than a neutral measurement. Share-of-total is how a marginal draw on a fully allocated system is always made to sound negligible: every acre-foot in the valley is already spoken for, so the question is not what fraction of the whole the new demand represents but whose existing use it displaces. On a river in structural deficit, small percentages of enormous numbers are precisely how the deficit was assembled.

Rucci has not confined his response to argument. IVCM sued the City of Imperial in federal court in January 2026. In April it filed a civil defamation suit against Jake Tison of Mt. Signal Media, against Jose Luis Olmeda Velez, executive director of the local nonprofit Comite Civico del Valle, and against the organisation itself. It sued the Imperial Irrigation District on 15 June. And in late July, Courthouse News reported, it went after the county's moratorium in court as well. The water case is live and unresolved. But the sequence has already established something the valley will not forget: an environmental assurance survives exactly as long as the cheaper option holds. Here was the future, asking the present to fallow itself, having spent a year promising it would not.

A Thirst Measured in Towns

The reason cooling water became the hinge of so many of these fights is physical, not rhetorical. A hyperscale data centre at full load is one of the thirstiest single structures a community can host. Independent analyses of the sector converge on a striking range: a large facility can consume somewhere between one and five million gallons of water a day, a draw comparable to a town of ten thousand to fifty thousand people materialising overnight on the edge of a farming district. The thirst is greatest where evaporative cooling is used — the technique of running water across hot equipment and letting it evaporate to carry the heat away. That evaporation is the point and the problem: the water does not come back. It leaves the watershed as vapour.

The numbers the industry itself reports make the scale concrete, and the direction of travel matters more than any single figure. Google's 2026 environmental reporting disclosed that the company consumed 10.9 billion gallons of water in 2025 — up 34 per cent year on year, and more than double its 2021 level. Set that against the 7.8 billion gallons its data centres withdrew in 2024, roughly 78 per cent of it consumed through evaporation rather than returned to the local system, and the shape of the problem is clear. A growth rate of that order, at a company that has spent two decades optimising for efficiency and has every commercial reason to bring the number down, is the story.

Google's answer to its own consumption figure is replenishment: projects returning water to stressed basins, which delivered roughly 4.5 billion gallons in 2024, about 64 per cent of freshwater consumption, against a stated target of 120 per cent by 2030. It is a serious commitment, seriously funded, and it is also a basin-level accounting claim rather than a return of water to the specific watershed drawn down. A wetland restored in one catchment does not put water back into the aquifer beneath the field next door. In Mesa, Arizona, a single proposed campus was projected to use up to 1.75 million gallons a day — a figure local officials compared to a meaningful fraction of the city's entire daily supply. The International Energy Agency, cited repeatedly by Canadian protest organisers, estimated that data centres worldwide consumed around 140 billion litres of water in 2023, before the current AI build-out had truly accelerated.

There is a genuine technical counter-current here that honest reporting must acknowledge. The industry knows water has become its political vulnerability, and many new builds are now specifying air-cooled, adiabatic or closed-loop liquid systems designed to slash evaporative loss — in some cases choosing the more expensive option precisely because the reputational cost of consuming water has come to exceed the savings. Closed-loop designs can dramatically reduce ongoing draw. But two facts blunt that reassurance. The first is that even reduced consumption is consumption, and in a closed basin like the lower Colorado, marginal new demand competes directly with everything already drawing on the same finite pool. The second is that water saved at the cooling tower is often water spent at the power plant instead, because the electricity these facilities devour is itself frequently produced by water-intensive generation. The thirst does not vanish. It moves.

That is why the op-ed published by Civil Eats on 24 June 2026, under the unambiguous headline “AI Data Centers Are a Threat to Local Food Systems,” landed with such force in agricultural circles. The piece argued that the extraordinary water, land and electricity appetites of hyperscale facilities place them in direct structural competition with the farms around them — not as an unfortunate side effect but as a defining feature of how and where the industry chooses to build. It opened not with a statistic but with a man named Carl Stover in Loudoun County, Virginia, listening to the unbroken hum of a newly built AI data centre beside his house, on ground where neighbours had recently tended tomato vines and pepper plants in a community garden now buried under concrete and server racks. The image — a vegetable plot paved into a machine hall — is the thesis. The harvested field of the title is the one that gets harvested only once, and never grows back.

The Other Appetite: Watts and Air

Water is the most visceral of the three resources, because a farmer can picture the river dropping. But the energy appetite is arguably the more transformative, and it is where the agricultural conflict acquires a second front. Training and running frontier AI models is astonishingly power-hungry, and the facilities now being proposed are sized accordingly — not in the megawatts of an industrial park but in the gigawatts of a regional grid. The Olds complex in Alberta is the clearest illustration: ten data-centre buildings of one hundred megawatts apiece, a gigawatt of computing load in total, paired not with a connection to the existing grid but with a dedicated 1.4-gigawatt natural-gas power plant built on site to feed it. That is generation on the scale of a major utility station, erected beside a farming town because the public grid cannot supply the load fast enough.

For an agricultural community, on-site fossil generation imports a hazard that has nothing to do with computing and everything to do with breathing. The Imperial Valley already carries among the highest paediatric asthma rates in California, a burden tied to its air quality, and residents at the El Centro hearing returned again and again to the cost of treating wheezing children. A representative of a local homebuilder told the board, in testimony reported by CalMatters, that it is expensive when families pay for emergency-room visits. When the cooling load is met by burning gas — or when the electricity is drawn from a grid that fires up gas peakers to cover the new demand — those emissions settle over the same fields and the same lungs. The competition is no longer only over the water in the soil. It is over the air above the crop.

There is a quieter form of competition here too, one that operates through price and queue rather than smokestack. Every gigawatt an AI campus claims is a gigawatt of grid capacity, transmission and generation investment that is not available, or not available as cheaply, to everyone else drawing on the same system — including the irrigation pumps, cold-storage warehouses and processing plants on which a modern food system runs. Electricity, like Colorado River water, is a shared and finite pool in any given region at any given moment. A data centre that jumps the interconnection queue with a billion-dollar cheque does not merely add demand; it reorders the priorities of an entire regional energy system around its own clock. The farm waits behind the server.

Fourteen Cities and a Word

The same week the Imperial County supervisors hit pause, the resistance went national in a second country. In late June 2026, Canada saw what organisers described as its first co-ordinated day of action against AI data centres, with demonstrations stretching from British Columbia to New Brunswick. The Council of Canadians, working with a constellation of grassroots groups, framed the day around three resources that the build-out threatens: water, power, and — the word appearing explicitly in community communications about proposals in Alberta and Ontario — farmland.

In Vancouver, hundreds marched from the Vancouver Art Gallery to City Hall, demanding a halt to two data centres tied to a partnership between the telecommunications company Telus and the federal government, in a metro region already living under Stage 3 water restrictions. But the more telling fact was the geographic spread. This was not a single city's grievance. The Council of Canadians' organising materials gathered voices from Vancouver, from an Alberta anti-data-centre group, from Hamilton in Ontario, from the small Ontario community of Burk's Falls, and from Lorneville on the East Coast — a roll-call of places with little in common except a sudden, shared sense of being approached by something enormous they had not invited.

