On 14 October, Labelbox said it secured Series E funding of more than $100 million through a mix of equity and a credit facility, following growth of more than fivefold year on year. Labelbox’s blog post put its pre-money valuation at $1.25 billion, FourWeekMBA reported.
Key points
- Infinitum Partners led the Series E, with Wipro Ventures and Andreessen Horowitz among the participants, Labelbox said.
- Labelbox said it reached profitability in 2026 and works with more than 400 customers.
- The company plans to fund work on Horizon for AI labs and Recursion Managed Agents for enterprises.
- Labelbox said it uses more than 50 agents across its own operations.
Infinitum leads equity and credit financing
Infinitum Partners led the round, according to Labelbox. The company named Wipro Ventures, Andreessen Horowitz, B Capital, Tau Ventures and Snowpoint Ventures as participants. It said the financing would support research and engineering for its two principal products, alongside expansion of the teams bringing them to customers.
The $1.25 billion figure is a pre-money valuation: the price assigned to Labelbox before the new financing. FourWeekMBA reported that the company did not divide the more than $100 million between equity and the credit facility. The publication also reported that Labelbox gave no revenue figure or post-money valuation. Those distinctions matter to the terms of the round because borrowing and newly issued equity are different forms of funding.
Labelbox said it reached profitability in 2026 while growing more than fivefold year on year. The company also said it serves more than 400 customers, a group it described as spanning AI labs, hyperscalers, startups and Fortune 500 companies. Manu Sharma, its chief executive and co-founder, said the funding would allow it to develop infrastructure for model training and agents that learn from their work.
Horizon supplies training environments to AI labs
Horizon supplies reinforcement learning environments, expert feedback and evaluation infrastructure to AI labs, Labelbox said. Its simulated tasks cover coding, research and work that takes multiple steps. Models can attempt those tasks, receive feedback and train on the results. Labelbox said it works with AI labs and hyperscalers on programmes conducted after a model’s initial training.
Labelbox described its customer reach as “over 90% of leading U.S. AI labs”. FourWeekMBA reported a different description in the company’s blog post: “90% of the leading frontier AI labs and hyperscalers” work with it through Horizon. The two descriptions differ both in the customer groups they count and in how they state the percentage.
The company said it would continue developing Horizon with proceeds from the financing. Its stated work spans the environments in which models practise tasks and the evaluations that provide feedback on their performance. Labelbox also offers data and training tools alongside its reinforcement learning infrastructure.
Recursion Managed Agents moves into enterprise use
Labelbox said it opened Recursion Managed Agents to enterprise customers earlier in October. The product places agents in applications that company teams already use. According to Labelbox, schedules, alerts and tickets can initiate their work, while agents can coordinate across tools and have their output assessed against defined standards.
Each run can contribute to an agent’s memory and affect how it handles later tasks, the company said. That ties Recursion’s enterprise product to Horizon’s training work: both use feedback from a completed task to inform a subsequent attempt. Labelbox said the financing would pay for further research and engineering on Recursion as well as Horizon.
Labelbox said more than 50 agents already work inside the company across engineering, research, product, sales, marketing and finance. Recursion Managed Agents extends that approach to customers deploying agents in their own applications, according to the company. Labelbox said those agents can be evaluated against standards set for their output.
