OpenAI raises $6.6 billion at a $157 billion valuation

OpenAI has closed a $6.6 billion funding round led by Thrive Capital, with participation from Microsoft, Nvidia and other investors. The deal values the company at $157 billion, placing it among the most valuable private companies in the world and confirming that generative AI continues to attract capital on an exceptional scale.

The round comes less than two years after ChatGPT turned OpenAI into generative AI’s most recognizable face. But the money is not funding a popular product alone: it is supporting an industrial race in which developing more capable models requires data centers, specialized chips, electricity and increasingly expensive research teams.

Thrive leads round with OpenAI’s major partners

Thrive Capital, the investment firm founded by Joshua Kushner, is leading the deal. Microsoft and Nvidia are among the participants, alongside other funds and technology investors. The presence of both companies illustrates the position OpenAI occupies in the AI value chain.

Microsoft has been OpenAI’s most important strategic partner since 2019 and had already committed close to $13 billion to the company. In addition to providing capital, it supplies the Azure cloud infrastructure on which OpenAI trains and distributes much of its models. The partnership has enabled Microsoft to integrate OpenAI’s technology into products such as Copilot, Bing and its enterprise services.

Nvidia, for its part, dominates the market for graphics processing units, or GPUs—the chips that have become the essential raw material for training language models. Its investment has significance beyond the financial: leading model developers need vast quantities of its hardware, while Nvidia benefits as demand for computing power continues to grow.

From $86 billion to $157 billion in less than a year

The new valuation marks a notable jump from the approximately $86 billion OpenAI reached in a secondary share sale at the beginning of 2024. In that transaction, current and former employees were able to sell part of their stakes; this time, the company is receiving new capital to fund its operations and expansion.

The distinction matters. A valuation is not the same as revenue or profit: it represents the price investors are willing to pay for a stake in the company. In OpenAI’s case, that price rests on two expectations that are difficult to separate. The first is that its paid products—ChatGPT Plus, team subscriptions and developer APIs—can become a large-scale business. The second is that the company can maintain a sufficient technical lead to drive the next generation of models.

That bet comes at a high cost. Training systems such as GPT-4 requires enormous computing resources, and serving them to millions of users also consumes processing capacity. ChatGPT’s popularity does not eliminate the problem: every conversation, generated image or complex response uses infrastructure that must be paid for continuously.

The funding buys time and computing capacity

The $6.6 billion gives OpenAI room to invest in research, hire talent and expand its access to data centers and chips. It also strengthens the company’s ability to compete with well-funded rivals, including Anthropic, Google, Meta and xAI.

The round comes as OpenAI considers changes to its corporate structure. The company was founded with a nonprofit entity overseeing a limited-profit subsidiary, an unusual architecture intended to combine commercial development with a safety mission. The arrival of capital on this scale increases the pressure to find a structure that can attract funding without weakening the original entity’s control over the company’s goals.

For the market, the deal sets a demanding benchmark. OpenAI will have to show that it can turn enthusiasm for generative AI into recurring revenue capable of covering increasingly expensive infrastructure. For its customers, the immediate consequence will be a company with more resources to expand its models and services. The outstanding question is whether that financial advantage will be enough to keep competitors at a distance when they too have their own clouds, chips and models.