Leadership Coachella Valley has been around for twenty-five years. It's an application-based program running September through June, moving through civic, healthcare, education, media, nonprofit, finance, and energy. It has a serious alumni network and a quiet reputation as the kind of program that actually changes how people think about this region. I was asked at the last minute to fill a slot on the May roster to talk about AI. I said yes. What follows is the condensed version — for the forty or so people who were in the room, and for the alumni who weren't.

Leadership Coachella Valley — May 15, 2026

I was introduced at 3:20 on a Friday afternoon. Last speaker of a full-day program. A room full of leaders across eleven sectors who had been sitting in chairs and taking field trips since morning and were, by any honest measure, thinking about happy hour. I also learned there had been a bus accident earlier that morning — nothing serious, but still the kind of thing that sits in a room.

I said: I know. I see you. Give me thirty minutes and I'll make it worth it.

Someone on the AV side asked if I needed the slides up. I said no slides. The room clapped. Not politely. Genuinely. Which told me everything I needed to know about how the previous seven hours had gone.

I also opened by asking: does anyone know how much a banana costs?

A few guesses. 39 cents. 49 cents (probably a Gelson's shopper.) 19 cents, which is Trader Joe's and correct. I said: somebody remind me at the end.

Then we got to work.

I always start the same way. The lay of the land. A show of hands.

Who's using ChatGPT? Much of the room. Forty people or so, hands up. Who's paying for it? About half the hands came down. Who's paying more than twenty dollars a month? Another handful down. Someone also asked: you can pay more than twenty dollars?

I filed that one away.

Who's using Claude? Three hands. Three. I looked at them and said: welcome, my fellow tech bros. Paying for it? Yes. More than twenty a month? Yes. If I had to guess, the four of us are probably spending two hundred dollars a month, mostly for Claude Code.

Here’s our first observation: when this question is asked in Los Angeles or San Francisco, the Claude hands go up across the entire room. Not three. Most everyone.

That is not a data point about Claude. It is a data point about where the Coachella Valley is relative to the places where these decisions are being made.

Moving on. Gemini — nearly the whole room uses it and most of them didn't know it. Gmail. Google Docs. Search. It's already inside the tools you're using every day. Hardly anyone is paying for it separately. That tracks.

Someone also asked about Copilot. Someone booed. I giggled.

Hold that image of the two rooms. We'll come back to it.

Incidentally, I had presented to their previous class in June 2025. Same program, different cohort. So when I walked in last Friday, I thought the most useful thing I could do — almost a year later, same room — was show them what a year actually looks like in the world of AI.

When I was last there in June 2025, ChatGPT had around 500 million weekly users. By May 2026, it's 900 million. Enterprise seats — businesses paying for it at scale — are up nine times in twelve months.

When I was last there in June 2025, Anthropic was valued at around sixty billion dollars. Today it's north of nine hundred billion. Revenue has gone from roughly one billion annualized to thirty billion. Claude Code — a product that didn't exist a year ago — is already at a two and a half billion dollar run rate. The CEO called the growth "crazy." His word. He ran out of adjectives. He's also running out of compute and energy — but that's for a different newsletter.

When I was last there in June 2025, big tech was spending three hundred and eighty billion dollars on AI infrastructure in the US. Today that number is seven hundred billion. Bank of America is projecting a trillion in 2027.

These companies are not betting on AI. They have already placed the bet. The concrete is being poured.

So I asked the room the only question that actually matters. Not what these companies are worth. Not what the next model can do.

What were you doing with AI when I was last here in June 2025 — and what are you doing now?

Organizations almost never fail at AI because they picked the wrong tool.

They fail because nobody made a real decision. Nobody addressed the fear. Nobody looked honestly at where they actually were. Nobody built the rules. Nobody owned what happened next.

Two days before this talk, Gartner published a global survey of 12,000 employees that confirmed what I've been watching for the better part of a year: the primary reason AI adoption fails inside organizations is not technical. It is cultural. Employees with a positive outlook toward AI are 3.4 times more likely to be highly productive. A software license doesn't create that outlook. Trust does. Transparent communication does. Leadership that has actually thought it through does.

