AMD has achieved a significant milestone in the semiconductor industry by overtaking Intel as the first customer for TSMC’s advanced 2nm manufacturing process, surpassing previous leader Apple and cementing its position as a front-runner in next-generation chip fabrication. Historically, Apple has been the earliest adopter of new TSMC nodes, but due to a slowing of Moore’s Law and a shift in AMD’s strategic urgency, the iPhone 18 will not debut Apple’s first 2nm chip until late 2026, leaving AMD to reach the market earlier than its rivals. This bold commitment from AMD is seen as a pre-emptive strike against Intel’s upcoming 18A server chips, a technology class intended to rescue Intel’s waning market position. Market data now shows AMD’s server revenue surpassing Intel’s, with Intel’s market capitalization dropping to just 60% of AMD’s value.
However, despite these manufacturing triumphs and marked success on the server processor front, AMD continues to fall further behind Nvidia in the discrete and Artificial Intelligence-focused GPU markets. Experts attribute this disparity not to a lack of technological capability alone, but rather to a decade-long underinvestment loop. Around 2015, AMD had to prioritize CPUs for its own survival, redirecting resources away from GPU innovation. The subsequent impact was a drop in GPU performance competitiveness, limited product variety, and declining discrete GPU production, now reportedly just one-third of its output from a decade ago. Meanwhile, Nvidia gained an enduring lead with its Maxwell generation, achieving a significant leap in performance per watt, setting a trajectory that AMD has not since matched.
In addition to R&D investment, industry commentary highlights how AMD’s leverage over Intel has been aided by its ability to benefit from the existing x86 and graphics software ecosystems. Conversely, Nvidia’s dominant position in Artificial Intelligence stems from its early and deep investments in specialized software and frameworks required for AI workloads. As the sector shifts, AMD faces challenges building a competitive software ecosystem for Artificial Intelligence, especially as integrated graphics and changing market dynamics reduce demand for external GPUs. Having previously relied on the broader Windows and Linux developer base, AMD’s challenge in AI is structural, not just technical, as replicating Nvidia’s software-driven network effects and dominance will be a formidable task.