The march did not translate into policy. On 16 July, Vancouver's council defeated by seven votes to three a motion from the OneCity councillor Lucy Maloney that would have required third-party assessment of water use, greenhouse-gas emissions, air contamination, noise and heat before any new data centre was approved. The defeat followed the majority held by Mayor Ken Sim's ABC party. Residents, union organisers, trade representatives and neighbourhood organisations had spoken in support, several observing that the city was at that moment under Stage 3 restrictions — a council declining to require a study of industrial water consumption in a month when its own residents were forbidden to fill a paddling pool.

Five days later the same council did something stranger. On 21 July it voted unanimously to reverse its own decision of 14 July to send the Telus-Westbank rezoning application for 111 East 5th Avenue and 2060 Quebec Street to a public hearing, returning the file for further consultation and leaving the ultimate decision to whichever council is elected in October 2026. A body that had just refused to require environmental assessment of data centres nonetheless declined to proceed to a hearing on one, and handed the question to its successors. That is less incoherence than instinct: a council reaching for delay because delay is the only instrument within reach, the same improvised brake as a moratorium applied by another route, in the absence of any standing to say no. Time is the resource an accelerated approval takes first, and buying some of it back is the first thing a community without power reaches for.

The Canadian fights have their own brutal specificity. In the town of Olds, Alberta — a community of roughly ten thousand people, home to an agricultural college, ninety kilometres north of Calgary — a company called Synapse proposed a complex described as potentially the largest AI data centre in the country: ten data-centre buildings of one hundred megawatts each, fed by a dedicated 1.4-gigawatt natural-gas power plant, on a single parcel, representing a private investment reported at around ten billion dollars. The plant would be among the largest in the province. According to coverage by CBC and The Energy Mix, the campus would sit on land in the town's northeast that was rezoned from agricultural to light industrial, with hundreds of homes within eight hundred metres of the site and some residences a short walk away. One resident told CBC the proposal was “consuming” the lives of those living next door. When the Alberta Utilities Commission opened a proceeding, roughly 685 individuals filed for and were granted standing — an extraordinary number for a rural application, and a measure of how directly the thing pressed on people's lives.

The regulator's own conduct underlined the speed problem at the heart of all this. The commission closed Synapse's original power-plant application altogether, finding it had not met the minimum standards required even to begin formal regulatory review, while leaving the door open to resubmission. Synapse resubmitted in early April 2026 and the process began again. In mid-June the commission suspended it, pending the company's answers to a second round of questions. Written evidence from intervenors was due by 4 August, ahead of a hearing.

The Town of Olds' chief administrative officer has drawn the obvious conclusion from that sequence, and it deserves to be met rather than ignored: the closure of the original application, he has argued, proves the process works. It is the strongest available counter to the claim that institutions are improvising, and it is half right. A regulator that refuses a filing failing its own minimum standards is a regulator doing its job, against a well-resourced applicant in a hurry. But what the process caught was a paperwork failure — an application incomplete on its face, rejected without anyone having to weigh a single competing interest. Nothing in that outcome demonstrates the harder capacity: to assess a gigawatt of generation and ten machine halls against the water, air and land of the community that will live beside them, on a timetable that community can participate in. Rejecting an incomplete form and adjudicating a contested future are different competences, and only one of them has so far been shown. That is the texture of 2026: institutions designed for a slower world, scrambling to evaluate gigawatt-scale, billion-dollar proposals on timelines dictated by an industry racing its competitors and its own valuations.

The Country-Wide Democratic Gap

The person who has named this condition most precisely is Anne Pasek, an associate professor of media, culture and the environment at Trent University in Ontario, who studies the climate impacts of the technology industry. Speaking in the Canadian press in support of a proposed data-centre moratorium in Hamilton — a pause that, to the best of her knowledge, would have made that city the first in Canada to impose one and to pledge a locally determined regulatory framework — Pasek diagnosed the underlying disorder with unusual clarity.

“Part of the reason why data centres have been so wildly unpopular,” she told Global News, “is because of a lack of transparency from the industry and a lack of clarity from regulatory bodies.” Cities, she argued, are “really uniquely positioned to help close that democratic gap and put this conversation on better footing.” Elsewhere in the Canadian coverage, the pattern she describes was rendered as a country-wide democratic gap: a structural mismatch in which communities lack the information, the processes and above all the time to assess what hyperscale AI infrastructure means for the land, water and food systems around them before the decisions affecting those systems are made.

Hamilton did not become that first city. On 15 July 2026 its council rejected the one-year moratorium by ten votes to six, to boos from the public gallery. The councillors who voted it down argued that a pause could threaten investment in a city whose industrial base is already flagging, and that a blanket bylaw would catch smaller proposals with research components alongside the hyperscale campuses the measure was aimed at. Burlington, along the same lakeshore, likewise voted down a moratorium of its own while municipal staff continued to review a proposal. Mississauga went the other way: on 29 July its council voted eleven to nil to support a one-year pause, with staff now drafting the bylaw and returning for final approval in September 2026. If it is adopted it will likely be the first data-centre moratorium in Canada, and observers expect it to prompt a wave of similar bylaws in neighbouring municipalities.

That mixed record is stronger evidence for Pasek's diagnosis than a clean run of victories would have been. Four Canadian councils faced versions of the same question inside a fortnight and answered it four different ways. The variable was not the merit of the proposals, or the vulnerability of the watersheds, or the depth of local feeling — Hamilton's gallery booed as loudly as Mississauga's applauded. The variable was the composition and the appetite of the council in the room. Where a community holds no standing to refuse, whether it is heard at all is contingent on who happened to win the last municipal election, and on how a dozen or so people weigh an industrial tax base against a watershed on a given afternoon. A protection that depends on the arithmetic of a particular chamber on a particular Wednesday is not a right. It is a run of luck, and that it ran in Mississauga's favour and against Hamilton's, Burlington's and Vancouver's within the same three weeks is the clearest measurement of the democratic gap yet taken.

Canada's own national posture has widened that gap rather than closed it. When the federal government launched its national artificial-intelligence strategy in June 2026, reporting by Canada's National Observer noted that the plan included no new protections for water or climate even as it called for a substantial build-out of public-private AI data-centre capacity by the end of the decade. The same outlet's investigations found that a large share of data-centre sites planned in Alberta sit in areas of high water stress. A strategy that promises infrastructure while declining to legislate safeguards for the resources that infrastructure will consume does not close a democratic gap. It institutionalises one, and leaves the whole question where the summer found it: on the desks of municipal councillors with no statutory tool for the job beyond delay.

What the Existing Frameworks Were Built to Do

It is tempting to assume that the law already contains the answer — that environmental review, zoning and farmland protection exist precisely to mediate conflicts like these. The reality is that the frameworks were built for a different shape of problem, and the AI build-out has found the seams.

