We are frequently asked some version of “Will someone acquire Intel?”. At this point, we think it is highly unlikely, but these are highly unpredictable times. Setting aside all the principal considerations – like money, strategy and regulatory approval – there are a few other hurdles. Chief among those are Intel’s license for x86.
Apparently, when Intel and AMD resolved their multi-decade dispute over x86, the cross-license agreement held provisions about transferability. If either company is acquired, the other has the right to cancel the license, effectively blocking the deal. When this was signed, we have to think everyone involved assumed it would be AMD getting acquired, but the provision seems to be mutual, and times change.
So just for fun, let’s think through how that would play out. Right now, the most likely potential buyer of Intel is Broadcom, what would happen if they approached AMD about this license?
AMD would have a few options. Their default position would likely be to just refuse and block the deal. Broadcom is a fearsome competitor, and even if they do not really go directly against Broadcom, a rejuvenated Intel underneath Broadcom would be a fearsome threat to AMD.
That being said, a failed Intel would also be a major problem for AMD, if for no other reason than the two are now linked at the hip for x86 which AMD definitely needs to remain healthy. All of which is to say there is room for negotiation.
What would AMD ask for to get grant that approval? Obviously they would take some cash. But there is only so much that Broadcom would be willing to pay before the Intel deal becomes unappealing. That probably equates to a few billion dollars. Moreover, there are things that AMD could get which would be worth far more than the cash.
For starters, they would probably ask for Broadcom for some help fighting Nvidia. This is an area of common interest between the two companies. And we imagine there are multiple ways in which the two would benefit from a deeper partnership. This could include more investment from Broadcom into Ultra Ethernet and AMD’s other networking initiatives. Or getting Broadcom to design an entire networking stack that only integrates with AMD’s Mi300 series of AI accelerators. Going further, AMD could ask Broadcom to incentivize its many AI ASIC customers to adapt AMD-friendly networking interfaces. We imagine there are plenty of novel ways that Broadcom could promote AMD’s interests in their combined fight against Nvidia.
Or AMD could ask Broadcom to throw a bunch of business towards ZT Systems. After AMD acquires ZT, they plan to sell off the manufacturing side of ZT. It sure would help if that business had a healthy backlog of Broadcom orders.
AMD would also want to make sure that Broadcom continued to spin off Altera, which competes with AMD’s Xilinx. Maybe go further, and get Broadcom to commit to using more Xilinx in future products and reference designs.
The point in all of these is that they would cost Broadcom very little and at the same time could prove very valuable to AMD strategically.
As much as we are conducting this thought experiment just for kicks, it does highlight one more hurdle for anyone buying Intel. There is almost no one else who could provide AMD with any comparably valuable strategic offsets. For instance, a private equity buyer for Intel, would only have cash to offer. The deal is unlikely, but if it happened, it would shake up many corners of the industry.
Leave a Reply