There are multiple reports that the U.S. Government is pushing AMD and Intel to merge. Given the difference in financial performance and valuations, surprisingly, AMD would be the company doing the acquisition.
AMD this year passed Intel (and Netflix) in brand recognition, and has been moving towards a stronger position than Intel in the critical server market for much of this year. AMD is doing impressively well under Lisa Su’s leadership, and it has become the new industry darling after NVIDIA.
Both companies use the same core CPU technology, x86, and while they are using different GPU technology, AMD’s has been by far the more successful, even though both lag NVIDIA. Together, they should be more effective against NVIDIA than as separate entities.
Let’s look at the viability of an AMD/Intel merger.
A Little History
Back when this current PC and server markets were conceived, IBM was the kingmaker. Choosing Intel for processors and Microsoft for the OS, the IBM PC, which was brought out to combat Apple’s surprising early success, became the industry standard. IBM didn’t like the idea of a single source for hardware, and Intel licensed x86 to AMD to create that second source.
So, in a way, AMD and Intel were IBM constructs because their future direction was initially set by IBM (which also licensed x86 technology for a while). IBM started floundering towards the end of the 90s, so companies like Dell stepped into the vacuum. This set up IBM to eventually exit the x86 market for PC and servers, and turned Lenovo into what IBM could have been which was competitive in both categories, globally.
Throughout the 90s, Intel dominated with marketing, IDF (Intel Developer Forum), and impressively strong execution. In the early 2000s, once Andy Grove retired and Dennis Carter, Intel’s CMO, retired, Intel started to lose its way. So AMD surprised Intel by focusing the company back on its core markets but not on restoring the funding for marketing. Intel survived largely because AMD screwed up, first by over-promising and under-delivering on a server part, and then by attempting to pivot to smartphones.
Fortunately, that pivot didn’t last, and AMD returned to its roots, but Intel made a hard attempt, and failed to pivot toward smartphones, took its eye off of its core market, and gave newly focused AMD a huge opportunity. IBM-trained Lisa Su took over AMD’s operations with Mark Papermaster, also IBM-trained, and AMD’s current success is due to AMD’s excellent execution, along with Intel’s collapsing marketing, IDF, and making a failed, underfunded, under executed pivot to smartphones.
Today, Intel still owns x86, but AMD is sharply out executing it. Together, the total could be greater than the sum of the parts for these two companies on the design and tech side.
FAB and Foundry
One of Intel’s greatest assets is its FAB business. This is what makes Intel too big to fail since it represents a critical part of national defense. Intel’s foundry business, which is being built, adds to this importance. However, it takes a long time to spin up FABS and foundries, and this effort has clearly become a distraction for Intel’s leadership as they attempt to turn the company around from the mistakes made by their predecessors.
Virtually every company that Intel competes with on CPUs and GPUs, including AMD, is fabless, which allows those competitors to be more agile and more focused on design. This means that while the foundry and FAB parts of Intel remain extremely important to the U.S. government and the West in general, they are a drag on Intel’s ability to compete.
If the FAB and foundry divisions were spun out of intel as independent entities (or as a single entity which would be more likely) while AMD and Intel were allowed to merge, the result would be a far more focused x86 platform with a vastly stronger GPU and AI component while eliminating the waste and market confusion of competing AMD and Intel products at a time when both face credible challenges from ARM.
The result would be a stronger ARM competitor (from a national security standpoint), a better and more focused processor manufacturing capability for the West (and particularly the U.S.), and a far more focused x86 company that could better compete with ARM and RISC.
Wrapping Up:
The merger of AMD and Intel makes more sense than a merger between Intel and Qualcomm mostly because AMD and Intel use the same core technology. Qualcomm is having problems with ARM that makes it not only a less capable buyer for Intel, but regulatory bodies, particularly in Europe, might take issue with this merger because that is where ARM resides. Intel’s resources added to Qualcomm would significantly strengthen Qualcomm’s ability to both fight ARM or switch from it. And Qualcomm leaving ARM might damn that company and technology, particularly with Huawei advancing so strongly in the smartphone design and microprocessor space.
So, from a U.S. perspective, and I expect from both an AMD and Intel perspective especially given their new Ecosystem advisory group, it makes sense for AMD and Intel to merge and spin out manufacturing (FABs and foundries) to create a more competitive company that can better focus on the emerging Huawei threat.