Consider California, where the tools are unusually strong on paper. The California Environmental Quality Act, CEQA, requires public agencies to study and disclose the significant environmental effects of projects they approve, including the conversion of agricultural land and impacts on water supply, and to consider alternatives and mitigation. The Williamson Act, in force since 1965, lets landowners contract with counties to keep land in agriculture in exchange for lower tax assessments, and today protects well over sixteen million of the state's thirty million acres of farm and ranch land. Mapped prime farmland and Williamson Act contracts can themselves trigger the deeper levels of CEQA review. These are not trivial protections. Indeed, it was a CEQA argument — that the Imperial project had sidestepped the review its proximity to homes and farmland demanded — that the City of Imperial reportedly raised in litigation, and CEQA-style concerns that helped justify the county's pause.

But notice what CEQA does and does not do. It compels disclosure and study; it does not compel consent. An agency can identify a significant impact, declare it unavoidable, adopt a statement of overriding considerations, and approve the project anyway. The act gives the public a right to be informed and to comment. It does not give the surrounding community a right to refuse. And crucially, the entire apparatus depends on time — the time to scope, study, circulate a draft, take comment, and respond — which is exactly the resource an accelerated approval is engineered to compress. When a county processes a billion-dollar facility at the speed Imperial initially did, the question is not whether CEQA exists but whether it was allowed to function.

The zoning story is starker still. In Olds, the path to the data centre ran through the reclassification of agricultural land to light industrial — an administrative act that, once complete, reframes a field as a site. Zoning is the most powerful lever a local government holds over land use, and it is also the most quietly exercised. A rezoning hearing rarely draws the crowd that 685 statements of standing represents; the fact that this one did is the exception that reveals the rule. Most farmland is not lost in a dramatic vote. It is lost in a line on a map that almost nobody reads in time.

The Idea of Asking First

If existing law guarantees disclosure but not consent, the obvious place to look for a richer standard is the body of practice that has grown up precisely around consent for large projects on contested land. The most developed of these is the principle of free, prior and informed consent — FPIC — articulated most fully in the context of Indigenous rights and endorsed by the United Nations in 2007 as a standard to be met before projects proceed on Indigenous lands.

The phrase is worth unpacking word by word, because each word answers one of the failures the data-centre fights have exposed. Free means consent given without coercion, manipulation or inducement — not the cornered acquiescence of a community told the jobs will go elsewhere if it hesitates. Prior means consent obtained before the project begins, before the land is rezoned and the water is contracted, not ratified after the bulldozers arrive. Informed means the community has access to the best available scientific, environmental, social and financial information — the actual water draw, the actual emissions, the actual grid impact — in time to weigh them. And consent means precisely what the existing frameworks withhold: the genuine possibility of saying no.

FPIC in its strict legal sense applies to Indigenous peoples, and the data-centre build-out is colliding with that right directly. The clearest case is unfolding in northern Alberta, around a project called Wonder Valley — an AI data-centre park pitched in late 2024 by the celebrity investor Kevin O'Leary and reported at seventy billion dollars. The Sturgeon Lake Cree Nation sought judicial review of Alberta's decision to grant the Municipal District of Greenview a licence under the Water Act permitting withdrawal from the Smoky River, on the ground that the Crown's duty to consult had not been discharged before the licence was issued. In April 2026 the Nation had already described the province's and O'Leary's handling of the file as an “absolute failure” of consultation. Arguments were heard in an Edmonton courtroom in June.

O'Leary's side then attempted to stop the review being heard at all. On 13 August 2026, Justice P. K. Johal of the Court of King's Bench dismissed that attempt, ruling the hearing should proceed on the question of whether the First Nation was consulted before the licence was granted. Set beside every other fight in this account, the case is the sharpest available demonstration of what FPIC is actually for. Nowhere else is a court being asked to adjudicate the words themselves — whether consultation was prior, whether it was informed — because nowhere else does the community hold a right that would give a court anything to adjudicate. In Imperial County and in Hamilton the question was what a board of elected officials would choose to do. Here the question is what the Crown was obliged to do, and the distance between those two questions is the whole distinction between a right and a courtesy. It is worth noticing, too, what the developer's first move was: not to demonstrate that consultation had occurred, but to argue that the question should never be put.

The underlying logic has a wider life beyond Indigenous law. The development-finance world has long used a softer cousin: the requirement, embedded in World Bank standards, that operators of certain projects secure “broad community support” before proceeding. The corporate world calls something adjacent the social licence to operate — the informal but consequential permission a community grants, or withholds, that no permit can substitute for. The lesson the mining and energy sectors learned the hard way, after years of blockades and reputational wreckage, is that a project lacking community consent is financially and operationally fragile no matter how many official approvals it holds.

The AI infrastructure sector is now learning the same lesson, parcel by parcel, in real time. The Imperial moratorium, the 685 Albertans claiming standing, the marchers in fourteen Canadian cities, the eleven-nil vote in Mississauga — these are communities asserting, in the absence of any formal right to consent, the social-licence veto they do possess: the power to make a project so contested, so litigated and so politically radioactive that the official approvals stop being worth what they cost. Hamilton, Burlington and Vancouver are the other half of the same lesson. Where that veto is the only instrument available, it works only when a majority of a particular council happens to be willing to pick it up. A social licence is not a legal one. It can be declined on a community's behalf by the very body elected to represent it, and from that there is no appeal.

What Meaningful Prior Consultation Would Actually Require

To say a community has a right to be asked first is easy. To specify what asking would actually entail is harder, and it is where the argument has to earn its keep. Drawing the threads together — the disclosure logic of CEQA, the consent logic of FPIC, the durability logic of social licence — a meaningful standard of prior consultation for hyperscale AI infrastructure on or beside agricultural land would have at least four properties that current practice lacks.

The first is full, early and binding disclosure of resource demand. Before any rezoning or approval, a developer would have to put on the public record the facility's projected water consumption — not withdrawal alone but consumptive, evaporative loss — its peak and average electricity draw, the source of that electricity and its own water and emissions footprint, and the impact on the local grid and watershed, all stated in units a farmer can compare against a crop. The word carrying the weight in that sentence is binding. Imperial Valley Computer Manufacturing disclosed its water plan early, publicly and at length: the cooling water would be recycled municipal wastewater, and the Colorado River would not be touched. So far as anyone has shown, the statement was true when it was made. It simply had nothing holding it in place, and when the cheaper route closed the company went to court for 260 million gallons a year of the river it had promised to leave alone. A disclosure regime that yields assurances a developer may abandon the moment they become inconvenient is a public-relations exercise with a filing cabinet attached. What a community needs on the record is not a statement of intent but a condition of approval: a number the permit is contingent upon, and which cannot be revised upward in a courtroom eighteen months later.

The second is time proportionate to scale. A proposal that will draw water at the rate of a small city and power at the rate of a small province cannot be evaluated honestly on a timeline measured in weeks. Imperial's improvised brake did become a real one: on 14 July 2026 the board of supervisors voted unanimously to extend the initial 45 days by ten months and fifteen days, carrying the moratorium through to 14 July 2027 and blocking the issuance of any data-centre permit for the duration. It acted under section 65858 of the California Government Code, on a finding of a current and immediate threat to public health, safety and welfare — a provision written for emergencies, pressed into service because nothing in the ordinary machinery of land-use law simply hands a county a year in which to think. Hamilton's council, offered a comparable pause, declined to take it. The moratorium is a crude instrument in either case, and its availability is wholly discretionary. A standing requirement that facilities above a defined water or power threshold trigger a mandatory minimum review period would convert that improvised brake into a designed one, and remove the question of whether a community gets to think at all from the discretion of the body deciding what it is allowed to think about.

