Smartbird, the recently rebranded AI infrastructure company formerly known as Allbirds, has tapped former AWS executive Nadia Carlsten as its new CEO. The company, which made headlines in April when the direct-to-consumer shoe retailer announced a dramatic pivot to artificial intelligence, has officially closed its footwear business as of this week and is now operating as a pure-play AI infrastructure provider. Carlsten, who holds an engineering PhD and most recently led European compute company DCAI, started the role from Amsterdam and is now focused on assembling an entirely new leadership team.
The transition from shoe seller to AI company followed a familiar pattern for troubled public companies chasing market momentum. Allbirds sold its shoe business for $43 million and raised an additional $100 million from the stock market, using the war chest to rebrand as Smartbird and chase the seemingly insatiable demand for AI compute. But unlike neoclouds that arbitrage GPU pricing or inference token costs, Smartbird is positioning itself as a provider of carefully managed deployments for customers who need direct control over the physical servers running their models. According to Carlsten, the ideal clients are those prioritizing data sovereignty and customized infrastructure over the elastic scalability of major public cloud platforms.
Carlsten identified pharmaceutical firms, energy companies, financial institutions, and public sector organizations as the primary targets, drawing on her experience at DCAI working with clients like Novo Nordisk on bespoke AI deployments. She acknowledged the market is still nascent, noting that many enterprises are only in the pilot phase with AI tools, and declined to estimate Smartbird's total addressable market. "We're going to be recruiting a brand-new team for the AI business, and we're going to be getting an office," Carlsten said, adding that her immediate priority is hiring a leader to head up infrastructure operations.
The challenge ahead is substantial: Smartbird is essentially a startup with a single founder and a sizable seed round, competing not against hyperscalers or neoclouds but against internal corporate IT teams building their own AI infrastructure. With no employees beyond its CEO, no office, and an unproven go-to-market strategy, the company will need to move quickly to establish credibility in a crowded and capital-intensive market where customers have plenty of alternatives for their compute needs.