Just over a month after Anthropic and the AI-infrastructure specialist Nscale disclosed a $45 billion agreement to secure compute resources, the pair appear to be negotiating a second deal of comparable size. The earlier contract marked one of the largest single-purpose AI compute purchases announced to date, underscoring the rapid scaling of model training demands.
Bloomberg cites sources who say the companies are on the brink of announcing the new $45 billion arrangement, while also noting that negotiations remain active and final terms could shift before any public filing. The report emphasizes that the information comes from individuals familiar with the discussions, and no official confirmation has been made.
If sealed, the contract would extend a pattern of multi-billion-dollar compute deals that Anthropic has pursued throughout the past twelve months, a strategy that has drawn scrutiny from investors wary of high capital outlays relative to near-term earnings. Critics argue that pouring vast sums into hardware and cloud capacity may not translate into proportional revenue growth, especially as competitive pressures intensify.
Anthropic has continued to forge such partnerships, maintaining that the scale of compute will ultimately deliver competitive advantage and market share gains. Company executives contend that the long-term payoff from accelerated model development outweighs the short-term expense, a view that aligns with their broader growth narrative.
Both firms are simultaneously preparing for public listings. Anthropic is slated to file for an IPO within the current calendar year, while Nscale has also signaled intent to go public. However, Bloomberg notes that Nscale faces investor resistance concerning its proposed $30 billion valuation, suggesting market skepticism about the pricing of such compute-centric businesses.
TechCrunch has reached out to representatives from Anthropic and Nscale for comment on the rumored agreement. The outlet indicated it will revise the story should the companies provide additional details or confirm the transaction, reflecting ongoing journalistic monitoring of large-scale AI financing activity in the sector.