The third is a genuine seat for food systems and the people who run them. The democratic gap is partly a representation gap: the boards that approve these projects are weighted toward growth, and the interests of farmland, watershed and food security have no guaranteed voice. It is notable that in the Imperial fight, a state senator moved to expand the local air-pollution board to include public-health and environmental representatives — an institutional fix aimed squarely at who gets to sit at the table.

Imperial County, in the event, went further and built something close to the body this argument calls for. Alongside the extension, the supervisors seated an Imperial County Data Center Advisory Committee, charged with bringing recommendations on data-centre zoning and building policy by January 2027. Its composition is the interesting part: two supervisors, John Hawke and Peggy Price; two community-interest representatives; one representative each from local government, the environmental sector, organised labour, healthcare, education, the nonprofit sector, business and the energy industry; and one from the Imperial Irrigation District, holder of the water at the centre of the entire dispute. County counsel Geoffrey Holbrook noted that the moratorium prevents permits being issued while the committee does its work, and that whatever the process produces “could change the different land-use regulations.”

That is a table with the right people around it. It is also a table assembled after the fight, by a county that had already approved the project once, under the pressure of a hearing that ran past the hour. Nothing obliged Imperial to build it, and nothing obliges the next county to build one. The distinction between a designed brake and an improvised one was never about the quality of the institution eventually produced; it is about whether that institution exists before the decision, or only after a community has spent months forcing it into being. Prior consultation worth the name would give agricultural and food-system stakeholders standing by right, not by the accident of who showed up to testify and how loudly.

The fourth, and most contested, is some real approximation of the power to refuse. Disclosure and time and representation still leave the final word with an agency that can override them all. The harder claim — the one the social-licence experience supports — is that for projects of this magnitude on this kind of land, a community's considered refusal should carry weight that an approving body cannot simply note and dismiss. Whether that takes the form of a supermajority requirement, a binding local referendum, a true consent threshold for affected agricultural districts, or something else, the principle is the same: the community that will lose the water and the land should be able to decide that it would rather keep them.

The Harvest That Does Not Come Back

There is a temptation, in writing about AI infrastructure, to frame the conflict as progress against nostalgia — the inevitable future versus people who would rather it didn't arrive on their doorstep. That framing is wrong, and the events of the summer of 2026 expose why. The Imperial Valley farmer defending Colorado River water is not resisting the future. She is defending a food system that the rest of the country depends on without thinking about it, in the months when no other system can do the job. The Albertans claiming standing before their utilities commission are not Luddites; they are residents asking, reasonably, what a gigawatt of gas turbines a few hundred metres from their homes will do to the air their children breathe. The marchers in fourteen cities are not against computation. They are against the proposition that computation's needs should silently outrank the needs of the land that feeds them, decided by processes too fast and too opaque for them to touch.

What makes the agricultural case distinct from the general backlash is the irreversibility at its core. A data centre can be decommissioned; its water can, in principle, be returned to other uses. But farmland converted to industrial land does not convert back, and water evaporated into the sky over the desert does not return to the river. Senior Colorado River rights accumulated over a century are not reissued. The phrase that keeps recurring — the harvested field — names a harvest that happens only once. The crop is the field itself.

The deepest finding of this strange convergent season is not that AI infrastructure is uniquely villainous. It is that the velocity of the build-out has outrun the institutions meant to govern it, opening the country-wide democratic gap that Anne Pasek named — a gap in which the decisions that determine whether a region keeps the means to feed itself are made faster than the people of that region can be meaningfully consulted. Imperial County's year-long pause, the 685 Albertans, the fourteen cities, the eleven-nil vote in Mississauga, a judge in Edmonton ruling that the question of consultation must at least be heard: these are not the end of the build-out. Neither are Hamilton's ten to six, Burlington's rejection, or the seven Vancouver votes against a demand as modest as an independent assessment of the water. Read together, the two columns say something more precise than either says alone. Through the summer of 2026, whether a community got to think before it was built upon turned not on any right it held but on the arithmetic of the chamber it happened to sit in. Communities are inventing, in the absence of any law that grants it, the right to be asked first — and discovering, parcel by parcel and hearing by hearing, that a protection some of them win and others are refused by comparable institutions in the same month is not yet a right at all. It is worth fighting for precisely because the harvest it protects does not come back.

