Nvidia is in talks to invest in AI data labeling firm Mercor as part of a funding round that could double the startup's valuation to $20 billion.

  • Existing investor General Catalyst is in discussions to lead a financing round for Mercor at a $20-billion valuation, doubling its October valuation of $10 billion. 
  • Nvidia paid Mercor millions of dollars last quarter to source specialized human-expert data for its open-source Nemotron models.
  • The potential deal aligns with Nvidia's aggressive investment surge, which reached $18.6 billion in the quarter ending in April.

Nvidia Corp. (NVDA) is in discussions to back the artificial intelligence data-labeling startup Mercor in a new funding round that would value the three-year-old firm at $20 billion.

Venture capital firm General Catalyst, an existing backer of Mercor, has been in talks to lead the investment round, according to a report published Wednesday by The Information. The proposed terms and precise size of Nvidia's potential contribution could not be determined.

NVDA stock eased 1% at the close on Wednesday.

Nvidia’s Open-Source AI Capabilities

The potential investment underscores Nvidia’s growing commitment to developing its open-source Nemotron AI models, designed to rival top open-weight systems globally. While Mercor historically derived the majority of its revenue from closed-source AI developers such as OpenAI, Google, and Anthropic, its business with Nvidia has expanded rapidly.

In the previous quarter alone, Nvidia paid Mercor tens of millions of dollars for high-quality data curated by domain experts across legal, financial, and scientific fields. According to company disclosures, Mercor's data was integrated into Nvidia’s two most recent Nemotron releases. To support this accelerating demand, a dedicated group of Mercor employees now focuses almost entirely on Nvidia operations.

Mercor’s growth trajectory has accelerated alongside this demand; the company’s annualized gross revenue reached $2 billion in June, doubling its figure from earlier in the year.

Hybrid Data Strategy And Multi-Sourced Training

Despite ramping up spending on human-evaluated data, Nvidia still relies heavily on AI-generated synthetic data to train the Nemotron family, as per the report. The chipmaker also uses data gathered when developers interact with Nemotron models via OpenRouter, a platform that lets developers access models across multiple providers.

Beyond Mercor, Nvidia sources human training data from external vendors such as Turing and Scale AI, as well as from its internal data operations. Nvidia previously participated in Scale AI’s 2024 funding round, which valued that startup at $14 billion.

Nvidia’s Surging Venture Capital Deployment

The discussions with Mercor reflect a broader surge in Nvidia's corporate venture activity. While the company primarily funds hardware customers and massive computing platforms—such as recent $2 billion investments into specialized cloud operators CoreWeave and Nebius, as well as optical component suppliers Lumentum and Coherent—it has dramatically scaled up capital deployment across the AI ecosystem.

In the fiscal quarter ending in April, Nvidia invested $18.6 billion into private enterprise investments and infrastructure funds—exceeding its total investment output for the entire previous year.

Mercor, valued at $10 billion in October, has raised about $520 million to date from prominent Silicon Valley backers, including Benchmark and Felicis, according to PitchBook data cited in the report.

NVDA Stock: Retail View 

Retail sentiment on Stocktwits was ‘bearish’ for the second consecutive week, along with ‘low’ message volumes. 

NVDA stock has gained nearly 16% year-to-date.

For updates and corrections, email newsroom[at]stocktwits[dot]com.<