Although no longer in the chair of Intel, Pat Gelsinger has no intention of remaining silent on the future of the world’s technological architecture. In a recent interview with the Financial Times, the former head of the Santa Clara giant presented a thesis that stands in stark opposition to the current market consensus. According to Gelsinger, the current ‘hype’ around artificial intelligence, although it will last a few more years, is merely a prelude to a much more profound change: the arrival of the era of quantum computing.
While Nvidia CEO Jensen Huang is building an empire based on GPU dominance and predicts that quantum technology will not impact the market for the next two decades, Gelsinger drastically shortens this horizon. In his view, quantum computers will enter the mainstream in just two years. This is a bold prediction, suggesting that the current hegemony of GPU accelerators could end before the end of this decade. Gelsinger visualises the digital future as a ‘holy trinity’ of coexisting technologies: classical processors, AI accelerators and quantum units. In this scenario, the role of GPUs, currently crucial for training large language models, would be significantly marginalised in favour of cubit-based chips.
Gelsinger’s perspective seems to be evolving as he moves away from Intel’s corporate structures towards the startup ecosystem. Working more closely with smaller, more agile quantum players, he sees a pace of innovation that may be invisible from the perspective of the incumbent silicon giants. He is also critical of the past in his assessment. He acknowledges that Intel under his helm struggled with a “loss of fundamental discipline”, resulting in delays in key processes such as 18A lithography. This experience of struggling with the limitations of classic silicon may have reinforced his belief that Moore’s Law in the traditional sense is depleting faster than the industry is willing to admit.