References

1. Garcia, Lauren. “This rural, California county approved a massive new data center. Then it changed its mind.” CalMatters, June 2026. https://calmatters.org/environment/2026/06/imperial-county-data-center/
2. inewsource. “Imperial County's data center moratorium will last one year.” 14 July 2026. https://inewsource.org/2026/07/14/data-center-moratorium-imperial-county/
3. KPBS Public Media. “Imperial Valley data center developer files lawsuit seeking access to Colorado River water.” 15 June 2026. https://www.kpbs.org/news/environment/2026/06/15/imperial-valley-data-center-developer-files-lawsuit-seeking-access-to-colorado-river-water
4. Courthouse News Service. “Desert data center developer challenges county moratorium in court.” July 2026. https://www.courthousenews.com/desert-data-center-developer-challenges-county-moratorium-in-court/
5. Tom's Hardware. “California's largest AI data center project suing for access to Colorado River water.” 2026. https://www.tomshardware.com/tech-industry/artificial-intelligence/californias-largest-ai-data-center-project-suing-for-access-to-287-million-gallons-of-colorado-river-water-0-03-percent-of-imperial-valleys-supply-plaintiffs-claim-project-equivalent-to-160-acre-farm-amidst-about-jobs-and-reallocation-of-farmland
6. Imperial Irrigation District. “Water” and “Water Conservation.” 2026. https://www.iid.com/water
7. Civil Eats. “Op-ed: AI Data Centers Are a Threat to Local Food Systems.” 24 June 2026. https://civileats.com/2026/06/24/op-ed-ai-data-centers-are-a-threat-to-local-food-systems/
8. Data Centre Magazine. “What Google's Environmental Report Says About Data Centres.” https://datacentremagazine.com/news/google-environmental-report-2025-the-data-centre-impact
9. The Council of Canadians. “From the Ground Up: Fighting AI Data Centres / National Day of Action.” 27 June 2026. https://canadians.org/event/from-the-ground-up-fighting-ai-data-centres/
10. Canada's National Observer. “Vancouver council squashes motion to regulate AI data centres.” 17 July 2026. https://www.nationalobserver.com/2026/07/17/news/vancouver-ai-data-centres-motion
11. Daily Hive Urbanized. “Telus-Westbank AI data centre decision punted until after Vancouver election.” July 2026. https://dailyhive.com/vancouver/m3-mount-pleasant-vancouver-telus-westbank-ai-data-centre-decision-deferred
12. CBC News. “Data centre proposal in Olds 'consuming' the lives of some living next door.” 2026. https://www.cbc.ca/news/canada/calgary/residents-pushback-data-centres-olds-9.7183912
13. The Energy Mix. “Application Delayed as 680 Local Residents Gain Standing on Olds, Alberta Data Centre Proposal.” 2026. https://www.theenergymix.com/application-delayed-as-680-local-residents-gain-standing-on-olds-alberta-data-centre-proposal/
14. The Albertan. “AUC's closure of data centre application proves process works, says Olds CAO.” 2026. https://www.thealbertan.com/olds-news/aucs-closure-of-data-centre-application-proves-process-works-12003027
15. Global News. “Hamilton advances proposed data centre moratorium in pushback to rapid AI buildout.” June 2026. https://globalnews.ca/news/11909502/hamilton-advances-proposed-data-centre-moratorium-in-pushback-to-rapid-ai-buildout/
16. CBC News. “No ban against data centre development after Hamilton council votes down bylaw.” July 2026. https://www.cbc.ca/news/canada/hamilton/data-centre-ban-voted-down-9.7271641
17. Canada's National Observer. “Mississauga votes for data centre moratorium that could set Canadian precedent.” 29 July 2026. https://www.nationalobserver.com/2026/07/29/news/mississauga-council-voted-data-centre-moratorium
18. CBC News. “As city staff review data centre proposal, Burlington, Ont., councillors vote down data centre moratorium.” 2026. https://www.cbc.ca/news/canada/hamilton/data-centre-moratorium-burlington-9.7282901
19. CBC News. “First Nation in court to challenge proposed Wonder Valley AI data centre project in northern Alberta.” June 2026. https://www.cbc.ca/news/canada/edmonton/sturgeon-lake-cree-nation-alberta-wonder-valley-9.7228016
20. The Energy Mix. “O'Leary Loses, First Nation Wins in Bid for Judicial Review of Data Centre Water Permits.” August 2026. https://www.theenergymix.com/oleary-loses-first-nation-wins-in-bid-for-judicial-review-of-data-centre-water-permits/
21. Canada's National Observer. “'Absolute failure': First Nation slams Alberta and Kevin O'Leary's data centre moves.” 24 April 2026. https://www.nationalobserver.com/2026/04/24/news/wonder-valley-kevin-o'leary-alberta-first-nation
22. Canada's National Observer. “Canada's AI strategy includes no new protections for water or climate.” 4 June 2026. https://www.nationalobserver.com/2026/06/04/news/canada-ai-strategy-water-climate
23. Canada's National Observer. “Three quarters of data centre sites planned in Alberta are in high water stress areas.” 23 March 2026. https://www.nationalobserver.com/2026/03/23/investigations/alberta-data-centres-water-supply
24. California Department of Conservation. “Land Conservation (Williamson) Act — Questions and Answers.” https://www.conservation.ca.gov/dlrp/wa/Pages/LCA_QandA.aspx
25. United Nations / FAO. “Free, Prior and Informed Consent — An Indigenous Peoples' right and a good practice for local communities.” 2016. https://www.un.org/development/desa/indigenouspeoples/publications/2016/10/free-prior-and-informed-consent-an-indigenous-peoples-right-and-a-good-practice-for-local-communities-fao/

---

Tim Green UK-based Systems Theorist & Independent Technology Writer

Tim explores the intersections of artificial intelligence, decentralised cognition, and posthuman ethics. His work, published at smarterarticles.co.uk, challenges dominant narratives of technological progress while proposing interdisciplinary frameworks for collective intelligence and digital stewardship.

His writing has been featured on Ground News and shared by independent researchers across both academic and technological communities.

ORCID: 0009-0002-0156-9795 Email: tim@smarterarticles.co.uk

Listen to the free weekly SmarterArticles Podcast

I've been working in the construction industry for more than 20 years now. It's a traditional sector, very much male-dominated. But that has never deterred me from it. I also went to engineering school, and the majority of the students were male (I think less than 15% were female).

I'm using an outdated binary classification here, I know. If women are considered a minority in construction, wait until you look at non-binary gender data. It's bad.

What inspired me to write about this was a meeting I attended the other day at my company. It was a senior management meetup between my company and a major subtrade we frequently do business with. It was basically for introductions and discussions on future opportunities. I entered the boardroom, and after I introduced myself to everybody and everyone had arrived, I looked around and realized I was the only woman in there!

There were 18 people in total in that room, giving us a 5.5% diversity representation.

I'm not saying there are no women in my company. There are! In my department, there are 17 of us in total, of which 6 are women, so about 35%. I've heard that women make up close to 15% of the total construction workforce across Canada. That's low!

My company is not the worst out there. We have inclusive diversity groups and initiatives, with regular webinars, events, and celebrations (like Pride Month, the Indigenous Employee Circle, and Black History Month, among others) to foster an inclusive workplace.

But I am always surprised when I look at the senior leadership and find it hard to spot a woman.

Anyway, this is a bit of a rant, I guess.

I don't exactly feel excluded, but... do I?

I don't have experience with how other industries are doing in terms of diversity in senior roles and high-level management (or anywhere else, to be honest). But I think we still have a long way to go!

Post 07 of #Blaugust #NoisyMusings #diversity

Starcraft Co. 9- Defense

This was the first mission where Starcraft Co. forces had to hold out against wave after wave of invaders. You have a base, but ironically, the command center has been destroyed and you lack the resources to build another. Its more about conserving resources, repairing, and avoiding attrition. Don't worry, a proper base will be coming... in a few missions.

---

Until then, the Zerg had been dangerous in the way a wildfire was dangerous. They spread, consumed and killed, but nobody had seen evidence that the outbreaks appearing across Terran space were being directed toward any larger objective. The zerglings in Varga's basement had attacked whatever was nearby. The infestations aboard the Horizons Beyond had nested wherever food and shelter were available. The ultralisks on Ardea had behaved like enormous invasive predators. Even the unprecedented swarm at Cressida had appeared more interested in feeding from the orbital city's power infrastructure than accomplishing anything strategically meaningful. The UIC's classified investigation had revealed that someone was breeding and experimenting on Zerg, but that was human strategy imposed upon animals. The Zerg themselves remained creatures following instincts they barely seemed to understand. Then they attacked Kestral Battery.

The installation occupied a barren plateau on a UIC perimeter world that had been strategically important enough to fortify and peaceful enough to neglect. Kestral's enormous energy complex supplied the primary power for the orbital defense network protecting the planet. Take the battery offline and the defense platforms above would retain emergency power for a time, but their heavy weapons, shields and long-range detection systems would rapidly degrade. Anyone wanting to attack the planet would therefore have an obvious first target. Unfortunately, the UIC garrison assigned to defend it existed more impressively in procurement reports than reality. Years without a serious attack had hollowed the base out. Personnel had been reassigned without replacement. Equipment disappeared into warehouses or onto black markets. Maintenance schedules became suggestions. Funds allocated for defensive expansion somehow produced neither defensive expansion nor funds. The Zerg had found it anyway. They had not wandered into Kestral. They had attacked it repeatedly. Wave after wave had come from the wasteland beyond the installation, withdrawing or dying before another arrived hours later. The garrison had already lost its command center, several barracks and most of its heavy equipment. The surviving SCVs could repair existing structures, but the destruction of the command center had taken the logistical systems and resource stockpiles required to construct replacements. The base was consuming ammunition faster than supply transports could safely land it. Intelligence estimated that another two or three serious attacks would finish the garrison.

