OpenAI President Brockman: By 2026, computational spending will reach $50 billion, and investment in AI infrastructure will continue to soar.

OpenAI co-founder and president Greg Brockman testified in court on Tuesday that the company is expected to spend up to $50 billion on computing power by 2026 to support the development and operation of its artificial intelligence software. Brockman revealed that as OpenAI develops more advanced AI models and expands its services to a wider user base, its computing costs have skyrocketed from around $30 million in 2017 to hundreds of billions of dollars this year. These statements were made on the second day of his highly anticipated court battle with Elon Musk. Since the release of ChatGPT at the end of 2022, OpenAI has sparked the current wave of generative AI. Since then, the company has aggressively acquired more chips and data centers for AI, driving a broader infrastructure investment boom. As a startup that has yet to turn a profit, OpenAI is currently at the heart of a complex investment and transaction network involving multiple leading cloud computing providers and chip manufacturers. OpenAI has previously stated that it has committed to investing over $1.4 trillion in AI infrastructure in the coming years. In February of this year, the company informed investors that it plans to spend a total of around $600 billion by 2030. Last month, reports indicated that OpenAI was falling short of its internal targets for new user acquisition and sales, intensifying concerns within the company about whether it can sustain its high AI infrastructure spending. In 2026, OpenAI failed to meet sales targets for several months in a row, primarily due to competitor Anthropic expanding its share in the programming and enterprise markets. This news has once again raised concerns about whether tech companies' massive investments in artificial intelligence (AI) infrastructure will yield reasonable returns, leading to a series of stock declines associated with OpenAI.