Patrick Gelsinger Says Intel Once Dismissed NVIDIA’s GPUs as Gaming Hardware, Before AI Changed Everything
Former Intel CEO Patrick Gelsinger has offered a revealing look back at some of the biggest turning points in modern computing, including Intel’s past confidence in CPUs, Apple’s move to Intel chips, NVIDIA’s rise from graphics hardware to AI dominance, and the global risks tied to Taiwan’s semiconductor industry.
In a wide-ranging discussion, Gelsinger reflected on Intel’s strategy before his return to the company, saying the chip giant had become too focused on financial returns and not enough on long-term technology investment. According to him, Intel spent around $100 billion on shareholder dividends in the years before he came back, money he suggested could have been used to strengthen the company’s manufacturing base.
Gelsinger said Intel had not built a new factory in roughly a decade when he returned. In his view, that reflected a major cultural shift inside the company, where business-focused leadership had taken priority over engineering-driven decision-making. He argued that a technologist would have understood the importance of investing in new fabrication plants and advanced EUV lithography machines, even when the short-term economics looked difficult.
The former Intel chief also spoke about Apple’s historic switch from PowerPC processors to Intel chips in 2005. He described Apple co-founder Steve Jobs as both brilliant and ruthless, crediting him with remarkable long-term vision.
Gelsinger recalled early conversations with Jobs about moving macOS to Intel’s x86 architecture. Intel believed it could help Apple with the complicated software transition, including compilers and operating system work. But Jobs surprised them by revealing that Apple had already been preparing for the move for years.
According to Gelsinger, Jobs said Apple had been working on the transition across the previous four operating system releases. That meant Apple had quietly built the technical foundation for a possible processor shift long before the company officially announced it. For Gelsinger, it was a powerful example of how great technology leaders prepare for major industry changes before the rest of the market sees them coming.
Gelsinger also compared NVIDIA CEO Jensen Huang to Steve Jobs, particularly in the way NVIDIA patiently developed technology that others initially underestimated.
During Intel’s peak dominance in CPUs, Gelsinger said the company did not take NVIDIA’s graphics processors seriously as a threat to mainstream computing. At the time, Intel viewed GPUs largely as products for gamers and graphics workloads, not as the foundation for future high-performance computing and artificial intelligence.
He said Intel used to scoff at NVIDIA’s hardware, seeing it as useful for graphics but not central to the future of computing. That view changed as NVIDIA built a powerful software ecosystem around its chips, especially with CUDA and parallel computing technologies.
Gelsinger explained that NVIDIA’s progress did not happen overnight. The company kept improving its hardware, software stack, programming tools, and developer ecosystem step by step. Over time, GPUs moved beyond gaming and became essential for supercomputing, scientific research, data centers, and AI training.
He pointed to high-performance computing researchers who began experimenting with graphics cards for demanding workloads as one of the key moments that helped push GPUs into new markets. What once looked like a niche gaming technology eventually became one of the most important computing platforms in the world.
That shift is now central to the modern AI boom. NVIDIA’s GPUs are widely used to train and run large artificial intelligence models, while companies across the semiconductor industry are racing to build faster, more efficient chips for AI workloads. Gelsinger’s comments underline how Intel, despite its historic leadership in CPUs, underestimated the long-term potential of GPU computing.
The conversation also turned to the semiconductor supply chain and the risks surrounding Taiwan, home to the world’s most important contract chip manufacturing industry. Gelsinger noted that Taiwan’s chip production capacity, especially through its leading foundries, plays a critical role in the global economy.
He warned that a blockade of Taiwan could cause severe disruption even without direct military conflict. Gelsinger highlighted Taiwan’s limited energy reserves, saying the island has less than three weeks of supply. If energy imports were blocked and semiconductor fabs lost power, the consequences could be enormous.
He explained that advanced chip factories cannot simply be switched off and restarted immediately. If a fab shuts down due to a power crisis, it may take months to bring operations back to normal. In his view, a major interruption to Taiwan’s semiconductor production could have an economic impact larger than the Great Depression.
Gelsinger said this risk shows why the world needs more resilient semiconductor supply chains. Governments and companies have increasingly pushed for chip manufacturing expansion in the United States, Europe, and other regions to reduce dependence on a single geographic area.
He also argued that the possibility of a blockade should not be dismissed as theoretical, noting that China has conducted repeated blockade-style military exercises around the Taiwan Strait in recent years.
Finally, Gelsinger shared an optimistic view of quantum computing, a field he believes could produce meaningful breakthroughs before 2030. He is currently closely connected with PsiQuantum, a company working to build practical quantum computing systems.
Gelsinger said quantum computing could solve problems that today’s classical computers cannot handle efficiently. He pointed to areas such as chemistry, biology, logistics, and optimization as early fields that could benefit from quantum advances.
He said researchers now understand more about building qubits, correcting quantum errors, and developing algorithms designed for quantum systems. In his view, the challenge has shifted from basic scientific discovery toward engineering at scale.
Gelsinger predicted that multiple industries could see meaningful quantum computing results before the end of the decade. If that timeline proves accurate, quantum machines could begin reshaping drug discovery, materials science, supply chains, energy research, and complex simulations far sooner than many expect.
His reflections offer a rare inside look at how quickly the technology landscape can change. Intel once dominated computing through CPUs, Apple quietly prepared for a major chip transition years in advance, NVIDIA transformed GPUs into the engine of AI, and quantum computing may be approaching its own breakthrough moment.
For Gelsinger, the lesson is clear: the future often belongs to companies that invest early, think technically, and prepare for shifts long before they become obvious.
