Anthropic and Nscale Reportedly Near Another $45 Billion Compute Deal
Bloomberg reports the two companies may soon announce a second deal of that size, while talks remain open and both firms prepare to go public
By AXY AI Desk ·
Key points
- Bloomberg reports Anthropic and Nscale are on the verge of announcing another $45 billion compute deal, citing people familiar with the matter.
- Sources cautioned that talks are still underway and the details could change.
- The reported deal would follow a $45 billion compute agreement the two companies announced a little over a month earlier.
- Investors have raised concerns that AI companies are spending heavily for too little profit.
- Anthropic is expected to go public this year, and Nscale is seeking a listing amid investor pushback over a proposed $30 billion valuation.
Anthropic and AI infrastructure provider Nscale are reportedly on the verge of announcing another $45 billion compute deal, according to Bloomberg, which cited people familiar with the matter. The sources cautioned that talks are still underway and that the details could change. TechCrunch reached out to both companies for more information and said it would update its report if they responded.
The reported agreement would be the latest in a series of very large computing contracts that Anthropic, the artificial intelligence company behind the Claude chatbot, has signed over the past year. Its scale and timing have drawn attention from investors, customers and rivals, because they show how quickly the cost of building advanced AI is rising.
What 'compute' means and why these deals are so large
In the AI industry, 'compute' refers to the computing power needed to train and run AI models. That power comes from specialised chips, the servers that hold them and the data centres that house, power and cool them. Training a frontier model, the most capable systems a company builds, and then serving millions of users, requires enormous amounts of this capacity. Companies usually cannot wait to build it themselves, so they sign long-term agreements with infrastructure providers to secure access.
That is why a single deal can be worth billions of dollars. The contract is not a one-time purchase of software. It is a commitment to pay for capacity over time, sometimes years, whether or not the AI company's revenue has grown to match. For the provider, a large signed commitment helps justify building new facilities and buying equipment. For the AI company, it secures the room to keep developing models while rivals compete for the same scarce hardware.
A pattern of giant commitments
The report places the new talks in the context of a deal that was announced a little over a month earlier. At that time, Anthropic and Nscale announced a $45 billion compute agreement. If the second agreement goes ahead, it would mean the two companies had struck two deals of the same headline size within a few months.
Bloomberg's sources described the second deal as a near-final step, not a finished contract. That distinction matters. In large corporate negotiations, reports that an agreement is on the verge of being announced sometimes lead to a formal announcement, but terms such as price, capacity, timing or the structure of payments can change before anything is signed. Until the companies publish a statement, the figure should be treated as reported rather than confirmed.
Analysts who follow the sector often describe this kind of rapid, large-scale contracting as a way for AI developers to lock in supply early. The logic is that if demand for AI services continues to grow, firms that secured capacity first will be better placed than those that waited. The risk is that commitments made today may prove larger than what the business can use or pay for later.
Investor unease about spending and profit
The scale of these commitments has made some investors uneasy. According to the TechCrunch report, a segment of the investment community worries that AI companies are spending heavily for limited profit. The concern is not that the technology lacks potential, but that the gap between enormous upfront costs and current earnings could be difficult to close, especially if prices for AI services fall or competition intensifies.
Anthropic has nonetheless continued to pursue these partnerships. The source material says the company is driven by the belief that the spending will pay off in the end. In other words, the company is making a strategic bet that future demand and revenue will justify the capacity it is securing now. Whether that bet succeeds will depend on how quickly AI products generate income relative to their cost.
This tension between investment and returns is one of the central questions for the wider AI sector. Companies that spend the most on infrastructure can gain technical and commercial advantages, but they also take on more financial risk. Investors are therefore watching not only the size of individual deals, but whether those deals translate into durable earnings.
Both companies are preparing to go public
The talks also come at a time when both companies are pursuing initial public offerings, or IPOs, the process by which a private company sells shares to the public for the first time. According to the report, Anthropic is still expected to go public this year. Nscale is also seeking a public listing, but Bloomberg reported on Monday that the company is facing investor pushback over its proposed valuation of $30 billion. A valuation is the total value that investors place on a company, and pushback suggests some potential buyers believe the price is too high.
The timing creates a complicated set of incentives. A large compute deal can signal growth and secure future capacity, which may appeal to public-market investors. At the same time, investors weighing an IPO will examine how much the company is committing to spend and whether those commitments are backed by revenue. Any significant new deal could therefore be read through the lens of an upcoming share sale, with supporters viewing it as proof of momentum and skeptics seeing it as added cost.
For Nscale, the investor response to its proposed valuation is a particularly visible test. A company seeking a listing needs buyers willing to accept its price. Pushback on valuation may lead to changes in terms, a different timetable or a lower figure, although the source material does not say which outcome, if any, is under consideration.
Why it matters beyond the two companies
For ordinary readers, deals of this kind can seem remote from daily life. Yet the capacity they secure shapes how quickly AI tools are developed, how widely they are offered and how reliable they are at peak demand. Large infrastructure commitments also influence where data centres are built and how much electricity and equipment the industry requires, which can affect communities, utilities and supply chains.
The deals also matter for the financial system. Major AI infrastructure contracts are closely watched by investors in chipmakers, cloud providers and data centre operators, as well as by the companies directly involved. If the reported agreement is announced, it could be read as a sign that demand for AI computing remains strong. If talks falter, or if IPO pricing comes under pressure, it could prompt a reassessment of how much the sector should be expected to spend before earning returns.
Competitors are also affected. When one leading AI developer secures large amounts of capacity, others may feel pressure to make similar commitments to keep pace. That dynamic can drive a cycle of escalating spending, which is why some observers see the current period as a test of how far the industry can push infrastructure investment before the economics are fully proven.
What to watch next
Several developments will determine how this story unfolds. The most immediate is whether Anthropic and Nscale formally announce the agreement, and if they do, whether the stated value and terms match the figure reported by Bloomberg. Neither company had responded to TechCrunch's requests at the time of the source report, so an official statement would be the first confirmation from the parties themselves.
Investors will also be watching the IPO process. Anthropic's expected listing this year and Nscale's valuation discussions will show whether public-market buyers are willing to back large AI infrastructure spending. How those offerings are received could influence the terms of future compute deals and the broader appetite for AI investment.
Finally, the performance of AI products will be central. The core question raised by the source material is whether the spending will pay off. Revenue growth, customer adoption and pricing over the coming months will provide the clearest signal, and will help determine whether the long string of giant compute deals is seen as a strategic advantage or as an expensive gamble.
Questions and answers
What is the Anthropic and Nscale deal about?
Bloomberg reports that Anthropic and Nscale are on the verge of announcing another $45 billion compute deal. Compute refers to the computing power needed to train and run AI models, and the sources said talks are still underway and details could change.
Has the second Anthropic and Nscale compute deal been confirmed?
Not in the source material. The report cites people familiar with the matter, and TechCrunch said it had reached out to both companies and would update its story if they responded.
Why are investors concerned about AI compute deals?
Some investors worry that AI companies are spending too much money for too little profit. Anthropic has continued these partnerships because it believes the spending will pay off over time.
Are Anthropic and Nscale going public?
Both companies are pursuing IPOs. Anthropic is still expected to go public this year, while Nscale is facing investor pushback over its proposed $30 billion valuation, according to Bloomberg.
Topics: Anthropic, Nscale, AI compute, AI infrastructure, AI IPO, Bloomberg, AI investment
Sources: TechCrunch AI