More than a decade ago, Intel led the charge to fund extreme ultraviolet lithography. They handed ASML over $4 billion in equity and R&D support. TSMC and Samsung followed with smaller checks. By 2015, ASML had working EUV machines and was ready for mass deployment.
Intel flinched.
They said the tools were too experimental, too expensive, too early. TSMC said yes.
By 2019, TSMC was using EUV in volume at seven nanometers while Intel was still tangled in delays on ten nanometers. That one decision handed process leadership to TSMC and put Intel years behind.
But ASML is only half the story.
Even if Intel had deployed EUV first, they faced another wall. The United States simply does not have enough skilled semiconductor labor to scale fabs at the pace needed. TSMC’s talent concentration in Taiwan gave it a structural advantage. Intel could not keep up.
Then came AMD.
While Intel stumbled, AMD moved fast. They adopted chiplet architecture early. Instead of one large monolithic chip, AMD began linking together multiple smaller chips. This improved manufacturing yields, cut costs, and boosted performance. Intel dismissed the approach. AMD ran with it.
The result: AMD gained massive market share across consumer and enterprise CPUs. Intel’s refusal to adapt in time has now halved their revenue in just five years.
Intel helped fund EUV. They ignored chiplets. They could not scale their fabs. And now they are trying to play catch-up in a game they once dominated.