The creators of ChatGPT have a problem. They will get three times less money

The creators of ChatGPT have a problem. They will get three times less money

NVIDIA is currently the undisputed leader in the market of equipment for working with artificial intelligence. Their accelerators are both the most efficient and the most frequently purchased, which has allowed them to become the highest valued company in the world. Thus, each move by the Greens is treated as a kind of barometer of what the future of AI will look like.

OpenAI will get $30 billion instead of $100 billion

A few months ago, it was announced that the company led by Jensen Huang would invest a staggering $100 billion in OpenAI. However, the topic went quiet and no final agreement was publicly announced. Additionally, recently there have even been rumors about the possible withdrawal of NVIDIA. This caused a wave of speculation. Is the partnership starting to falter?

The Reuters news agency claims that the Greens still plan to fund ChatGPT developers, but to a lesser extent. Currently, we are talking about USD 30 billion, although it is possible that this will be only one of several stages of cooperation and the total will amount to USD 100 billion. After all, we are talking about the largest investment partnership of a company with Santa Clara.

For comparison, NVIDIA’s previous financial moves included a $20 billion licensing deal with Groq (not to be confused with xAI’s Grok) and the purchase of about 4% of Intel for $5 billion.

OpenAI is a key customer that was not entirely satisfied

Additionally, OpenAI is to be one of the first customers of the upcoming NVIDIA Vera Rubin platform, and in the coming years it plans to contract up to 10 GW of computing power in systems from the Greens. This volume means gigantic data centers and hundreds of thousands of accelerators.

At the same time, there was information that Sam Altman’s team was not completely satisfied with the delays in NVIDIA hardware. The company was to consider alternatives offered by entities opting for SRAM-based architecture, such as Cerebras. NVIDIA cannot afford to lose a key partner, because it could cause an avalanche. And concluding last year’s agreement – even partially – reduces this risk.

Similar Posts