The call to StarCraft Co. contained no negotiation. A UIC colonel appeared on the commander's screen and said, “We need you at Kestral immediately.” The commander asked what the government was offering. “Whatever it costs.” There was no time to argue over numbers. StarCraft Co. loaded two elite infantry teams aboard transports, added every weapon and crate of ammunition they could fit, and launched. They landed north of the main installation because the southern approaches were already contested. The first surprise came almost immediately. A secondary UIC supply post appeared on their tactical map less than five kilometers from the landing zone, but nobody at Kestral had mentioned it. The reason became obvious when StarCraft Co. reached the compound. Its long-range communications array had been destroyed during the opening attacks, leaving the personnel there isolated from the rest of the battlefield. Several marines remained alive alongside a handful of Goliath combat walkers, all of them waiting for orders that hadn't come. The officer commanding the post initially demanded to know under whose authority a private military company was giving instructions to UIC personnel. The commander forwarded the emergency authorization from planetary command. That ended the discussion. Ten minutes later, StarCraft Co. had acquired reinforcements.

They reached Kestral just as the next wave arrived. Zerglings came first, pouring across the southern approaches in a mass large enough to look almost liquid from the bunker cameras. Hydralisks followed behind them, stopping beyond the first defensive line to launch spines into firing ports and exposed infantry. The defenders barely had time to recognize the arriving StarCraft Co. transports before the mercenaries joined the line. Marines filled damaged bunkers. Goliaths moved between defensive positions and began firing over them. StarCraft Co.'s elite infantry established overlapping fields of fire across the approaches while medics dragged wounded government troops away from the firing line. The attack lasted nineteen minutes. When it ended, hundreds of Zerg corpses covered the ground outside the installation. Kestral was still standing. Barely. The commander toured the defenses immediately afterward and discovered exactly how bad the situation was. Two bunkers had partially collapsed. Another had no functioning ventilation system. Ammunition stores were dangerously low. Automated turrets existed on the base schematic but had apparently never been installed. The surviving SCVs could repair damage, but without the command center there was no practical way to rebuild the base. So they stopped thinking about rebuilding it. They fortified what remained.

The SCVs reinforced damaged bunkers with whatever metal could be stripped from destroyed structures. Wrecked vehicles became barricades. Engineers dismantled sections of the ruined command center and welded the armor plating onto firing positions. Narrow approaches between buildings were deliberately obstructed to funnel zerglings into kill zones. Marines drove metal stakes into the ground and stretched industrial cable between them at heights intended to catch charging creatures. Fuel from destroyed vehicles was transferred into drums and buried along likely attack routes. Explosives intended for demolition work became command-detonated mines. Ammunition caches were distributed throughout the base so a destroyed storage building could not eliminate their entire supply. The commander imposed one rule on every defensive plan: nothing depended on stopping the Zerg. Anything strong enough to physically stop an ultralisk did not exist at Kestral. The defenses were designed to slow, divide and redirect. Every second a zergling spent climbing over wreckage was another second a marine could shoot it. Every hydralisk forced into a narrow approach became a target for three bunkers instead of one. Every obstacle existed to buy seconds.

They received forty-seven minutes before the next attack. This wave came from two directions. Zerglings hit the western perimeter while hydralisks approached from the south. The defenders allowed the zerglings through the first barricade before detonating buried fuel behind them. Fire severed the leading creatures from the rest of the swarm. Goliaths and marines destroyed the trapped group while the remaining Zerg tried forcing their way around the burning ground. On the southern line, StarCraft Co. snipers and heavy rifle teams concentrated on hydralisks as soon as they entered range. The attack collapsed before reaching the bunkers. The third wave came before repairs were finished. Then the fourth. The Zerg never attacked exactly the same way twice. That bothered the commander. One wave sent hydralisks first, forcing the defenders into cover before zerglings charged. Another divided around the main barricades and attacked the eastern maintenance sector that had previously seen little fighting. When the defenders shifted Goliaths east, the next attack concentrated on the west. The behavior wasn't brilliant. It didn't need to be. Something was testing them.

By nightfall, mutalisks appeared. The first group came low over the southern ridge while zerglings attacked the perimeter. The Goliaths immediately redirected their weapons skyward, leaving the ground troops without their heaviest support. That seemed intentional too. Mutalisks swept across the base, their glaive wurms ricocheting between clustered defenders. One pass killed four marines and wounded seven. StarCraft Co. responded by spreading infantry positions farther apart and relocating the Goliaths so no single ground attack could pin all of them simultaneously. The next mutalisk wave encountered overlapping missile fire before reaching the bunkers. Nobody slept for more than twenty minutes at a time. The SCVs became as important as the soldiers. The instant an attack ended, they moved onto the battlefield. Armor was welded. Bunker walls were patched. Firing ports were rebuilt. Wreckage was dragged into new barricades. Several SCVs worked while medics were still collecting casualties around them. One crew repaired the same bunker four times in eleven hours. The men started calling it Lazarus.

The next morning, the ground began shaking. The commander knew what that meant before reconnaissance confirmed it. Ultralisks. Three accompanied the next wave. The soldiers who had fought on Ardea understood exactly what was coming. Everyone else learned quickly. The first ultralisk hit an outer barricade and went through it without noticeably slowing. The second struck a bunker. Its kaiser blades penetrated the reinforced wall and ripped half the structure open. Marines escaped through the rear while Goliaths concentrated fire on the creature's legs. The third charged directly through a minefield. Two explosions tore chunks from its underside. It kept coming. StarCraft Co. used the same principle that had worked on Ardea. Nobody tried to stand against them. Positions abandoned the moment an ultralisk committed to attacking them. Goliaths and marines fired from the flanks, concentrating on joints while the creatures chased retreating targets. The problem was everything around them. Zerglings followed in the ultralisks' wake. Hydralisks fired between them. Mutalisks swept overhead whenever the Goliaths became occupied. The first ultralisk died twenty meters inside the perimeter. The second reached the ruined command center before concentrated fire shattered one of its forelegs. The third killed eleven people before it went down. Kestral held.

Another attack came six hours later. Then another. The battlefield surrounding the base became unrecognizable. Zerg corpses accumulated in heaps until SCVs had to push them aside to restore fields of fire. The air smelled of burned flesh, propellant and industrial coolant. Rifles failed from continuous use. Marines took weapons from the dead. Medics stopped distinguishing between StarCraft Co. personnel and government troops because there was no longer any practical reason to. The commander watched the numbers decline. Every attack cost the Zerg hundreds. Every attack cost Kestral people it could not replace. Eventually reconnaissance detected something different: movement across the entire southern horizon.

The final wave was enormous. Hundreds of zerglings formed the bulk of it, spreading across the wasteland around clusters of hydralisks. Several ultralisks moved at the front. Mutalisks circled overhead. There were enough organisms that the tactical computer stopped trying to display individual contacts and replaced them with a single red mass. Nobody needed to be told that Kestral could not survive a direct engagement, so the commander ordered everyone to abandon the outer perimeter. The defenders withdrew into the installation's inner structures. SCVs completed the last preparations and retreated with them. The Zerg advanced without resistance. That was the point. The outer approaches had been transformed during the previous hours. Every remaining demolition charge had been buried there. Fuel drums sat beneath wreckage. Damaged ammunition that could no longer be safely fired had been packed into improvised explosive caches. Mines overlapped with mines.

