OpenAI Considers $7 Trillion Round, Thinks It Can Double Revenue by 2025: Reports

OpenAI CEO Sam Altman is in talks with investors, including from the United Arab Emirates, to raise between $5 trillion to $7 trillion in funding. The goal, according to a report in The Wall Street Journal, is to increase the world’s chip manufacturing capacity and enhance AI capabilities.

The fundraising efforts are part of a broader strategy to address OpenAI’s growth constraints, particularly the scarcity of AI chips needed for training large language models like ChatGPT.

Altman’s proposal is said to include forming a partnership with investors, chip manufacturers, and power providers to finance the construction of chip foundries, which would then be operated by the chip manufacturers.

The Semiconductor Industry Association (SIA) recently forecast the global chip industry would see a 13% increase in sales in 2024, rising to $595.3 billion. That’s a huge leap from last year, when sales dropped by 8%.

Even with the growth constraint of chip production, OpenAI has hit some staggering milestones.

Two unnamed sources told Financial Times that OpenAI’s revenue has reached more than $2 billion on an annualized basis due to the success of ChatGPT. That makes the San Francisco startup one of the fastest-growing tech companies.

With a yearly revenue run rate of $2 billion as of December 2023, OpenAI thinks it can double that figure by 2025, fueled by strong business demand for its generative AI tools.

Despite being founded as a non-profit AI lab in 2015, OpenAI’s transformation into a commercial entity has seen rapid sales growth, even as it continues to face significant training and development costs.

Meanwhile, the company is now facing stiffer competition from one of its biggest rivals: Google.

Google just rebranded its Bard AI as Gemini and is now selling access to it through a premium subscription of $20 a month.

The rebrand and price point signals that Google intends to compete directly with OpenAI’s ChatGPT and its underlying models, such as GPT-4, which have gained substantial traction for their ability to generate code, text, images, video, and analyze information based on user prompts.