Every few weeks, some new number from the AI infrastructure boom makes me do a double-take. This week it was Nscale — a company most people in India have never heard of — filing paperwork to go public on the New York Stock Exchange under the ticker NSCL. The London-based firm rents out Nvidia GPUs for training and running AI models, and its customer list reads like a who's-who of the AI race: Anthropic, OpenAI, Microsoft, and even Figure, the humanoid robotics company. Here's what caught my eye in the filing, and why it matters well beyond Wall Street.
01.From crypto mining to AI hyperscaler in two years
The origin story is the part I find most fascinating. Nscale didn't start as an AI company at all — it emerged from stealth mode in 2024 as a spinoff of Arkon Energy, a cryptocurrency mining firm. When crypto cooled, the founder, 32-year-old Josh Payne, pivoted the company's power and data center expertise toward the one customer with an even bigger appetite for energy and GPUs: the AI industry. In a letter to prospective investors, Payne wrote that the company was "built with an infrastructure-first thesis, building against contracted customer demand." That early bet now sits at the center of the AI buildout.
02.The numbers: explosive growth, scary losses
Let's be honest — the revenue line is wild. Nscale reported $140.6 million in revenue for the first six months of 2026, up 1,252% from $10.4 million in the same period a year earlier. But the net loss over that stretch widened to $1.02 billion, from $368.9 million. Total assets stand at $17.25 billion against $12.36 billion in liabilities.
The reason both numbers can be true at once is Nscale's business model: long-term take-or-pay contracts, where customers commit to paying for capacity whether or not they end up using it. Active and contracted total contract value hit $103.4 billion as of August 31, 2026, up from $38 billion at the end of 2025. The company says margins should improve as its data centers come online and ramp up. Reports around the filing suggest Nscale is seeking to raise up to $3 billion at a valuation of roughly $30 billion — some reports say up to $35 billion. For context, it was valued at $14.6 billion in March.
03.The Anthropic deal anchoring it all
The single biggest name on that contract list is Anthropic. Last month, the two signed a six-year infrastructure agreement worth up to $45 billion, under which Anthropic leases around 460 megawatts of computing capacity at Nscale's flagship Monarch Compute Campus in West Virginia — a site of roughly 2,250 acres that Nscale says has a clear expansion path past 8 gigawatts. It's part of the same infrastructure scramble we've been tracking: we covered how Anthropic and OpenAI are also hunting smaller, faster-to-deploy data center deals, because every megawatt matters when demand is this hot. And with Anthropic itself reportedly weighing an IPO, the money in this ecosystem is starting to chase itself in circles.
04.Nvidia is everywhere in this filing
Here's the detail that says the most about how the AI economy actually works: Nvidia is simultaneously Nscale's GPU supplier, a customer, and one of its biggest financial backers. The filing revealed that Nscale recently issued $3.1 billion worth of convertible notes, of which at least $1 billion is going to Nvidia. Nvidia has also agreed to guarantee up to $860 million on a Texas data center lease. The chip giant has reportedly invested over $2 billion in Nscale cumulatively, alongside investors like Dell, Nokia, Fidelity and Point72. When your supplier is also your banker, the incentive to keep buying chips is, let's say, well aligned.
05.The boardroom is a Big Tech reunion
One more thing worth knowing: Nscale's board includes Sheryl Sandberg (former Meta COO), Nick Clegg (former Meta president and UK deputy prime minister), and Fidji Simo, who stepped down as OpenAI's product and business chief in July and joined the Nscale board just last week. The company also plans to close its $1.65 billion acquisition of Anyscale — the startup behind the Ray distributed computing framework — at or around the IPO. Compute plus software, all under one roof.
Worth reading alongside this: Anthropic's $2 Trillion IPO: Everything You Need to Know and Anthropic Is Opening a Singapore Office — Its Fifth in Asia-Pacific.
Key Takeaways
- Nscale has filed to list on the NYSE under ticker NSCL, with Goldman Sachs, J.P. Morgan and Morgan Stanley leading the offering
- H1 2026: $140.6 million revenue (up 1,252%), $1.02 billion net loss, $103.4 billion in active and contracted contracts
- Its anchor deal is an up-to-$45 billion, six-year Anthropic agreement at the Monarch campus in West Virginia
- Nvidia is supplier, customer and major financier — including $1 billion of a $3.1 billion convertible note issue
- The company spun out of crypto-mining firm Arkon Energy in 2024 and is reportedly targeting a ~$30 billion valuation
Frequently Asked Questions
What is Nscale and why is its IPO significant?
Nscale is a UK-based AI cloud company that rents out Nvidia GPUs for training and running AI models, with customers including Anthropic, OpenAI, Microsoft and Figure. It has filed its S-1 to list on the New York Stock Exchange under the ticker NSCL, reportedly targeting a valuation of around $30 billion — roughly double its March private valuation.
What are Nscale's key financial numbers in its IPO filing?
Revenue grew 1,252% to $140.6 million in the first half of 2026, up from $10.4 million a year earlier — but the net loss widened to $1.02 billion. Its active and contracted contracts are worth $103.4 billion, anchored by an up-to-$45 billion Anthropic compute deal.
How did Nscale get started?
Nscale emerged from stealth in 2024 as a spinoff of Arkon Energy, a cryptocurrency mining company. When crypto cooled, 32-year-old founder Josh Payne pivoted the company's power and data centre expertise toward AI customers, building what he described as an infrastructure-first business built against contracted customer demand.
What This Means for You
If you're an Indian student or professional watching the AI job market, this is a signal worth reading carefully. The money in AI is flowing into infrastructure — data centers, power, GPU fleets — and Nscale's filing shows those operations need finance, legal, operations and engineering talent that isn't strictly "AI research." For founders and freelancers, there's a quieter lesson: compute is the single biggest line item in almost every AI business, and the gap between Nscale's $140 million revenue and $103 billion contracts is a reminder that costs scale brutally fast in this space. That's exactly why the cost-optimization chapter in the GPT-6 Astra — Zero to Hero Guide is the chapter readers tell us saved them the most money — knowing how to control your token spend matters whether you're running one API key or negotiating megawatts. Beyond that, treat this IPO season as a masterclass: read the S-1 filings of AI companies if you want to understand where the industry is actually going, not where the hype says it's going. And if parsing hundred-page filings isn't your idea of a weekend, our free Digitalaikart AI assistant can break down AI business news in plain language — no sign-up needed to start.
By the way, if terms like agents and computer use still feel fuzzy, our GPT-6 Astra guide for beginners explains them all without the jargon.