The ultralisks entered first. The commander waited. Zerglings flooded behind them. Hydralisks reached the perimeter. Still he waited. One government officer finally asked whether the detonators had failed. “Not yet.” More Zerg entered. The red mass filled the outer base. “Now.” Kestral disappeared behind fire. The first explosions triggered the second line of charges. Those detonated fuel stores. Burning wreckage ignited ammunition caches. Mines erupted beneath the swarm. One ultralisk vanished inside an explosion powerful enough to throw pieces of its armor across the installation. Another staggered forward missing half a leg. Hundreds of zerglings died in seconds.

The survivors kept coming. Every remaining weapon at Kestral opened fire. The wounded ultralisk reached the inner barricade and smashed through it. Goliaths fired until their autocannons glowed. Marines poured rounds into the breach. Hydralisk spines struck the bunkers like artillery. Mutalisks dove through missile fire and attacked the defenders from above. Lazarus finally died. An ultralisk struck the bunker broadside and collapsed the entire front wall. The marines inside kept firing through the opening until zerglings reached them. The defensive line folded inward. One position fell, then another. The commander committed the reserve. There were barely enough people left to call it one. They counterattacked through the maintenance yard, retaking a firing position that had been overrun minutes earlier. Medics carried rifles. Wounded marines fired from the ground. An SCV with a damaged manipulator arm charged a hydralisk and crushed it against a wall.

The last ultralisk reached the power installation. Every Goliath still functioning turned toward it. Missiles and autocannon fire struck the creature from three directions. It continued forward. Marines joined in. One leg failed. It dragged itself onward. The commander ordered the explosives beneath the access road detonated. The ground disappeared beneath the ultralisk. It fell into the crater. The defenders fired downward until it stopped moving. Then they turned around. There were still zerglings inside the base. The fighting became close and ugly. Squads no longer existed in any meaningful sense. Whoever remained alive formed groups around whatever weapons still worked. Marines cleared buildings room by room. Goliaths stomped through courtyards firing at anything moving. The last mutalisk was killed by a missile from a Goliath whose other weapon system had already been destroyed.

Eventually the gunfire became sporadic, then stopped. The Zerg were dead. For nearly ten minutes, nobody moved beyond what was necessary to treat the wounded. The commander began counting. Fewer than half of StarCraft Co.'s deployed personnel remained combat capable. Some of the missing were wounded. Many weren't. Of the UIC troops who had occupied Kestral when the attacks began, fewer than one in four were still alive. Two Goliaths remained operational. Most of the bunkers were destroyed. The base's command center was already gone. Nearly everything else was burning. But the power installation was intact. The orbital defense network remained online. They had held.

Then a sentry reported movement. Six contacts. Everyone still capable of carrying a weapon took a position. Six hydralisks emerged from the smoke. There were no zerglings with them. No mutalisks. No ultralisks. Just six hydralisks walking toward the ruined base. A few soldiers laughed. After what they had just survived, six hydralisks seemed almost insulting.

The first one fired. A marine's chest exploded. The spine crossed the distance faster than anyone expected. “COVER!” The second volley struck the barricades. These weren't normal hydralisks. They moved faster. Much faster. Their carapaces were thicker and darker, layered in ways the veterans had never seen. Gauss rounds struck them and shattered pieces of armor without penetrating deeply enough to stop them. Their spines hit with velocities that turned cover designed to withstand ordinary hydralisk fire into shrapnel. The defenders concentrated on one. It went down, then stood back up. Wounds were closing across its body. “WHAT THE FUCK ARE THESE THINGS?” Nobody knew.

The six creatures spread out. That was worse. Hydralisks normally attacked with aggression and crude coordination. These used cover. One fired while another repositioned. Two deliberately targeted a Goliath's damaged leg until the machine collapsed, then shifted fire toward the crew escaping it. The exhausted defenders suddenly found themselves in another battle, and people started dying again. One creature reached the interior of the base and killed three wounded marines before a medic emptied an entire magazine into its head at point-blank range. Another absorbed concentrated fire from seven rifles and continued advancing until a Goliath autocannon tore it nearly in half. Four remained, then three. The defenders learned. Head shots worked better. Destroying limbs worked better. Overwhelming damage worked best. One hydralisk took a grenade beneath its body and lost both legs on one side. It continued firing from the ground until a marine walked around the barricade and emptied his rifle into its skull.

Two remained. They almost reached the power station. The last Goliath stopped one. The other killed its pilot. Rourke's surviving infantry surrounded the final creature from three directions. It fired at one group. The others fired back. Its wounds began closing. They kept shooting. Armor broke. Tissue ruptured. The creature tried to retreat. That was something else nobody had ever seen. Mercer followed. The hydralisk turned and fired. A round struck Mercer in the shoulder and spun him around. Rourke's squad opened fire. The creature collapsed. They didn't stop when it stopped moving, when its head came apart, or until there was almost nothing left to regenerate.

Silence returned to Kestral. This time it stayed. One of the surviving government marines eventually walked toward the corpse. He nudged it with his boot and looked at the other five scattered across the ruins. “Hunter killers,” he muttered. Someone nearby heard him. The name stuck.

The UIC response fleet arrived fourteen hours later. It found a battlefield. Thousands of Zerg corpses surrounded the installation. The base itself looked almost indistinguishable from them. Fires still burned among collapsed structures. SCVs crawled through the ruins searching for survivors. Medics had converted the remains of a supply warehouse into a field hospital. The power installation was still operating. The UIC investigators were less interested in the thousands of dead Zerg than they were in six particular corpses. The commander watched them seal the remains inside biological containment units. “You've seen these before?” he asked. The intelligence officer from the clandestine briefing looked at him. “No.” That answer worried the commander more than yes would have.

The attack on Kestral changed the UIC's assessment of the crisis overnight. Zerg had not wandered onto the planet. They had identified a strategically significant target and attacked it repeatedly. They had varied their tactics. They had escalated when earlier attacks failed. Then, after the main swarm was destroyed, six organisms unlike anything previously documented had arrived. Someone—or something—was directing them. The UIC no longer treated StarCraft Co. as merely a preferred contractor. The retainer they offered was unprecedented. StarCraft Co. would receive guaranteed funding sufficient to maintain and expand its forces regardless of individual contracts. In return, the UIC received priority access to the company for operations related to the expanding Zerg crisis. The agreement included procurement privileges, access to restricted military technology and intelligence sharing that would previously have been unthinkable for a private organization.

The commander signed. Then the UIC handed him the target list. It was thick. Colonies. Research facilities. Missing ships. Suspected nests. Pirate stations dealing in biological material. Unexplained attacks. Coordinates where reconnaissance had simply stopped returning. The commander turned page after page. Kestral had answered one question. The Zerg outbreaks weren't random anymore. Unfortunately, it had replaced that question with a much worse one.

