Thinking Machines Lab, the artificial intelligence company established by former OpenAI chief technology officer Mira Murati, is advancing toward the close of a substantial new funding round.
According to reporting from multiple sources familiar with the discussions, the San Francisco-based startup is raising approximately $5 billion at a $40 billion pre-money valuation, with venture firm Accel positioned to lead the investment.
The round could finalize as soon as the coming days, potentially elevating the company’s post-money valuation near $45 billion if completed at the reported terms.
This prospective raise would more than triple the valuation achieved in Thinking Machines Lab’s initial major financing. In July 2025, the company secured roughly $2 billion in a seed round that valued it at about $12 billion post-money.
That earlier financing, led by Andreessen Horowitz and joined by participants including Nvidia, Accel, AMD, Cisco, ServiceNow, and Jane Street, ranked among the largest early-stage investments in the AI sector.
Subsequent reports indicated the company had explored a follow-on round targeting valuations around $50 billion in late 2025, though those talks did not result in a completed deal.
Founded in early 2025 after Murati’s departure from OpenAI, Thinking Machines Lab has emphasized the development of AI systems designed for greater customizability and collaboration with human users.
Its primary commercial offering is Tinker, a platform that enables organizations and researchers to fine-tune large language models using their own data through a managed API, charging based on compute usage.
The company has also released open-weight models under the Inkling name, including a recent Inkling 2 Small variant that supports multimodal inputs and aims to deliver competitive performance with greater efficiency.
A new iteration of Tinker was introduced in early November 2026, with the company noting improvements in AI-assisted coding capabilities.
The lab has faced notable personnel shifts since its founding.
Roughly one-third of its original team has departed, with several co-founders—including individuals who previously contributed to early ChatGPT development—moving to other organizations such as OpenAI and Meta.
These transitions have occurred against a backdrop of intense competition for AI talent, including large compensation packages offered by established labs.
Despite the turnover, the company has continued to attract specialized talent and has formed a multi-year strategic partnership with Nvidia.
Announced in March 2026, that arrangement includes a significant undisclosed investment from Nvidia and a commitment for Thinking Machines Lab to deploy at least one gigawatt of Nvidia’s Vera Rubin systems beginning in 2027, supporting both model training and scalable customizable AI platforms.
Revenue generation remains in an early phase relative to the proposed valuation.
Sources have indicated an annualized run rate exceeding $100 million, primarily driven by Tinker’s usage-based model.
At a $40 billion pre-money figure, this implies a high revenue multiple, reflecting continued investor willingness to back teams with strong pedigrees even as product traction and competitive dynamics evolve.
These developments carry implications for the fintech industry, where AI capabilities increasingly underpin core functions such as credit assessment, fraud detection, algorithmic trading, personalized financial advice, and regulatory compliance.
Platforms like Tinker that lower the barriers to fine-tuning models on proprietary datasets could enable financial institutions to adapt general-purpose AI more effectively to domain-specific needs—such as incorporating internal risk models, transaction histories, or compliance rules—without building full-scale training infrastructure from scratch.
Open-weight releases and customization tools may further support fintech innovators seeking greater control over model behavior, interpretability, and data privacy compared with fully closed systems.Indirect effects on the broader financial technology ecosystem are also noteworthy.
Jane Street’s participation in the seed round already illustrates interest from quantitative trading firms, which could benefit from advances in models capable of processing complex financial data or generating synthetic scenarios for stress testing.
The substantial capital inflow and compute partnerships signal sustained investor confidence in AI infrastructure, which may influence funding availability and valuation expectations for adjacent fintech startups focused on AI-driven payments, insurance technology, or wealth management tools.
Talent mobility between frontier labs and industry applications could accelerate the transfer of techniques into financial services, while the high multiples attached to early AI revenue may encourage similar ambitions among fintechs integrating advanced models.
Over time, more accessible customization platforms might also shape initiatives in RegTech and risk management by allowing institutions to iteratively refine AI systems in response to evolving regulations and market conditions.
The ultimate terms and timing of the round remain subject to completion, and company representatives have declined to comment publicly on the reports.
As Thinking Machines Lab continues product development and scaling efforts, the outcome will provide further indication of how capital markets are pricing the combination of research ambition, commercial tools, and competitive pressures in the current AI landscape.
