Labelbox, a San Francisco company that supplies data and training infrastructure to AI developers, said on 14 October 2026 that it had raised more than $100m in a Series E round. The package combines equity with a credit facility, and the company did not say how the total splits between the two.
Infinitum Partners led the round. Wipro Ventures, Andreessen Horowitz, B Capital, Tau Ventures and Snowpoint Ventures also took part, along with other investors the company did not name.
Labelbox's press release did not give a valuation. Unite.AI and Business News Today both reported a pre-money valuation of $1.25bn; the post-money figure has not been disclosed.
“Every company will soon run on AI systems that learn from their own work. Getting there takes innovation at every layer, from sandbox infrastructure and the agent harness to reinforcement learning post-training. Labelbox brings those layers together in one platform, and this funding lets us push each one further, so that frontier labs and enterprises can build AI that gets better with every run.”Manu Sharma, CEO and co-founder, Labelbox
What Labelbox does
Labelbox describes itself as a provider of data, reinforcement learning infrastructure and a managed agent platform. It sells two main products. Horizon gives AI labs reinforcement learning environments, expert feedback and evaluation tools. Its environments simulate multi-step tasks in areas such as coding and research, so that models can practise, be scored on their work and improve through reinforcement learning. The company says it works with leading AI labs and hyperscalers on post-training programmes.
The second product, Recursion Managed Agents, lets companies run groups of AI agents inside the software their staff already use. Agents can start work from schedules, alerts or tickets, use several tools, and have their output checked against standards the customer sets. Each run can feed the agent's memory so that it handles later tasks differently. Labelbox opened the product to enterprise customers earlier in October, and says it runs more than 50 agents of its own across engineering, research, product, sales, marketing and finance.
Growth and profitability
The company says it has more than 400 customers, including over 90% of the leading US AI labs, as well as hyperscalers, startups and Fortune 500 companies. It also said it reached profitability in 2026 while growing more than five times year on year. It did not publish revenue or profit figures, and it has not disclosed its headcount.
Chief executive Manu Sharma, who co-founded the company, said Labelbox brings agent infrastructure, sandbox environments and reinforcement learning post-training into one platform, and that the new money would let it push each of those layers further.
What the money is for
Labelbox said the funding will pay for more research and engineering on Horizon and Recursion Managed Agents, and for larger teams to bring both products to customers.
According to Business News Today, its previous round was a $110m Series D in January 2022 led by SoftBank Vision Fund 2, with Databricks Ventures, Snowpoint Ventures, B Capital and Andreessen Horowitz among the participants. Unite.AI reports that the company was founded in 2018. Labelbox did not give a figure for its total funding to date.
Where Labelbox fits
Labelbox argues that as AI models take on more complex work, the next problem is getting them to learn from experience and perform reliably in real settings. It is pitching itself both to the labs that train frontier models and to companies that want to deploy agents, with the same idea of systems that improve from their own work running through both products.
Labelbox did not say how large the credit facility is or which lender provides it. Unite.AI reported that the company also runs a network of more than 4m experts who supply feedback for training data.
Company website: labelbox.com