4 Minutes
Former Intel CEO Pat Gelsinger has stirred the tech world by arguing that a quantum computing breakthrough could puncture the current AI frenzy — and that GPUs, the backbone of today’s AI stacks, may not survive the decade.
Why Gelsinger believes quantum will reshuffle AI’s deck
In a wide-ranging interview with the Financial Times, Gelsinger called quantum computing part of a new computing “trinity” alongside classical and AI systems. Drawing on his work with venture firm Playground Global and direct exposure to quantum research, he suggested qubits could render today’s GPU-centric approach obsolete much faster than many expect.
Gelsinger’s take is provocative: rather than a slow evolution, he sees the potential for a relatively rapid shift if a meaningful quantum milestone arrives. That change, he warned, could deflate the AI investment bubble that’s been inflating around expensive GPU compute and model scaling — especially where commercial valuations depend on those chips remaining unrivaled.
Two-year shock or a decades-long drift? The debate heats up
Not everyone agrees on timing. Nvidia CEO Jensen Huang has previously said it will take decades for quantum to go mainstream. Gelsinger, by contrast, suggested the timeline could be dramatically shorter. Whether it’s two years or twenty, both views agree on one thing: the next decade will be decisive for computing.
Why timing matters: AI development today leans heavily on GPUs for model training and inference. If quantum systems begin to provide genuine advantages for certain workloads, capital will reallocate quickly — and companies built around GPU ecosystems may need to pivot or face rapid market pressure.
An industry echo: Microsoft, OpenAI and old software plays
Gelsinger also drew a historical parallel, comparing the Microsoft–OpenAI partnership to Bill Gates’ alliance with IBM in the 1990s. He framed OpenAI as a distribution partner leveraging Microsoft’s massive compute capacity — a reminder that strategic cloud and compute deals shape which technologies win commercially, not just what performs best in the lab.
Intel’s internal story: discipline, delays and the 18A saga
The interview didn’t focus solely on quantum. Gelsinger candidly reflected on his time at Intel, describing a period where what he called “basic disciplines” had been lost. He told FT reporters that in the five years before his return, not a single Intel product shipped on schedule — a decay he said ran deeper than he’d realized.
Among the casualties was Intel’s ambitious 18A node. Gelsinger said that although he had promised leadership a five-year timetable to deliver 18A, organizational problems and delays meant the company missed its internal targets. After he left, the new CEO chose to discontinue the project within that timeframe, underscoring how technical roadmaps and leadership changes can rapidly reshape a chipmaker’s fate.
What this means for technologists and investors
Whether you’re a researcher, engineer, or investor, Gelsinger’s comments are a reminder to watch multiple technological axes at once. AI today is deeply tied to GPU economics and data-center scale. Quantum raises the possibility that some of the most computationally intense problems could be reimagined on a different substrate, shifting winners and losers across hardware, cloud services, and AI platform providers.
Imagine a world where certain optimization problems or simulations are more naturally solved by quantum devices — that would change software stacks, procurement decisions, and strategic bets. For now, the debate over timing continues. But Gelsinger’s voice, backed by his industry credentials and recent hands-on exposure to quantum startups, adds weight to the argument that we should be ready for disruptive surprises.
Source: wccftech
Leave a Comment