Thinking Machines to Raise $1B at $40B Valuation

Thinking Machines is in discussions to raise $1 billion in funding. The valuation is said to be at least $40 billion. According to reporting by Marina Temkin of TechCrunch AI, existing investor Accel is in talks to lead the round. The report was published on September 3, 2026.

The company is an AI lab founded early last year by former OpenAI Chief Technology Officer Mira Murati. In its previous round, it raised $2 billion at a $12 billion valuation. If completed, the valuation would more than triple in a short period. Annualized revenue is said to exceed $100 million, focusing scrutiny on whether it justifies the high multiple.

Thinking Machines, the AI lab founded early last year by former OpenAI CTO Mira Murati, is in discussions to raise $1 billion at a valuation of at least $40 billion

The above is a quotation from the original source. This article is a summary and analysis of reporting based on TechCrunch AI. The facts rely on the primary reporting and testimony from sources.

Structure of the Fundraising Talks at a $40

Billion Valuation

At the center of the negotiations is existing shareholder Accel. It is said to be in talks to put together the $1 billion raise as the lead. Reporting by The Information and testimony from sources are said to align. Both companies did not immediately respond to requests for comment.

The focus is on the level of the valuation. It is below the $50 billion the company reportedly sought late last year. Still, more than $40 billion is well above the previous $12 billion. The $1 billion raise itself is also large for foundation-model development.

The previous round was positioned as one of the largest seed financings in history. It was led by Andreessen Horowitz, with participation from Nvidia, GV, Lightspeed, and Conviction Partners. The basis for the investment was the backgrounds of Murati and former OpenAI researchers. This time as well, the scarcity of talent is seen as underpinning the valuation.

Unlike public markets, the terms of large private raises are undisclosed. It is difficult for outsiders to verify whether the valuation and revenue multiple are reasonable. Even so, the $40 billion figure will become a market benchmark. It is expected to spill over into other companies’ next raises.

Revenue Base and the Role of Tinker

Annualized revenue is said to exceed $100 million. The figure is based on financial information from sources. A $40 billion valuation implies an extremely high multiple on revenue. It is a level that raises questions about revenue sustainability.

The pillar of monetization is Inkling, released in July. It is offered as an open-weight model. It charges usage-based compute fees for adapting the model with proprietary data. The delivery platform is Tinker.

Tinker is a platform that opens training and adaptation work to outside users. Users run fine-tuning processes using their own proprietary data. Costs are incurred according to the computing resources used. Its structure is close to cloud-style pay-as-you-go billing.

Unlike conventional API billing, the adaptation work itself becomes a source of revenue. It is designed to disclose the model weights while earning compensation from operations. It is characterized by making reproducibility and verifiability easier to ensure. It is seen as aiming to lower adoption barriers for enterprise use.

The scale of revenue is still at an early stage. More than $100 million is small compared with major labs. Even so, reaching that level in a short time after product launch is fast. Going forward, contract renewal rates and unit prices will be the focus.

Founding Team Departures and Organizational

Restructuring

At its founding, its centripetal force was the depth of its researchers. Former OpenAI researchers gathered at its core. Some of them have departed. Lilian Weng and Luke Metz and others are said to have returned to OpenAI.

Departures of founding members are not uncommon for early-stage labs. Differences in research direction and productization speed are often factors. Thinking Machines is also in a transition to a product organization. Authority and responsibilities are seen as being redistributed.

The expertise of the remaining team is still assessed as high. Market confidence in Murati’s leadership remains strong. However, dependence on specific individuals is an operational weakness. It is at a stage where reproducibility as an organization is required.

Talent mobility is intensifying among AI labs. Large players seek to retain staff with compensation and computing resources. Emerging players counter with autonomy and equity stakes. Thinking Machines’ next hires are seen as indicating the stability of its structure.

Demand for Compute Infrastructure and the

Open-Model Trend

Behind the large raise is securing computing resources. Demand for training and inference continues to grow. Building high-performance computing infrastructure is an essential requirement for labs. Advances in computing infrastructure, such as Featuring Intel Xeon 7 Diamond Rapids with 256 Cores and 1.28TB Cache, support the development race.

Releasing open weights is an industry trend. Publishing weights advances external verification and derivative development. Development originating from public resources is also active, as in AI Video Generator MoneyPrinterTurbo Surges on GitHub. Inkling is seen as positioned in this trend.

At the distribution end for generated content, display technology is also evolving in parallel. Video-quality initiatives such as LG OLED evo, Creator Original Picture Mode with Prime Video provide a basis for evaluating generated video. Both model performance and viewing environments affect the experience. The market is seen as forming through a chain from infrastructure to devices.

Tinker-style adaptation billing sits in the middle of this chain. It monetizes model provision and compute provision together. It aims to capture demand for optimization for specific uses. The degree of vertical integration is assessed to determine competitiveness.

Heated AI Investment and the

Revenue-Multiple Debate

A $40 billion valuation corresponds to 400 times revenue. It is on a different order of magnitude from typical corporate revenue multiples. It is a figure that prices in sustained growth. It can be interpreted as a bet on a future dominant position.

The rise from the previous $12 billion valuation is sharp. Product releases and revenue generation are said to be drivers. Still, it did not reach the requested $50 billion level. It is seen as the result of the market demanding a degree of discipline.

Investors emphasized backgrounds and technology assets. Nvidia’s participation indicates a link to compute supply. Participation by GV and Lightspeed indicates expectations for product rollout. Accel’s continued lead is seen as meaning confidence from the inside.

On the other hand, justifying the revenue multiple requires accumulating contracts. Usage-based billing is susceptible to fluctuations in use. Continued use by large customers is essential. Verifying undisclosed renewal rates remains a challenge.

Editorial Opinion

In the short term, concentration of funding into generative AI infrastructure is expected to advance further. A valuation on the $40 billion scale is assessed to push up fundraising terms for other companies. If Tinker-style usage-based billing takes hold, price competition for training infrastructure is expected to intensify.

In the long term, the validity of the revenue multiple is expected to become a touchstone for the market as a whole. Vertical integration of proprietary models and development platforms is assessed to become mainstream. Whether researcher-led organizations can shift to sustainable operations will be the focus.

As a question from the editors, whether a revenue path commensurate with a $40 billion valuation is clear remains an issue. Whether usage-based billing is compatible with customers’ desire to reduce fixed costs is unverified. The impact of founding-team departures on technological advantage also needs close examination.

References

• “Accel reportedly in talks to lead $1B round for Thinking Machines at $40B valuation”, by Marina Temkin — TechCrunch AI, 2026-09-03T19:36:29.000Z (ARR)
• Source URL: https://techcrunch.com/2026/09/03/accel-reportedly-in-talks-to-lead-1b-round-for-thinking-machines-at-40b-valuation/