Gelsinger: Intel Missed the GPU Wave, a Taiwan Blockade Would Devastate the Global Economy
Miles Bennett
Former Intel CEO Pat Gelsinger told the RAISE Summit in Paris that Intel dismissed GPUs as gaming tools and missed the compute wave, warning that a Taiwan blockade could deal a blow to the global economy worse than the Great Depression.
Why did Intel dismiss GPUs?
Gelsinger said Intel, at the peak of its CPU dominance, treated Nvidia's GPUs as graphics tools for gamers and never imagined them doing general-purpose compute.
This means → Intel did not lack the capability to build GPUs — it made a strategic misjudgment, ignoring the opportunity when it had the most resources to compete.
He credited Nvidia founder Jensen Huang with refining the CUDA software stack — a programming platform that lets developers run general compute on GPUs — and the SIMT architecture, calling Huang's foresight on par with Steve Jobs's.
He also noted that Japanese high-performance-computing researchers were among the first to bring GPUs into general workloads, helping spark the GPU market's explosive growth.
No new fabs for a decade — what did management do?
Gelsinger blamed Intel's decline on pre-return management: non-technical executives distributed over $100 billion in dividends to shareholders in five to six years.
In plain terms = the company handed shareholders money that should have gone into building fabs and buying equipment.
The direct result: Intel built no new fabs for a decade and failed to procure EUV lithography tools — machines that etch chip circuits with extreme-ultraviolet light, essential for the most advanced process nodes.
This reflects what can happen when a technology company lets non-technical leaders control capital allocation — it can lose manufacturing capability at the most critical window.
How serious is the Taiwan blockade risk?
Gelsinger noted that TSMC's wafer capacity once reached five times Intel's.
He cited reports showing Taiwan's on-island energy reserves, including natural gas, last fewer than three weeks; if a blockade forced power rationing, restarting fabs would take roughly 90 days.
This means → a Taiwan capacity shutdown would create a months-long vacuum in global chip supply, with an economic impact that could exceed the Great Depression.
How did Taiwan's government respond?
Taiwan's Ministry of Economic Affairs issued a statement on July 19, 2026, pushing back. It stressed that Taiwan's semiconductor supply chain has weathered multiple major earthquakes, COVID-era logistics disruptions, and several international energy crises without a single full interruption.
The ministry noted that Taiwan's highly integrated industrial clusters and robust infrastructure underpin its critical role in the global semiconductor and AI supply chain.
It added that during the earlier global chip shortage, Taiwanese firms accelerated capacity expansion and boosted supply, and urged caution toward what it called "overly speculative" claims.
Can quantum computing deliver real results before 2030?
Gelsinger has close ties to quantum-computing firm PsiQuantum. He said quantum computing has solved its core challenges in qubit error correction — technology that lets a quantum computer automatically fix its own calculation errors — and at the algorithm level.
He sees the field entering an engineering scale-up phase and expects meaningful quantum-advantage results in chemistry, biology, and logistics before 2030.
Put simply = in his view, quantum computing has moved past the "can it work?" stage; the question now is "how to scale it up."
Content is for reference only, not financial advice.