When Allbirds announced it was pivoting to AI back in April, I laughed. It felt like a plotline from Silicon Valley — the show, not the place. A direct-to-consumer shoe company that helped define the whole Patagonia-vest-and-lame-kicks aesthetic suddenly deciding it was an AI play? That’s the kind of move that usually ends up as a punchline.
Except it worked. At least financially. Allbirds sold its shoe business for $43 million, raised another $100 million from the stock market, rebranded as Smartbird, and now it’s got a CEO with a real resume and a real challenge.
Nadia Carlsten started yesterday. She’s a former AWS exec with an engineering PhD, most recently running the European compute company DCAI. She’s based in Amsterdam, she has no employees, no office, and a very large pile of cash. “We’re going to be recruiting a brand new team for the AI business, and we’re going to be getting an office,” she told TechCrunch. Her first task: rounding up a leadership team, starting with someone to lead infrastructure operations.
So it’s a startup with a sole founder and a $100 million seed round. That’s an unusual combination, and it’s going to be interesting to watch.
Smartbird’s pitch is AI infrastructure — specifically, managed compute for companies that want direct control over the servers running their models. That’s a different lane from the hyperscalers (AWS, Azure, GCP) and the neoclouds (CoreWeave, Lambda, etc.) that are all about arbitraging chip prices and scaling as fast as possible. Carlsten is aiming at customers who need data sovereignty, either for political reasons or because their business model demands it. Think pharma companies like Novo Nordisk, energy firms, financial institutions, public sector orgs. These are customers who are piloting AI but aren’t ready to hand their data over to a public cloud.
Carlsten couldn’t estimate the size of that market yet, which is honest but also a little worrying. She argues it’s nascent because many companies are still just piloting AI tools. That’s fair, but it also means Smartbird is betting on a market that might take years to mature.
She says Smartbird isn’t competing with hyperscalers or neoclouds, but with internal company projects — the “should we build this ourselves?” question. That’s a real distinction, but there are already established players in this space. Hewlett Packard offers single-tenant managed AI compute. Equinix does too. So Smartbird isn’t exactly breaking new ground.
What’s less clear is the growth potential. Carlsten expects to have compute clusters deployed for several customers by the end of the year. Compare that to a startup like General Compute, which announced a $300 billion chip order when it came out of stealth last month. Different ambitions, different scales. Carlsten says her customers need hundreds to thousands of chips, not tens of thousands. “It’s not about large scales and huge numbers of GPUs,” she said. “It’s more about agility of these clusters, and more about having control of the infrastructure stack.”
That’s a defensible niche, but it’s also a smaller one. And Smartbird won’t compete on price either, because cloud services optimize chip usage 24/7 to offer the cheapest compute. Carlsten’s bet is that companies with specialized workflows will be more efficient on their own servers. Maybe. But efficiency isn’t the same as cost.
One thing that quietly disappeared in the pivot: Allbirds’ public benefit corporation status. That was supposed to enshrine the company’s sustainability commitments, the whole reason the shoes existed in the first place. PBC charters are often used by companies to signal non-financial priorities — OpenAI is a PBC with a focus on AI safety. But this change suggests those charters are about as ironclad as a wet paper bag when the board decides to chase a new trend.
Carlsten insists this wasn’t a cynical move. “It wasn’t, ‘Let’s just do AI, because it’s AI, and it’s hot,'” she said. She’s getting paid $700,000 a year plus about $9 million in stock to make that case. The board has made a long-term commitment to her strategy, she says.
I want to believe her. The AI infrastructure market is real, and there’s room for players that aren’t just racing to build the biggest cluster. But Smartbird is starting from zero — zero employees, zero customers, zero track record. The $100 million gives them runway, but it doesn’t give them credibility. That’s what Carlsten has to build, one managed cluster at a time.
“There are some companies out there chasing AI,” she said. “But at the end of the day, what matters is, is there actual weight behind the chasing?”
We’re about to find out.
Comments (0)
Login Log in to comment.
Be the first to comment!