Who was giving the orders?

This is an update to the petition that was posted here:

https://www.change.org/p/demand-forensic-audit-and-transparency-for-kwantlen-student-association/u/35151837

---

When this petition was launched in October 2025, we called for something very specific: an independent forensic audit of the Kwantlen Student Association and greater accountability for the millions of dollars in mandatory fees collected from KPU students.

That demand is no longer hypothetical.

On March 13, 2026, the B.C. Minister of Finance formally launched an investigation into the Kwantlen Student Association under the Societies Act. The investigation is intended to determine whether there has been misuse of funds or other problematic conduct. While it is underway, a ministerial order prohibits the KSA from disposing of, dealing with, or diminishing the value of its assets except as authorized by the order, while permitting reasonable spending on proper operational expenses.

Then, in May, another major development occurred: Finance Minister Brenda Bailey confirmed that PricewaterhouseCoopers (PwC) would conduct the forensic audit, with PwC's forensic services team appointed to lead the investigation.

This represents a major step toward the accountability this petition has been calling for.

But an investigation is not a finding of wrongdoing. Allegations remain allegations unless and until the investigation establishes otherwise. What matters now is that these questions are finally being examined independently.

And the work is not finished.

This petition was never solely about discovering what happened at one student association. From the beginning, it also raised a larger structural question: what safeguards should exist when organizations collecting millions of dollars in compulsory student fees fail to govern themselves effectively?

Our petition therefore called not only for a forensic audit of the KSA, but also for stronger transparency and accountability requirements for student unions under B.C.'s legislative framework.

Following the announcement of the provincial investigation, the British Columbia Federation of Students (BCFS) issued a statement supporting the investigation but arguing that situations involving alleged financial mismanagement such as this are “extremely rare in British Columbia.” The BCFS emphasized the democratic processes and financial safeguards that exist within most student unions and argued that government intervention should remain rare.

We agree on an important point: students' unions perform valuable work, and it would be wrong to suggest that every students' union is poorly governed or financially mismanaged.

But it would be equally wrong to treat the problems at the KSA as though serious failures of student-union governance are without precedent.

The KSA is not an isolated Canadian example.

Ryerson University / Toronto Metropolitan University

At Toronto Metropolitan University — formerly Ryerson University — the student union has experienced serious financial controversies on more than one occasion.

In 2019, the university's independent student newspaper, The Eyeopener, obtained financial statements showing substantial expenditures on Ryerson Students' Union credit cards. A student-union director said the statements totalled approximately $250,000, with reported purchases including nightclubs, entertainment venues, Airbnb, clothing and other expenditures.

The controversy became serious enough that Ryerson requested that the student union undertake a forensic audit. The union subsequently retained PwC to conduct a financial review into allegations of financial mismanagement and later filed a police report. No systemic change occured.

Yet the problems did not end there.

In 2023, the renamed Toronto Metropolitan Students' Union filed a $900,000 lawsuit against five former executives from the 2021-22 academic year. The lawsuit alleged fraud, financial mismanagement, breach of fiduciary duty, breach of contract and civil conspiracy, claiming approximately $250,000 in financial losses. Among the allegations were nearly $75,000 withdrawn for Toronto Raptors tickets, payments to third-party vendors without board approval and numerous transactions allegedly lacking appropriate documentation or justification.

Those allegations have not been proven in court.

There is also a noteworthy connection to Kwantlen: current KSA Executive Director Timothii Ragavan is himself a Ryerson alumnus, having graduated from the university with a bachelor's degree in business/managerial economics. There is no suggestion that Ragavan had any involvement whatsoever in the Ryerson Students' Union controversies; the point is simply that the problems now being debated at Kwantlen are not unique to one institution or one student-union environment.

University of Regina

More recently, serious financial and governance concerns emerged at the University of Regina Students' Union (URSU).

After reviewing URSU's 2024 audited financial statements, the University of Regina raised significant concerns about the organization's financial position. The university said it attempted to work with URSU leadership and offered financial and governance assistance, but that recommended steps were not taken.

The university ultimately stated that it had “lost confidence in URSU's ability to serve and represent the University's students” and terminated the agreement under which it collected student-union fees on URSU's behalf, effective August 31, 2025.

Reporting on URSU's finances indicated that its 2024 financial statements showed a $1.25-million deficit, while separate legal disputes involved allegations that hundreds of thousands of dollars in student fees collected for campus organizations had been withheld.

The situation escalated further in September 2025, when the Regina Police Service confirmed that URSU was under investigation regarding the possible misappropriation of funds.

And this is not unprecedented in British Columbia We do not even need to leave British Columbia to find another example.

In 2017, the UBC Students' Union Okanagan (UBCSUO) dismissed its general manager and began an external accounting review after two consecutive annual deficits exceeding $300,000.

Reporting by The Ubyssey had previously uncovered evidence of financial mismanagement and potential conflicts of interest. The deficits were particularly significant because the union's own bylaws prohibited it from operating at a deficit, and the union had an annual budget of approximately $1.4 million.

These organizations are independent of one another. Their circumstances are different, the allegations involved are different, and it would be irresponsible to suggest that every students' union suffers from the same problems.

But when similar governance failures repeatedly emerge at independent student unions in different provinces, at different universities and under different leadership, we should at least be willing to ask whether the problem is entirely individual — or whether there is also a structural weakness in the system intended to prevent and address these failures.

That distinction matters.

The argument for stronger oversight does not depend on proving that most student unions are mismanaged. Regulation exists precisely for the circumstances in which ordinary internal controls fail. As they have, time and time again. 

Students' unions can administer millions of dollars collected through compulsory student fees. They are independent organizations with their own elected leadership, bylaws and internal accountability mechanisms. But when these lax and easily-manipulated mechanisms fail, the examples above demonstrate how serious a situation can become before an external institution, court, police service or provincial government becomes involved.

That is why this campaign has always been about more than the particular individuals currently running the KSA. The issues have been systemic for decades.

Independent financial oversight, accessible financial information, enforceable conflict-of-interest requirements, meaningful consequences for breaches of fiduciary obligations, and effective mechanisms for intervention when internal governance fails do not undermine well-run students' unions.

They protect them.

They protect students from misuse of compulsory fees, and they protect responsible students' unions from having public confidence in the entire sector damaged by organizations that fail to meet those standards.

The question is therefore not whether every students' union is mismanaged.

Clearly, they are not.

The question is whether the regulatory framework is strong enough when one is.

And that brings us back to Kwantlen.

The forensic audit we asked for is now underway. But the next questions are just as important:

1. What will PwC find?

2. Will the findings be made public?

3. What corrective action will follow if wrongdoing or governance failures are identified?

4. And, perhaps most importantly, will the Province use what it learns to consider reforms that could prevent another student association from reaching the point where extraordinary government intervention becomes necessary?

To Everyone Involved, Thank You.

To everyone who signed this petition, shared it, spoke publicly, contacted government officials, reported concerns, or simply refused to accept that compulsory student fees should be collected without meaningful accountability: thank you.

This investigation is an important milestone.

But accountability does not end when an audit begins.

Now we need transparency about what it finds... and meaningful reform from whatever lessons emerge.