Anthropic just dropped the mic without even saying anything.

Don't look now, but "somehow," it leaked that investor interest in Anthropic has reached "feverish" new levels, and the company is turning down investors who are "clamoring" to invest a "fresh $50 billion" at a valuation in the "$850 billion to $900 billion range."

Same company that brought you: All code will be written by AI by February, 2026.

OK, I mean, it's May and all, but you know what? I believe them this time. Fool me twice. And my first and only reaction to this new news, and likely yours too, is that we've now entered the final lap of the Great AI Race.

So get your AI projects and AI agent teams and AI code farms in order now. Because we are nearing the terminal point of the race track, and I see a Wile-E-Coyote-style brick wall that's painted to look like a tunnel.

Here's when the AI race ends, why, and what to do when it happens.

Anthropic Takes The AI Driver's Seat

You just know Sam Altman is sitting uncomfortably in the back seat now, shouting at Dario to slow down, he's braking too hard, left, left, LEFT, you missed it, etc.

But seriously, Anthropic is having another moment.

It started just over a month ago when a study showed "Claude gaining paid subscribers in record numbers" on the consumer side, but also, "Claude Code and Claude Cowork — developer and productivity tools released in January — have been drivers of subscriptions."

And I'm hearing the same in my own circles. I've always used Claude Code, only hopping to other models when I needed a "second opinion" or didn't want to leave whatever IDE I was currently poking around in.

Yo, I still got Atom loaded up.

So, yeah, I'm not a great user persona, certainly not hardcore. But as early as mid-2025, it kinda seemed obvious hearing from those around me that Anthropic was in the sweet spot to satisfy both the hardcore coders and the newly minted AI billionaire wanna-bes.

Say what you want about enterprise adoption, but AI still isn't ready for prime-time. It's getting there, and so now Anthropic can ride with the AI consultant crowd as they warm up their big hit: "Oh, I see what your problem is. It's the vendor you just fired."

Real consultants know.

$1 Trillion Is a Huge Psychological Milestone

Here's the trap that's about to be sprung on all of us. Go get your tin foil hat.

I don't know about you, but I keep looking at the sentence about Anthropic being poised to raise another "$50 billion" and my lizard brain keeps mistakenly reading that as only "$50 million."

And while I know $50 million is couch cushion change these days, $50 billion still seems unreal as a "round" of fundraising — even if it is the final round before the IPO.

I want the IPOs to happen. I think when one of them AI boys like Altman or Dario utters the term "IPO" in an official public setting, it's going to unfreeze a lot of innovation spending that's been iced or hoovered into AI since 2021.

But while none of these privately-held AI companies has "officially" hit the $1 trillion valuation mark, it's coming. And if I'm Anthropic, I wait, what, a month? And then we parlay that "feverish clamoring" to raise a little more to hit the magic $1 trillion valuation number.

And then brains melt. And then the Great AI Race comes to a screeching halt. And we win.

Here's why.

AI Investment Capital Is Not Infinite

Over the last few months, I've been talking about the massive amounts of spending that companies like Amazon, Microsoft, and Meta have committed to AI for 2026, and my take on that is that they can't all be winners.

The same is true for the AI companies receiving said investments. One company, exactly one, is going to turn a couple hundred billion dollars investment into $1 trillion. The runner up might eke out survival and revival in the 2030s, and the rest of them will pivot.

Proof? Microsoft used to have a mobile OS, and Blackberry led the sector forever. Until it didn't.

But a more recent lesson in rugs-being-pulled can be taken from the very recent demise of LIV Golf, the Saudi-PIF-backed upstart competition to the thousand-year-old-or-whatever PGA Tour. The Saudis literally threw $5-or-$6 billion at LIV Golf, until they didn't.

Or, more pointedly, "until last Thursday."

"Wait. You mean we haven't taken over the world yet? Well, eff this. We're outty. See you at Putt-Putt."

In other words, even the largest, deepest, and most spend-crazy investment fund in the world has its limits. And when those limits are reached, there's no saving the venture, because literally no one else can step in and replace that lifeline.

AI wasn't nickeled-and-dimed into existence. Companies went from zero to one to nine-figure seed rounds overnight. When the AI race ends, the "follow-on" money is either going to the winner, to the runner-up, or to something else entirely.

"Hey guys! We're ready for our $500 million Series A now! Who's in?"

Maybe The Big AI Evolution Already Happened

OK. Now put that tin foil hat on.

A colleague recently brought up to me the fact that these big Magnificent 7 companies are making investments into the next "step jump," as AI evolves in big leaps and then incremental steps.

Agreed. But that got me thinking. What if the final big step jump was a couple years ago, and they all know this and nobody is saying anything?

I don't know if you remember, but at the end of 2024, everyone was talking about the "AI wall" that we had hit. There was a lot of speculation about what the AI wall was, why we hit it, and what would happen next.

I've been building and working with AI for 15 years, and so I knew what that wall was. I tried to explain it simply, and for a simple example I used my writing and RAG models, which were hot at the time:

"I've written around 1,000 posts over the last five years. An AI platform could ingest another 10,000, or 100,000, and it wouldn't make much of a difference in the return to the end user."

What if we hit that wall in 2024, the wall that says, yeah, AI will keep getting better, but there's only so much productivity we can generate before we start overproducing what today's society of information consumers needs? What if the rest of AI evolution happens only incrementally because we don't actually need anything more than what it can do today? The demand for AI has a hard limit, not the supply.

How come no one thinks like this? Well, some do. Real techies know.

Again, look at mobile phones. At some point, the quest for the next evolutionary step in a technology gets punished instead of rewarded, so incremental improvement becomes all that the market will pay for. Apple and Tim Cook figured this out, and if I were to guess, I think they applied that same theory to AI.

So kids, get those AI apps out to production and grab your market share now. Maybe I'm wrong about $1 trillion in valuation being the end of the Great AI Race, but I do know this. It will end, and it'll end suddenly, and second place doesn't get to go to Disneyland.

Now would be a good time to join my email list, and get a heads up whenever I say something that will totally make sense 16 months from now.