Over the past two weeks, NVIDIA, a key facilitator and primary beneficiary of the artificial intelligence boom, has made investments in quantum computing. Richard Shannon, an investment analyst at Craig Hallum, remarked, “I’m a little surprised they haven’t done it before.”
These investments, made by NVIDIA’s venture capital division and other backers—collectively valuing quantum computing startups , , and at over $17 billion—signal a change in stance for NVIDIA’s CEO, Jensen Huang. In January, Huang asserted that practical quantum computers were 15 to 20 years away, causing a downturn for publicly traded quantum computing companies. He reversed this position in , in June, declaring that quantum computing had reached an “inflection point” and could “solve some interesting problems in the coming years.” An NVIDIA spokesperson declined to comment.
By positioning itself as the central provider of hardware essential for AI companies to run their models, NVIDIA has ascended to become the world’s most valuable company, currently valued at $4 trillion. NVIDIA designs GPUs (chips specialized for AI algorithms), develops CUDA (software enabling chip communication), and integrates these components into supercomputers, roughly the size of refrigerators, which AI companies are eagerly acquiring for their data centers.
Quantum computing is unlikely to assist NVIDIA’s AI clients. Pete Shadbolt, Chief Science Officer at PsiQuantum, one of the startups NVIDIA invested in, stated that “Quantum computing and AI are sort of diametrically opposed.” AI systems derive their power from learning patterns within vast datasets. Quantum computers, in contrast, “hate data, and they love precision,” Shadbolt explained.
Proponents envision quantum computing ushering in a new computing era. Instead of executing numerous simple calculations in parallel, as NVIDIA’s GPUs do, quantum computers could tackle a few, exceptionally valuable equations. Quantinuum achieves this using ions (charged atoms), PsiQuantum uses photons (light particles), and QuEra employs neutral atoms. These minuscule particles adhere to the peculiar laws of quantum mechanics: they can exist in multiple states simultaneously, allowing quantum computers to explore various paths of a complex calculation concurrently. This leads to solutions that contemporary “classical” computers cannot arrive at within any reasonable timeframe. One example is reducing the millions of years classical computers would take to break the encryption schemes underpinning much of the digital economy to , prompting banks to urgently seek “quantum-resistant” cryptography.
There are also beneficial applications. Hsin-Yuan Huang, a senior research scientist at Google Quantum AI, noted that quantum computers can simulate quantum mechanical systems, a feat impossible for classical computers. Given that quantum mechanics provides our most fundamental description of the physical world, the ability to simulate it could aid in designing novel drugs, materials, and chemical processes. Hsin-Yuan Huang stated, “That’s not going to be solvable with just many GPUs. It’s just inherently too hard.”
For instance, quantum computing may offer greener methods for producing ammonia, a process currently accounting for 2 percent of global energy consumption. PsiQuantum has partnered with Mercedes-Benz to investigate how quantum computers could simulate lithium-ion battery electrolytes—potentially accelerating electric vehicle battery design—and is collaborating with Boehringer Ingelheim, a pharmaceutical company, to comprehend an enzyme involved in the human body’s drug metabolism.
However, without sufficiently large quantum computers to implement these concepts, the tangible utility of quantum computers remains unproven, according to Jan Ole Ernst, a Ph.D. researcher in quantum information and computation at the University of Oxford. While it is possible this will change when larger quantum computers become available for application development, Ernst commented, “But I think that’s also kind of far-fetched, because there’s so much research going into applications, and we haven’t really found anything really clear-cut, apart from factorizing large numbers.”
The timeline for quantum computers to operate at a scale where these questions can be definitively answered is unclear. Timelines in quantum computing, much like in many technically challenging fields, often tend to stretch. PsiQuantum reports that it has commenced work at sites in Australia and Illinois and is testing the equipment designed to cool its chips to operational temperatures. The company expects to be running a quantum computer large enough to be useful by 2027, two years later than it stated in 2021.
Should quantum computers become functional, NVIDIA is likely to be at the forefront. Ernst remarked, “You’ll never be able to run a quantum computer without a ton of classical processing.” Classical computers are necessary for controlling quantum computers, performing error correction, and analyzing their outputs. PsiQuantum currently uses NVIDIA’s hardware to prepare its quantum computations and process their results. In 2022, the company introduced CUDA-Q, intended to facilitate communication between quantum and classical computers. This year, it established the NVIDIA Accelerated Quantum Computing Research Center in Boston, “dedicated to shortening the timeline to useful quantum computing.”
“I think it’s fair to say that they’re doing everything but the quantum computer,” Shadbolt observed.
This situation could evolve. Shannon believes NVIDIA’s recent investments will provide the company with “advance notice about which platforms are scaling better.” He concluded, “I believe, given enough time, it’s a virtual certainty that NVIDIA buys one or more quantum companies.”