What follows is the five-step framework I normally walk through in a four and a half hour workshop. I gave them the fifteen-minute version. This is the newsletter version. Let’s go.

It’s not complicated. Take what's useful. Skip it if you already have this down.

The through-line across all five: the technology is not the bottleneck. The culture is. The communication is. The decision that nobody quite made is.

Near the end I asked the question I always ask.

Does anyone in this room have an AI playbook at their organization — something written down, something you can point to?

One hand. Someone from KESQ. And before I could say anything, they said: your talk. June 2025. You said we needed a playbook. I took it back to my bosses and that's why we have one.

Well. This stuff actually works. Who knew.

(I knew. I did not always believe it.)

Here's the thing. I have been doing this for two years. Thirty workshops last year. Chamber events. Council presentations. Entrepreneurial classes. The slides were professional but nothing of note follows. The feedback loop in this work is long and slow and often silent. You give the talk, you hand out the one-pager, you drive home down 111. Most of the time you do not find out what happened next.

And most of the time, if I'm being honest, I'm not sure I moved the needle the way I thought I would.

I spend a fair amount of time in SF/LA. On any given weeknight there are two, three, four AI meetups, hackathons, dinners, and salons happening simultaneously. Weekends too. The density of conversation, experimentation, and collision is constant. You come back to the Coachella Valley and there is one conference a year. The valley is falling behind. I don't say that to be harsh. I say it because pretending otherwise doesn't help anyone in the valley.

Which is part of why 2026 looks different for me. Fewer workshops. More building. Because at some point you have to stop explaining what these tools can do and start showing it. mirage. The AI lobbying tracker. The AICV intelligence network. These exist because talk without proof is just a slide deck. The tools are good enough now that the most honest thing I can do is build, show, and tell.

And here’s the thing about that one hand going up at 4pm on a Friday.

Think about what that actually means. Someone sat in a room in June 2025, heard the framework, went back to their organization, and made the case. Leadership said yes. They built the document. A local news station — one that is simultaneously covering AI, being disrupted by AI, and competing in an industry being turned upside down by it — now has a written policy for how they handle it. Steps one through four of the AI Playbook. One person. One conversation.

That is what one conversation can do. Imagine what twelve of them a year could do. Or twenty-four. Or one a week.

Someone remembered at the end — what's up with the banana? — so I told them.

Last Monday, the developers behind a large data center project trying to get fast-tracked in Coachella held a community meeting. The kind of meeting developers hold when they need to demonstrate they listened. City staff was there. Legal was there. And many residents too.

A Coachella Valley resident wanted to make a point about who these outside developers are. Where they come from. What they actually know about the people who live in the neighborhoods where they want to build.

She asked the CEO: do you have any idea how much a banana costs?

He said: sure. Ninety-nine cents.

I was watching the live stream. Chaos, laughter, shouting, someone calling out 'you're out of touch’. The CEO had answered confidently and precisely wrong.

The AI economy is arriving in this valley whether we organize around it or not. The question is whether we show up to that conversation as participants or as a location. Whether the people making decisions about our infrastructure, our water, our power grid, our workforce — whether or not they know what a banana costs.

And whether we decide, collectively, that this is the moment to actually get serious. Not conference-serious. Not task-force-serious. Actually serious. The kind of serious where we educate our workforce, equip our students, and make sure the people who live here are the ones who benefit from what gets built here — not just the ones who absorb the costs of it.

We are currently moving at the speed of one conference a year. The world is moving at the speed of $700 billion in annual capital expenditure.

I don't think those two speeds are compatible.

Sat Singh builds with AI, experiments with AI and runs AI workshops in Rancho Mirage. Just one person raising their hand might be enough to keep going.

If this resonates, pass it along to someone in the valley who'd benefit. This is a community project — it grows the same way communities do, one conversation at a time.