First AMD, Now Broadcom. How OpenAI Is Ruining NVIDIA's Party

and there's likely more strategic partners on the way....

In what has recently become an almost weekly development, OpenAI announced its latest strategic partnership, this time with Broadcom, on October 13 last, details here.

According to the OpenAI blog on the topic, this partnership will involve both companies co-developing systems that include accelerators and ethernet solutions from Broadcom, with the latter being responsible for deploying racks of AI accelerator and network systems targeted to start in the second half of 2026, and to complete by end of 2029.

This must be a most interesting development for NVIDIA CEO Jensen Huang. Having just signed his own strategic partnership with OpenAI on September 22 last, details here, he was apparently blindsided by the announcement of OpenAI’s further strategic partnership with AMD on October 6, details here. Why do I say blindsided? Well, because he was asked about the deal on multiple occasions, most notably on CNBC’s Squawk Box, details here. The following is a ChatGPT generated summary of those comments:

Jensen also appeared on a separate CNBC interview with Jim Cramer where this topic was raised and you can find the transcript of that discussion here. This also includes Jensen’s rationale for the deal he recently cut with Intel, where he noted the following:

Our our partnership with Intel is because I could imagine a future for the both of us where we could both win and and not one of us has to win, the other one to lose. And this partnership brings them into our ecosystem. And so they’re going to build a custom microprocessor for us. We’re going to also build a custom microprocessor for a new market.

While each of OpenAI’s three strategic partnerships (thus far) are fantastic news for said partners in their own right, the AMD and Broadcom deals are most unwelcome developments from NVIDIA’s perspective. Let’s dig in…