What the Allbirds CEO Knows About AI on the Product Roadmap

The Pivot That Should Make Every PM Nervous

In April of 2026, a company that used to sell wool sneakers announced it was going to start renting out graphics processing units to artificial intelligence developers. It is one of the more extreme examples of what happens when the pressure to put AI on the product roadmap overrides the discipline to ask why.

The company is Allbirds. Picture the shoes a venture capitalist wore to brunch in 2019. Wool. Sustainable. B Corp certified. The kind of thing you bought because it said something about you. They went public in 2021 at a valuation of around four billion dollars. By late 2025 the company was warning shareholders it might not survive.

In March of 2026, they sold the entire shoe business (brand, inventory, intellectual property, all of it) for thirty-nine million dollars. About one percent of what the company was worth at its peak.

What was left was a publicly traded shell. A ticker symbol on the Nasdaq, a small amount of cash, and a CEO named Joe Vernachio.

Two weeks later, Vernachio announced what was left would become NewBird AI. A company that buys graphics processing units and rents them to AI developers. Fifty million dollars in financing. No GPU experience. No data center. No customers. No utility relationships. The press release used the phrase “fully integrated AI-native cloud solutions provider.”

The stock went up five hundred and eighty percent in a day.

The next day it fell twenty-nine percent. By then the headlines had been written, the LinkedIn posts had been posted, and somewhere a CEO was forwarding an article to a product manager with two question marks and no other context.

You read this and the obvious thing to think is that’s ridiculous. A shoe company is going to lease GPUs. Sure. You laugh and I laugh, the way the internet laughed. You feel briefly superior. You feel a little better about your own roadmap.

That’s the comfortable thing to do. I’m not interested in comfortable.

Everyone Acted Rationally. That’s the Problem.

Joe Vernachio is not a fool.

Look at his situation in March. Cash burning down. Stores closing. Revenue dropping every quarter. Nasdaq compliance pressure. A board that wants a plan. He has two doors. Door one is an orderly wind-down, where he becomes the CEO who closed Allbirds. Door two is a dramatic pivot, where he becomes the CEO who tried something bold.

Door one ends his career. Door two extends it. Door two also happens to be the one the market is paying for.

Now look at the investor who put up the fifty million. They aren’t betting on GPU leasing. They’re buying an option on a publicly traded shell, in an environment where putting “AI” in a company name produces reliable, immediate stock gains. If the narrative holds long enough to raise more money at inflated prices, the return is enormous. If it doesn’t, the downside is capped.

And look at the market itself. An economist named Owen Lamont published research in April showing that since 2023, thirty-three companies on major U.S. exchanges have added “AI” or some AI-flavored term to their names. Almost all of them got a stock gain on the announcement. The same thing happened in the late 1990s with “.com.” Some were real. Most were not. The market did not sort that out in time.

So here is the picture. A CEO with rational personal incentives. An investor with rational financial incentives. A market with a measurable, repeatable pattern of paying premiums for the letters A and I. Thirty-three companies before Allbirds noticed the pattern and acted on it. Allbirds is the thirty-fourth.

This was not a moment of madness. This was the system working exactly as designed.

There is one detail in the filings I keep coming back to. As part of the pivot, the company is asking shareholders to vote in May to remove all references to operating as an “environmental conservation public benefit” corporation from its charter. The B Corp language. The thing the customers were buying. The reason the shoes cost what they cost.

They are not just changing what they sell. They are voting to delete who they were.

That detail isn’t making the LinkedIn posts. But that detail is the entire story.

The Same Machine Is Running in Your Building

Now let’s talk about you.

Your CEO is not a fool either. Your CEO is responding to the same incentives Vernachio responded to. The board wants an AI story. The investors want an AI story. The competitor down the street has one, or at least a press release that sounds like one. The easiest, fastest, most legible move is to put AI on the product roadmap and ship something with that label on it before the end of the quarter.

You have already had this meeting. The one where someone said “we need an AI feature.” You know how it ended. The feature was greenlit. The discovery cycle was not. You said something in the room and were politely steamrolled, and an engineer started building.

