So, wait, now TSMC is supposedly pitching a joint venture with Nvidia, AMD and Broadcom to run Intel’s ailing chip fabs?

In the on-again, off-again saga of TSMC’s supposed designs on Intel’s chip factories, there’s a new twist. Reuters reports that TSMC has pitched Nvidia, AMD and Broadcom with the idea of a joint venture to take control of Intel’s struggling fabs.

Immediately, we should point out that the story just a few weeks ago was that Intel’s fabs would be spun off into a joint venture between Intel and TSMC but run by the latter. Then a few days later, that rumour was dismissed by some industry observers, while others backed it up and added Broadcom into the mix.

Long story short, what’s actually going on, if anything, is anyone’s guess. But Intel is certainly in trouble, what with making a $18.8 billion loss in 2024 and ditching the CEO that was supposed to be its saviour. So, it’s hardly surprising there are rumours regarding its future.

Anywho, back to this latest yarn. According to Reuters, TSMC is proposing to Nvidia, AMD and Broadcom that it would run the fabs but that they would be owned by the joint venture with TSMC’s stake limited to no more than 50%.

That latter detail reflects the claimed fact that the Trump administration would not support any deal that saw majority ownership of Intel go overseas. Reuters also claims that it may well ultimately be the Trump administration that’s behind also these machinations.

“U.S. President Donald Trump’s administration requested TSMC, the world’s leading contract chipmaker, assist in turning around the troubled U.S. industrial icon, the sources said on condition of anonymity because the talks are not public,” Reuters claims.

Unsurprisingly, Reuters says that none of the alleged parties deigned to comment on the story. But if the story is to be believed, negotiations are quite advanced. “Intel board members have backed a deal and held negotiations with TSMC, while some executives are firmly opposed, according to two sources,” Reuters says.

It’s ultimately all a bit of a mystery. However, what seems to be likely is that much if not all of this of this hinges on Intel’s 18A node. If we had to guess, we’d say that if Intel can deliver on its claims that 18A will be up and running, banging out chips later this year and competitive with TSMC’s N3 and upcoming N2 nodes, then the preference at the company will be to hang on to its fabs.

On the other hand, if 18A suffers further delays and isn’t looking like a clear winner, the company may begin to run out of time and money. For now, it may make sense to explore the option of divesting the fabs on some level while still intending to go it alone should 18A deliver.

All told, it feels like, and has indeed been years since Pat Gelsinger rolled out his Intel 2.0 rescue plan. We’re still waiting to find out whether it’s worked. In the meantime, Intel’s fate seems to be balanced on a knife-edge. At least it’s not boring!