When Allbirds announced it was selling its shoe business and pivoting into AI infrastructure back in April 2026, the internet did what the internet does best: it laughed. A company famous for flimsy footwear made from wool and sugarcane was suddenly going to build hyperscale compute clusters. It read like a sketch from Silicon Valley that had somehow escaped the writers' room and landed on a Nasdaq listing.
But here's what most people missed in the mockery: the company actually raised $100 million from the stock market after the pivot, sold its shoe business for $43 million, and brought in Nadia Carlsten as CEO — a former AWS executive with an engineering PhD who'd most recently led the European compute company DCAI. This isn't a meme stock play anymore. It's a real business with real money and a real leader who knows the infrastructure stack inside out.
Smartbird, as the rebranded company is now called, isn't trying to be another neocloud. It's not going to arbitrage chip prices against GPU time the way General Compute or other inference-focused startups are. Carlsten's vision is narrower, more deliberate, and honestly, probably more defensible than the race to the bottom on compute pricing.
The Data Sovereignty Thesis
The core of Smartbird's strategy is data sovereignty. Not the buzzword version that gets tossed around in boardrooms — the actual, legal, compliance-driven need that pharmaceutical companies, energy firms, financial institutions, and public sector organizations have for keeping their data on infrastructure they control.
Think about it. A company like Novo Nordisk — one of Carlsten's former clients at DCAI — doesn't want its drug discovery models running on a multi-tenant public cloud where the data residency boundaries get fuzzy. They need single-tenant compute, predictable performance, and absolute control over who touches their intellectual property. The same goes for energy companies managing grid optimization models, or banks running fraud detection systems that can't leave their jurisdiction.
Carlsten told TechCrunch she couldn't estimate the size of this market yet, and that's fair — many enterprises are still piloting AI tools rather than running production workloads at scale. But the demand is real, and it's growing. The question isn't whether this market exists; it's whether Smartbird can build the operational muscle to serve it before someone else does.
Not Competing With Hyperscalers
Here's where Carlsten's thinking gets interesting. She doesn't see Smartbird as competing with AWS, Google Cloud, or Microsoft Azure. Those companies are playing a different game — one centered on massive scale, multi-tenancy, and optimizing chip usage around the clock to offer the cheapest possible compute.
Smartbird's customers don't want cheap. They want control. They're looking at clusters in the range of hundreds to a few thousand chips — enough to run serious AI workloads, but nowhere near the planet-scale deployments that define the hyperscaler story. Carlsten put it bluntly: "It's not about large scales and huge numbers of GPUs; they're more about agility of these clusters, and more about having control of the infrastructure stack."
This positioning also means Smartbird isn't really going head-to-head with neocloud startups either. Those companies are chasing volume and price optimization. Smartbird is chasing specificity — serving organizations with specialized workflows that can actually work more efficiently on their own dedicated servers than on a shared public cloud.
The real competition, Carlsten argues, is internal IT projects. Companies that are building their own AI infrastructure in-house because they don't trust anyone else with their data. Smartbird wants to be the alternative to that self-built approach — a managed service that gives you the sovereignty of owning your stack without the operational headache.
The Competitive Landscape
Smartbird isn't entering a vacuum. Hewlett Packard already offers a single-tenant managed AI compute service. Equinix, the data center giant, has been quietly building out infrastructure services for enterprises that need more control than a public cloud offers. These are established players with deep relationships and proven track records.
But there's a meaningful difference. HPE and Equinix are infrastructure providers that happen to offer managed services. Smartbird is an AI-native company from the ground up — or at least, it's trying to be. The pivot from shoes to servers is jarring, sure, but Carlsten's background at AWS and DCAI gives her credibility in a way that a pure infrastructure company pivoting into AI might not have.
The challenge is execution. Carlsten admitted she's starting from zero on the people side: "We're going to be recruiting a brand-new team for the AI business, and we're going to be getting an office." Her first task is rounding up a leadership team — looking for someone to lead infrastructure operations, for example. It's essentially a startup with one founder and a very large seed round.
That's both a vulnerability and an opportunity. Vulnerability, because building a team from scratch takes time, and the AI infrastructure market isn't waiting. Opportunity, because there's no legacy culture to overcome, no shoe-business DNA holding the company back. Smartbird can build its operational identity around AI from day one.
The Numbers Behind the Pivot
Let's talk about what this actually costs. Carlsten will be paid a $700,000 annual salary and was awarded stock worth about $9 million to take the job. The company raised $100 million from the stock market after the pivot announcement. That's a serious war chest for what is, functionally, a company with no employees and no customers yet.
Compare that to General Compute, the inference cloud startup that came out of stealth last month with a $300 billion chip order. Smartbird isn't playing that game. Carlsten doesn't need massive chip commitments because her target customers need hundreds to thousands of chips, not the tens of thousands that define the hyperscaler and neocloud deployments.
This is a deliberate choice, and it reflects Carlsten's understanding of where the actual demand is. The $300 billion chip orders make for great headlines, but they're betting on a future where every company needs planet-scale inference. Smartbird is betting on the present, where many organizations need competent, controlled, single-tenant compute right now.
What Happens Next
Carlsten expects to have compute clusters deployed for several customers by the end of the year. That's an ambitious timeline for a company that literally has no employees, but she's clearly confident in the demand side. The question is whether she can build the operational capability fast enough to deliver.
There's also the matter of Smartbird's public benefit corporation status, which Allbirds had held to enshrine sustainability commitments. That status was dropped during the pivot — a move that suggests PBC charters aren't as ironclad as companies might hope. For some observers, this raises questions about whether Smartbird's commitments to its customers will hold up the same way.
Carlsten pushed back on the idea that Smartbird is just chasing AI for the hype. "It wasn't, 'Let's just do AI, because it's AI, and it's hot,'" she said. "It was really about, do we have a chance to build a business over time that is going to find this niche in the market and be able to grow over time?" She added, "There are some companies out there chasing AI, but at the end of the day, what matters is, is there actual weight behind the chasing?"
The board's commitment to her strategy, combined with the $100 million raise and Carlsten's own $9 million in stock compensation, suggests there is weight behind the chasing. Whether that translates into a viable business remains to be seen.
Why This Matters Beyond Smartbird
The Smartbird story is interesting not just because it's bizarre — a shoe company becoming an AI infrastructure provider — but because it highlights a real tension in the AI market. Everyone is focused on scale. Everyone is racing to build bigger clusters, secure more chips, and offer cheaper compute. But there's a segment of the market that doesn't want cheap or big. They want control.
Data sovereignty isn't going away. If anything, it's getting stricter as governments around the world introduce more regulations around data residency and AI governance. Smartbird is positioning itself at the intersection of that regulatory pressure and the practical need for dedicated AI infrastructure.
Whether Carlsten can build this into a sustainable business is the open question. She has the experience, the funding, and a clear thesis. What she doesn't have is a team, customers, or any operational history in this space. The next six to twelve months will tell us whether Smartbird is a genuine play on data sovereignty or just another case of a company that got lucky with a pivot and ran out of ideas.
