Allbirds, the once-celebrated eco-friendly shoe brand popular among tech CEOs and celebrities, is undergoing a dramatic pivot, transitioning its business entirely to artificial intelligence infrastructure. The San Francisco-based company announced a definitive agreement with an unnamed institutional investor for $50 million in financing, specifically earmarked to facilitate this shift. With this transformation, Allbirds will also adopt a new identity: NewBird AI. The proceeds from the financing are intended for the acquisition of graphics processing units (GPUs), essential components for high-performance computing, with the transaction expected to finalize in the second quarter of this year.

According to the company, this strategic redirection is a response to the "unprecedented structural demand for specialized, high-performance compute that the market is struggling to meet," a gap that NewBird AI aims to help close. However, this abrupt change in business model has raised eyebrows among industry experts. Bill Kleyman, an AI infrastructure expert, described the move as a "strange pivot," noting that a transition from footwear to AI infrastructure lacks "a very natural adjacency" within the industry.

The specific nature of Allbirds' new venture involves reinventing itself as a "GPU-as-a-service" provider, a business model centered on renting out computing power to AI companies. This entails offering access to a vast number of graphics processors or other specialized AI computer chips, typically designed by major players like Nvidia or AMD, which operate within large data centers managed by cloud computing giants such as Amazon or Oracle. Kleyman, CEO and co-founder of Apolo, emphasized the inherent complexities of running physical AI infrastructure, highlighting the need for "access to GPUs in a constrained market, long-term power agreements, advanced cooling strategies, and a credible operating model."

This significant announcement follows a period of decline for Allbirds. Just over two weeks prior, the company sold its intellectual property and certain other assets and liabilities to American Exchange Group, a prominent accessories design, licensing, and manufacturing firm, for $39 million. This marks a stark contrast to Allbirds' peak valuation of $4 billion in late 2021. The company had also previously stated it would not be releasing its quarterly earnings report scheduled for March 31.

The brand's original mission, founded in 2015 by former professional soccer player Tim Brown and renewable resources expert Joey Zwillinger, was to create footwear from natural materials rather than synthetics, leading to the launch of its iconic wool runner shoe a year later. However, Allbirds faced challenges common to many dot.com brands, including overexpansion through physical stores, and a subsequent decline in consumer interest. In February, the company shuttered most of its remaining stores to prioritize e-commerce, retail partnerships, and international distribution, though it continues to operate two outlet stores in the U.S. and two full-price stores in London.

News of the AI pivot sent Allbirds' shares soaring by more than 600% on Wednesday, with the stock hovering near $18 in late afternoon trading, a dramatic increase from its $3 valuation just days prior, though still far from its previous high of $520 per share. Kleyman viewed this stock market surge as "more like initial excitement and speculative momentum tied to anything AI rather than validation of execution." He also cautioned that $50 million is a modest sum for entering an infrastructure-heavy market and suggested that while some shifts to AI are strategic, others, like Allbirds', appear "reactive" and perhaps "a bit desperate," offering a "compelling narrative reset" for a struggling underlying business.

Jim Piazza, former head of computing infrastructure at Meta and now chief AI officer at IT services firm Ensono, reiterated the sentiment, acknowledging the real demand for AI computing power but also the pervasive "hype." Piazza stressed that building a robust AI infrastructure business demands "deep capital, technical expertise and disciplined execution," a challenge he considers "crazy hard for tech-savvy companies" and potentially "an impossible challenge" for those outside the core tech sector.