AI Industry
Labelbox Raises $100M+ at $1.25B Pre-Money
Oct 14, 2026·3 min read·By Rahish Saifi
On October 14, 2026, Labelbox announced it has raised more than $100 million in Series E financing—including equity and a credit facility—at a $1.25 billion pre-money valuation. In a post by CEO and cofounder Manu Sharma, the company said the round was led by Infinitum Partners, with participation from Wipro Ventures, a16z, B Capital, Tau Ventures, Snowpoint Ventures, and others.
The headline is not just the unicorn mark on paper. Labelbox says it reached profitability this year while growing more than 5× year over year—a combination that still stands out in AI infrastructure, where heavy compute and R&D often push growth-stage companies deep into the red.
What Labelbox is building now
Labelbox started in 2018, founded by Brian, Dan, and Manu Sharma, as a training-data platform. Sharma’s Series E letter frames the company around three product chapters that mirror how frontier labs actually ship models today:
- Data: labeling and post-training data workflows—the core business many teams still associate with the Labelbox name.
- Environments: Horizon, Labelbox’s reinforcement-learning environment stack for frontier labs. The company says it now works with 90% of the leading frontier AI labs and hyperscalers—positioning Labelbox as infrastructure for RL-heavy post-training, not only static datasets.
- Agents: Recursion Managed Agents, which Labelbox opened to customer teams last week. Internally, Labelbox says it already runs 50+ agents on the same stack it sells.
That arc—from annotation to simulated worlds to managed agents—tracks where model builders spend marginal dollars after pretraining: curated data, RL loops, and agent orchestration at scale.
Where the new capital goes
Labelbox plans to accelerate R&D and engineering for Horizon and Recursion. Sharma also stresses compute access as strategically important for customers training and evaluating frontier systems—an area where investor networks can matter as much as cash.
Infinitum Partners’ John Yetimoglu, noted in Labelbox’s announcement as an early Groq investor now involved with GroqCloud, underscores that theme: Series E capital here is partly about tying Labelbox’s environment and agent products to the compute supply chain labs already fight over.
Round snapshot
| Detail | Announced terms |
|---|---|
| Round | Series E (equity + credit facility) |
| Size | More than $100 million |
| Pre-money valuation | $1.25 billion |
| Lead investor | Infinitum Partners |
| Other participants | Wipro Ventures, a16z, B Capital, Tau Ventures, Snowpoint Ventures, and others |
Labelbox’s prior $110 million Series D (January 2022, led by SoftBank Vision Fund 2) brought total funding to $189 million with no disclosed valuation (GlobeNewswire); $1.25 billion pre-money is its first disclosed unicorn mark.
Why this matters for AI teams
For data and eval leads: A profitable, fast-growing Labelbox reduces the odds that post-training tooling gets starved during a funding winter—but it also raises expectations that labeling, RL environments, and agent ops stay on one platform. If you already use Labelbox for production labels, Horizon and Recursion are the upsell path Sharma is funding with this round.
For RL and post-training engineers: Horizon’s claim of broad frontier-lab adoption matters if you are choosing environment vendors for reward modeling, tool-use sandboxes, or long-horizon agent evals. The Series E bet is that RL environments become as standardized as labeling pipelines were five years ago.
For agent builders: Recursion opening to every team “last week” plus 50+ internal agents is Labelbox arguing it eats its own cooking. Compare that posture to generic agent frameworks that lack managed runtime SLAs—Labelbox is selling operated agents, not just SDKs.
For investors and competitors: Profitability at 5×+ growth with a $1.25B pre-money sets a benchmark for AI infra vendors; expect data, RL sim, and MLOps incumbents to respond with partnerships or pricing—not only press releases.
Practical takeaways
- Revisit your data + RL stack: If post-training mixes offline labels and online RL, map whether Horizon replaces bespoke env code you maintain today.
- Pilot Recursion early: Newly launched products often ship with design-partner pricing or support—worth a scoped trial if you already trust Labelbox for data governance.
- Model compute contracts: Yetimoglu’s Groq/GroqCloud tie-in is a signal to ask Labelbox (and your cloud vendors) how Series E partnerships affect reserved capacity or credits for heavy eval runs.
- Do not chase the valuation: The company disclosed pre-money, not post-money; size your partnership on product fit and unit economics, not the billion-dollar headline alone.
Bottom line
Labelbox’s Series E backs data, RL environments, and managed agents—not labeling alone. More than $100 million at $1.25 billion pre-money, led by Infinitum Partners, funds Horizon and Recursion while the company claims profitability on 5×+ growth. Frontier labs are shifting spend from static datasets to simulated worlds and operated agents; Labelbox raised to meet them there.