Anthropic said to plan employee tender offer at $350b valuation

Anthropic is reportedly planning a secondary share offering that would allow some employees to sell shares at a valuation of at least US$350 billion, according to Bloomberg sources.

The company is also in the process of raising over US$20 billion in a funding round, which is linked to this plan.

The valuation mentioned is pre-money and aligns with the ongoing fundraising discussions.

Details of the tender offer, including the final transaction amount and timing, have not been finalized.

The move aims to provide liquidity options for staff amid a competitive AI talent market, similar to practices at other tech firms like Stripe, SpaceX, and OpenAI.

Anthropic has not publicly commented on the matter.

🔗 Source: Bloomberg

🧠 Food for thought

Implications, context, and why it matters.

Employee share sale comes after a surge of funding and heavy spending

The company’s valuation jumped from US$183 billion in September 2025 1 to around US$350 billion in recent reports 2.
That rise has come with steep cash needs, including a pledge to buy US$30 billion in Azure cloud compute from Microsoft 3 plus a US$50 billion plan to build data centers with Fluidstack (a company that designs and operates large-scale computing infrastructure) in Texas and New York 4.
Revenue growth helps cover the bill, with Anthropic’s annual run-rate revenue moving from about US$1 billion at the start of 2025 to more than US$5 billion by August 2025 1.
A secondary share offering (tender offer) would let some employees sell shares at a valuation of at least US$350 billion, with a possible IPO as early as 2026 based on reports 2.

Compute and capital deals reshape how frontier AI gets funded

The structure suggests some frontier AI work is getting financed in new ways that go past standard venture capital.
Microsoft and Nvidia have said they plan to put billions into Anthropic, while Anthropic has committed to spend tens of billions on cloud and chip capacity, including Azure and NVIDIA systems 3.
The arrangement locks in large, long-term customers for cloud providers and gives AI labs money to keep pace, while some investors worry it looks like an “AI bubble” 5.
Anthropic runs compute across Google TPUs (Tensor Processing Units, Google’s AI chips), Amazon Trainium (Amazon’s AI training chips), and NVIDIA GPUs (graphics processing units widely used for AI), which the company says spreads risk and reduces reliance on one vendor 6.