FPGA Dominance & StickinessWide moat

The heart of the embedded franchise is AMD's leadership in field-programmable gate arrays, the business it acquired as Xilinx, which for decades has been one of the two dominant FPGA makers alongside Intel's Altera. FPGAs are chips whose logic is reconfigurable after manufacture, prized in applications that need custom hardware behavior without the cost and rigidity of a custom-designed chip — and they are among the stickiest products in all of semiconductors. A customer designing an FPGA into a piece of industrial, aerospace, communications, or medical equipment invests years architecting the system around it and writing the specialized low-level programming the part requires.

That investment creates a switching cost so high it approaches lock-in: moving to a competitor's FPGA would mean redesigning the hardware and rewriting the programming, re-validating and re-certifying the entire product — prohibitively expensive for equipment with lifecycles that run a decade or more. So an FPGA design win throws off high-margin revenue for many years with little competitive threat, and the installed base compounds as new designs layer on top of old ones. This is a genuine structural moat — switching costs and long lifecycles — of exactly the kind the rest of AMD lacks. The check is that FPGAs face pressure at the margins from custom ASICs (cheaper at high volume) and from GPUs and specialized accelerators encroaching on some workloads, and the market is mature rather than fast-growing. But within its domain the FPGA franchise is durable, defensible, and highly profitable — the widest-moat business AMD owns, bought with Xilinx for ~$49 billion¹.