I Own a Pair of Allbirds

I own a pair of Allbirds. They're fine. When I needed shoes again, I didn't go looking for another pair.

Today, Allbirds announced it's pivoting to "GPU-as-a-Service."

Two weeks ago they sold what was left of the shoe business for $39 million to the brand holding company that also owns Aerosoles and Ed Hardy. The remaining corporate shell took a $50 million cash injection, rebranded as NewBird AI, and the plan is to turn that into GPUs.

Stock popped 400% on the announcement. Not on revenue. Not on a single GPU spinning. On the press release.

That's the tell.

Wired nailed it in one line: "Startups used to make things; now they buy processors." Bitcoin miners pivoted first. A supersonic airliner company is now selling natural-gas turbines to data centers. Now a shoe brand. None of these companies particularly believe in AI. They've noticed that announcing AI pays better than running a real business. And right now, it does.

Here's why that matters if you're trying to actually use this stuff:

The gap between announcing AI and running AI in production is enormous, and it's widening every week. Uptime is hard. Rate limits are real and tightening. Model behavior shifts under your feet (an AMD senior director just filed a GitHub issue documenting Claude Code regressing across 234,760 tool calls). None of that shows up in a press release. Which is exactly why press releases keep getting funded over operations.

Your job isn't to join the theater. Your job is to find the parts of your business an agent can reliably handle today, measure it honestly, and let the sneaker companies fight over the rack space.

(Warby Parker, you're on notice.)

Start boring. Ship reliably. Let the pivots be somebody else's story.