The recent cooperation agreement between NVIDIA and OpenAI has become the focus of the market. There is actually a little-known story behind this deal-of-the-century cooperation. According to the "Wall Street Journal" report, the agreement between NVIDIA and OpenAI actually stalled for a time. It was not until CEO Huang Jensen put out the fire and re-engaged with Sam Altman that the deal almost fell through.
According to foreign media WCCFtech, NVIDIA recently reached a "huge agreement" with OpenAI and will spend US$100 billion in OpenAI, while the latter will obtain millions of AI chips with a total power consumption of 10GW.
However, according to the Wall Street Journal, the agreement between NVIDIA and OpenAI was in a deadlock. It was not until the outside world spread that OpenAI planned to integrate Google TPU chips to support AI computing that Huang Renxun quickly reconnected with Sam Altman in an attempt to save the deal.
It is reported that OpenAI signed its first cloud service contract with Google at that time, triggering speculation that it might start using Google's AI chips. The report quoted people familiar with the matter as saying that after the news was exposed, Huang immediately called Altman to inquire about the authenticity and expressed his willingness to restart negotiations. He believed that NVIDIA could assist OpenAI through direct investment to ensure the smooth progress of the cooperation between the two parties.
NVIDIA’s direct investment not only helps Altman obtain the computing resources it needs, but also ensures that NVIDIA chips remain at the core of the AI ecosystem.
In addition, it can be seen from Huang Renxun’s rapid change in attitude that ensuring cooperation with OpenAI was one of NVIDIA’s highest priorities at the time, because OpenAI is currently one of the most influential AI companies in the world. If it turns to Google’s self-developed AI chips, NVIDIA’s technology stack will face a real competitive threat. This investment of up to US$100 billion has essentially formed a "supplier lock-in" mechanism. For the foreseeable future, OpenAI will rely on NVIDIA's hardware, including the upcoming Vera Rubin system.
WCCFtech believes that through this transaction, Huang Renxun has become one of OpenAI’s largest computing resource providers, and it has also prevented Google ASIC chips from further penetrating the market and protected NVIDIA’s technical advantages.
NVIDIA’s CEO Jensen Huang ‘Scrambled’ To Rescue a Stalled Deal From Collapse With OpenAI After the AI Giant Considered Turning to Google TPU Chips Further reading: Arm joins the OCP board of directors and jointly promotes the development of AI data center standards with AMD and NVIDIA To accelerate the semiconductor design and test process, Siemens launches new solutions