Over the past two weeks, NVIDIA, power and most importantly the beneficiary AI Craze, bought quantum calculations. “I am a little surprised that they did not do this before,” says Richard Shannon, investment analyst Craig Hallum.
Investments from NVIDIA venture capital and other investors who collectively value quantum calculations of startups QuantinIN Moan And PSIQUANTUM In the amount of more than $ 17 billion – the presented change in the tone for the general director of NVIDIA, Jensen Huang. In January, Juang said These useful quantum computers took place from 15 to 20 years, sending publicly accessible quantum computing companies. He entered this MarchIN statement In June, quantum calculations reached the “Point of Peregun” and could “solve some interesting problems in the coming years.” The representative of NVIDIA refused to comment.
Positioning itself in the center of hardware necessary for AI companies to launch its models, NVIDIA has become the most valuable company in the world, worth $ 4 trillion. NVIDIA develops graphic processors (chips specializing in the management of AI algorithms), develops CUDA (which allows the chips to talk to each other), and the package is all into supercomputers the size of refrigerators that the company falls on each other to place data processing centers.
Quantum calculations are unlikely to help customers AI from NVIDIA. “Quantum calculations and AI are a kind of diametrically opposite,” said Pete Shadbolt, Chief Director for Science in Psique, one of the startups in which NVIDIA invested. Quantum computers, on the contrary, are “hatred, and they love accuracy,” said Shadbolt.
Supporters believe that quantum calculations can open a new computing paradigm. Instead of performing numerous simple calculations in parallel, as the NVIDIA graphic processors do, quantum computers can solve several extremely valuable equations. Quantinuum does this using ions (charged atoms), PSIquantum uses photons (particles of light), and the quera uses neutral atoms. These tiny particles obey the strange laws of quantum mechanics: they can be in several conditions at the same time, allowing quantum computers to explore multiple paths simultaneously complicated calculation, reaching solutions that current “classic” computers cannot reach any reasonable time. This includes a reduction in millions of years that will require classic computers to break the encryption schemes that underlie most of the digital economy before watchwhich has Sent Banks trying to “quantum-resistant” cryptography.
There are also useful applications. Quantum computers can model quantum mechanical systems that cannot be done using classic computers, said Hsin-Yun Huang, Senior Researcher Google Quantum AI. Quantum mechanics is our most fundamental description of the physical world, and the ability to simulate it can help develop new medicines, materials and chemical processes. “This will not be solved only with a large number of graphic processors. It’s just too difficult, ”said Xin-Juan Juang.
For example, quantum calculations can unlock More green methods for the production of ammonia, which currently account for 2 percent of global energy consumption. PSIQUANTUM collaborates with Mercedes-Benz to understand how quantum computers can simulate electrolytes of the lithium-ion battery, which can accelerate the design of the electric car battery and work with the pharmaceutical company Boehringer Ingelheim to understand the enzyme that is associated with the metabolism of the body in the human body.
But without fairly large quantum computers, to put these ideas in practice, a specific utility of quantum computers has yet to be demonstrated, said Jan Ole Ernst, Doctor of Philosophy. Researcher in the field of quantum information and calculations at the University of Oxford. Perhaps this will change when larger quantum computers will be available for application development. “But I think that it is also far -fetched, because there are so many research in the applications, and we have not found anything really clear, in addition to factorization of a large number.”
It is not clear how much time it will take for quantum computers to work on a scale, where these questions can be answered completely. Dates in quantum calculations, as in many technically complex areas, have the habit of stretching. PSIQUANTUM says that he began to work in areas in Australia and Illinois, and tests equipment that cools their chips to operating temperatures. The company claims to launch a quantum computer, large enough to be useful by 2027, in two years than it thought in 2021.
If and when quantum computers begin to work, NVIDIA will probably be in the center of action. “You can never start a quantum computer without a ton of classical processing,” said Ernst. Classic computers are necessary for managing quantum computers, performing errors and analysis of their testimony. PSIQUANTUM uses NVIDIA hardware to prepare its quantum calculations and process its outputs. In 2022 the company Runned Cuda-Q, which will allow quantum computers to communicate with classic computers. This year it is announced NVIDIA accelerated the research center of quantum calculations in Boston, “dedicated to the reduction of graphs for useful quantum calculations.”
“I think it’s fair to say that they do everything except a quantum computer,” said Shadbolt.
This may change. Recent NVIDIA investments will give the company “Pre -Notification of which platforms are scaled better,” Shannon said. “I believe that, given enough time, this is virtual confidence that Nvidia is buying one or more quantum companies.”
