Meta Meta Turns Muse Into an App Platform: New Spaces Build Persistent Hubs From a Prompt and Update Themselves in the Background Security Censys Finds 1,465 Openly Exposed Consoles for Artex, the AI Hacking Framework Behind the Korean Bank and ConnectWise Attacks Business Alphabet's Filing Reveals a $548 Million Qualcomm Stake as Google Backs a Chipmaker Chasing Nvidia's Data Center Business News AI Physics Startup PSI Says a Non-Public Anthropic Model Cracked an 18-Year-Old Quantum Conjecture in Three Days News Tsinghua-MIT Startup TensorWorld Says It Raised a $105 Million Seed Round to Build Physics World Models Security METR Audit Finds Anthropic's Model Progress Nearly Doubled and May Have Crossed Its Own Warning Threshold in July Business Microsoft Winds Down Its Oracle AI Cloud Deal and Puts Outside GPU Suppliers Under Review AI Agents After Blocking Muse, Amazon Now Lets US Users Text Alexa to Shop Ahead of the Holidays Meta Meta Turns Muse Into an App Platform: New Spaces Build Persistent Hubs From a Prompt and Update Themselves in the Background Security Censys Finds 1,465 Openly Exposed Consoles for Artex, the AI Hacking Framework Behind the Korean Bank and ConnectWise Attacks Business Alphabet's Filing Reveals a $548 Million Qualcomm Stake as Google Backs a Chipmaker Chasing Nvidia's Data Center Business News AI Physics Startup PSI Says a Non-Public Anthropic Model Cracked an 18-Year-Old Quantum Conjecture in Three Days News Tsinghua-MIT Startup TensorWorld Says It Raised a $105 Million Seed Round to Build Physics World Models Security METR Audit Finds Anthropic's Model Progress Nearly Doubled and May Have Crossed Its Own Warning Threshold in July Business Microsoft Winds Down Its Oracle AI Cloud Deal and Puts Outside GPU Suppliers Under Review AI Agents After Blocking Muse, Amazon Now Lets US Users Text Alexa to Shop Ahead of the Holidays

Thinking Machines Lab Is Raising $5 Billion at a $40 Billion Valuation and Just Shipped Inkling 2 Small, a 594B-Parameter MoE Going Open Weight Next Week

Thinking Machines Lab Is Raising $5 Billion at a $40 Billion Valuation and Just Shipped Inkling 2 Small, a 594B-Parameter MoE Going Open Weight Next Week

Business Insider reports Mira Murati's Thinking Machines Lab is closing a $5 billion round led by Accel at a $40 billion pre-money valuation, more than triple its last public mark. The same day it unveiled Inkling 2 Small, a 594B-parameter MoE with 17 billion active parameters, whose weights arrive next week under Apache 2.0.

Thinking Machines Lab, the artificial intelligence startup founded by former OpenAI chief technology officer Mira Murati, is raising $5 billion in a round led by Accel at a $40 billion pre-money valuation, according to people familiar with the matter cited by Business Insider. The transaction could close as soon as this week, and would more than triple the company's valuation from its last public financing. A company representative declined to comment, and Accel did not respond to a request for comment.

The round marks a striking recovery for a startup that has been through a turbulent year. Thinking Machines raised a $2 billion seed round last year at a valuation exceeding $10 billion before it had shipped a product. It then sought a round at a $50 billion valuation, but those discussions were abandoned. AI Weekly reported that NVIDIA is separately in talks to invest roughly $2.5 billion, a detail that has not been confirmed by either company and should be treated as provisional.

The capital push comes despite notable departures. As Business Insider reported in May, about a third of the founding team has left the company, driven in part by large compensation packages at rivals such as Meta and OpenAI. The exits included founding members Barret Zoph, Luke Metz and Andrew Tulloch. The Information reported in September that a new financing was in the works.

On the same day the funding report landed, the lab released its second-generation model, Inkling 2 Small. It is a sparse mixture-of-experts model with 594 billion total parameters and about 17 billion active per token. The company says it went through four stages of pre-training followed by nine days of joint reinforcement training, and that it scores far above its predecessors on analytical intelligence across coding, knowledge work and reasoning. Those figures are company-reported and have not been independently verified.

Third-party benchmark aggregator Artificial Analysis placed Inkling 2 Small alongside Gemini 3.8 Flash (high) on its index, and noted that if the weights ship as promised it would become the strongest open-weights model available outside China. That is the number that will matter more than the funding headline: the open-weights arena is currently defined by Chinese labs such as DeepSeek, Qwen and Kimi, and a credible Western challenger changes the competitive picture.

The weights are scheduled for release next week under the Apache 2.0 license, after the model completes safety testing. The company also shipped a new version of Tinker, its fine-tuning platform, on Thursday, with what it describes as improved AI coding capability.

The round is a signal that despite weeks of market volatility around AI stocks, investors remain willing to fund frontier-adjacent labs at steep premiums. The test now shifts to execution: Inkling 2 Small's open-weights release next week will show whether Thinking Machines can translate a $40 billion paper valuation into real presence on the leaderboards that developers actually check.

Comments (0)

Log in to join the discussion

Log In

No comments yet