The reason you laughed at Allbirds is that the pivot was loud and stupid and obvious. The reason you should pay attention is that the same machinery is operating, more quietly, in your building. The pressure to put AI on the product roadmap is not coming from a villain. It is coming from a board, an executive team, and a CEO all responding rationally to a market paying for the label. You are inside the same system. And that pressure is precisely why product discovery for AI features keeps getting skipped.

That is what AI fatigue is. Not that you’re tired of AI. That you can feel the gravity pulling everyone in your building toward the same kind of decision Vernachio made, and you don’t yet have the words to slow it down.

What AI on the Product Roadmap Requires (Realistically)

So let’s think about this differently.

The job of the product manager in this environment is not to resist AI. That argument loses, and it should lose. The technology is real. The opportunity is real. There is a company called CoreWeave that started out renting computers to cryptocurrency miners and is now worth more than twenty billion dollars renting computers to AI companies. Some pivots work.

Your job is to be the person in the building whose incentive is to ask the questions everyone else is incentivized to skip.

• Who is the customer.
• What problem are we solving for them.
• What outcome are they trying to reach.
• What evidence do we have that this is real.
• What do we believe that we cannot prove yet.

These questions are not anti-AI. They are the only thing that makes an AI investment worth the money. A leader who has good answers should be supported, fast and visibly. A leader who doesn’t have answers does not have an AI problem. They have a product management problem, and you are the answer to it.

This is how we lead quietly. We demonstrate that product discovery for AI features is not the slow lane. It is the only lane that doesn’t end in a thirty-nine-million-dollar fire sale.

What to Do Tomorrow Morning

If you are a product manager and you read this far, here is what you do tomorrow morning.

Forward this piece to your leadership team. Not as a complaint. As a starting point.

Ask one question with it.

What customer problem are we solving with the AI work on our roadmap, and what evidence do we have that solving it matters? That question is the foundation of product discovery for AI features, and it is the most important question you can bring into that room.

If your leaders have answers, your job just got clearer. You stop being a passive recipient of strategy and become the person who turns it into outcomes. That is the most valuable thing you can do this year.

If your leaders don’t have answers, you’ve also learned something useful. You are the most important person on the team for helping them define why that strategy matters.

If you are a leader and you’re reading this because someone on your product team forwarded it to you, the first move is not to walk down the hall and quiz your PM. The first move is in the mirror.

Ask yourself why the business needs to pivot toward AI. Not the answer you’d give an investor. The answer specific enough that a thoughtful person could disagree with it. If you can’t get past “the market is moving” or “our competitors are doing it,” that’s not a strategy. That’s the same incentive gradient that produced NewBird AI.

Ask yourself whether you have told your team what outcomes you want. Not the feature. Not the headline. The outcome. The customer behavior or business result that would tell you the bet paid off. If you haven’t said it out loud, in language clear enough that they could chase it without you in the room, they aren’t being passive. They’re being careful.

And ask yourself whether you are giving them the space to do the work that matters. Figuring out who you’re serving. What those people are trying to accomplish. Which problems are worth solving. If the answer is “we don’t have time for that, we need to ship something with AI on it by Q3,” you already know what kind of company you are about to become.

Joe Vernachio knows something about putting AI on the product roadmap that you should know too. He knows the market will pay for the label. He knows the incentives all push the same way. He knows the pressure feels enormous from inside the room.

What he doesn’t know is whether anyone will still be there to buy what he’s selling six months from now.

You can know that. That’s the product manager’s job.

• The pressure to put AI on the product roadmap isn’t going away. What changes is whether you have the frameworks to redirect it. Our AI Product Management course is live online and instructor-led: built for PMs and product leaders who want to move from reactive to deliberate, across discovery, prioritisation, strategy, and product ops.
• Are you the person in your building asking the hard questions about AI on the product roadmap? Tell us how you’re doing it on LinkedIn and tag @Productside. We want to know what’s working (and what’s getting steamrolled).