The Allbirds-to-AI Pivot
How a sustainable shoe company accidentally proved the Compute Heat Rate thesis
There’s a moment in every market cycle when the universe hands you a case study so hilariously perfect you couldn’t have scripted it. Today is that moment.
Allbirds, the sustainable footwear company, announced this morning that it’s selling the brand for $39 million and pivoting to become an AI compute infrastructure company. The stock instantly surged 800% in one day.
They’re calling it NewBird AI. Lol.
….
Allbirds went public in 2021. It was a certified B Corporation. It enshrined environmental conservation as a public benefit in its Delaware certificate of incorporation. The pitch was beautiful: shoes made from eucalyptus fibers, sugarcane-based soles, a carbon footprint label on every box.
Then revenue cratered. By Q3 2025, net revenue had fallen 23% to $33 million with net losses of $20 million. The stock, once above $28, was trading around $2.
So the board sold the brand, kept the public company shell, raised $50 million in convertible notes, and announced a pivot to “GPU-as-a-Service.” They’re also asking shareholders to vote to remove the environmental conservation language from the corporate charter. Yikes!
The CHR Connection
For those who haven’t been closely following my recent work, the Compute Heat Rate1 is a framework I’ve developed for understanding how AI compute demand structurally reprices wholesale electricity markets. The core insight is that data centers fundamentally change the supply-demand dynamics of every power market they enter. Check this site out for more context on this.
The typical players here are hyperscalers: Microsoft, Google, Amazon, Meta among others. Companies with the balance sheets and engineering talent to build data centers at scale.
I’ve kind of been waiting for other market signals to show up: why not convert an entire aluminum smelting operation into an AI data center if the returns on using electricity for AI are so much better? Besides the obvious reasons…
And here we are! The Allbirds pivot signals the emergence of another wave of demand from entities that essentially have no business being in this space. The perceived opportunity in AI compute is now so large it’s attracting capital from the most improbable sources imaginable.
What CHR Tells Us About the NewBirds of the World
I don’t think NewBird AI will become a meaningful compute provider, but we’ll see. A $50 million raise doesn’t buy much when a single NVIDIA DGX SuperPOD costs north of $60 million.
But that’s not the point. The point is what their existence tells us about the demand curve.
CHR models compute demand as a function of three variables: workload intensity (power per unit of useful compute), capacity growth (new MW coming online), and market penetration (compute’s share of regional load). The “NewBird effect,” let’s coin that, is a leading indicator for variable two. When non-endemic capital starts flowing into GPU acquisition, it signals that the market believes demand will exceed supply long enough to generate returns even for the most random operators.
Final Thoughts
I’ve spent a good amount of time building the CHR framework to explain how AI compute demand would structurally transform energy markets, but I did not think a shoe company would become one of my best case studies.
But here we are. NewBird AI, a company seemingly with no technology team, no previous data center expertise, and a corporate charter that until this morning was dedicated to environmental conservation, just demonstrated the most important thesis in energy markets today: the demand for AI compute is so overwhelming it’s pulling in capital from every corner of the economy.
That demand needs power. That power needs to come from somewhere. And the structural repricing of wholesale electricity markets that CHR predicts will likely follow.
Hans Royal is the originator of the Compute Heat Rate™ (CHR) framework. All views are his own and do not represent those of any employer or affiliated organization.
Royal, Hans, The Compute Heat Rate: Quantifying AI-Driven Electricity Price Tolerance
and Its Implications for Wholesale Market Repricing (February 28, 2026).
Available at SSRN: http://dx.doi.org/10.2139/ssrn.6322318