Recommended Articles
Fundraising

Thinking Machines Lab and the $40 Billion Valuation Question

By Angeled · November 10, 2026 · Fundraising
Thinking Machines Lab and the $40 Billion Valuation Question
Thinking Machines Lab and the $40 Billion Valuation Question

What is reported

Thinking Machines Lab, the AI research lab, is reportedly raising a $5 billion round led by Accel at a $40 billion pre-money valuation. The report came from Business Insider on November 6, 2026. The round is unconfirmed, and the lab declined to comment.

That framing matters. Nothing in this article treats the round as closed. What can be analysed is the reported structure, the company's known history, and what a valuation of that size would imply.

The lab previously raised a $2 billion seed round in July 2025 at a $12 billion valuation. A move from $12 billion to a reported $40 billion pre-money in about sixteen months would be a more than threefold increase before the company has shipped a broad commercial product suite.

The revenue multiple

TechCrunch reported in September 2026 that the lab's annual revenue was above $100 million. If that figure is taken as the base, a $40 billion valuation implies a revenue multiple of up to around 400 times.

To put that in context, established software companies typically trade at single-digit or low double-digit revenue multiples. High-growth AI companies have commanded far higher multiples, but 400 times is at the extreme end even by that standard. It implies that investors are not valuing current revenue at all. They are valuing a belief about future revenue: that the lab will become one of a small number of frontier AI providers, and that its revenue will grow by orders of magnitude from its current base.

That belief may prove right. It is also, by definition, a bet on a future that has not yet arrived. Founders watching this round should not read a 400-times multiple as a benchmark for their own pricing. It reflects a specific scarcity — a frontier lab with a notable founding team — not a general market rate for AI revenue.

The product so far

The lab's first product is Tinker, a tool for fine-tuning large language models. Fine-tuning lets a customer adapt a general model to a specific task, dataset or style, and it is a natural first commercial step for a research lab: it monetises model capability without requiring the lab to win the full consumer chatbot market on day one.

On November 5, 2026, the lab shipped Inkling 2 Small, a model with 594 billion total parameters and 17 billion active parameters. The distinction matters. Total parameters describe the full size of the model; active parameters describe how much of it is used for any given token. A design with a large total and a smaller active count is built for efficiency — more capability per unit of compute at inference time.

The Nvidia dimension

Nvidia is reportedly considering an investment of around $2.5 billion in the round. Separately, the lab is reported to have a multi-year deal for at least 1 gigawatt of Vera Rubin systems from early 2027.

One gigawatt is an enormous amount of compute capacity. It signals that the lab intends to train and serve models at frontier scale, and that it has secured the hardware supply to do so. For Nvidia, an investment alongside a large hardware commitment aligns the chip supplier with a major future customer. For the lab, it reduces the risk that compute scarcity constrains its roadmap.

Founders should note the structure: the deal is not only about capital. It is about access to scarce infrastructure. In frontier AI, compute access is itself a competitive moat, and partnerships that secure it are as strategically important as the funding round.

The earlier $50 billion attempt

The lab earlier attempted a $50 billion raise that was abandoned after co-founder departures. That history is relevant to the current report in two ways.

First, it shows that even the most sought-after AI labs face limits. A raise can fail, or be withdrawn, when the team changes or when the terms do not align. Second, it shows that the current reported $40 billion figure is lower than the earlier ambition — a recalibration, not simply an escalation.

What this means for founders

For AI founders in particular, the article also illustrates the cost of the frontier game. Competing at this level requires gigawatt-scale compute and billions in capital. Most AI companies will not compete there. Their opportunity is to build on top of frontier models, or in verticals where a smaller, specialised model wins.

What this means for investors

A reported $40 billion pre-money valuation with revenue above $100 million demands a clear thesis about how revenue scales by a factor of ten or more, and how quickly. Diligence questions should include: how much of the revenue is recurring, who the customers are, what Tinker and Inkling adoption looks like beyond early users, and how the Nvidia compute deal affects margins.

Angeled.com connects founders with investors.

Business Insider, November 6, 2026
TechCrunch, September 2026

Read Next