AI labs around the world are gearing up to receive a quantum jolt that can slingshot artificial intelligence (AI) to new heights. But it may take longer than you think.
A radically new paradigm, quantum computing that has quietly slipped under the radar – largely overshadowed by AI’s halo – has been incubating for years, and is on its way to unlock massive computation power far beyond the scope of classical computers.
Word is, quantum computing’s parallel processing capabilities can be the way to create robust AI models capable of more complex problem-solving at much faster rates. But the technology in its current state is neither highly reliable, nor practically useful.
“It’s not here today, but the interesting thing is the software stack that needs to be built for quantum is getting built actively,” said an optimistic Satya Nandella at the World Economic Forum in 2023. “In fact, if anything, the entire fact that you can simulate quantum algorithms on classical [computer] is one place where we can already start benefiting.”
Powerhouse companies and a growing body of well-funded startups have already thrown their weight behind the research and development performing breakthrough experiments harnessing quantum attributes – and China, like the U.S., is ready to stake its claim in the race.
In 2023, a group of researchers in China announced a breakthrough in quantum computing. The team led by the highly lauded quantum physicist, Pan Jianwei, developed a prototype of a photonic quantum computer that they claimed is “180 million times faster on AI-related tasks” than the most powerful supercomputer in the world.
With 255 detected photons that greatly elevates computing speed and complexity, makers said that the Jiuzhang computer needs less than a second to solve a task that the fastest classical supercomputer takes five years to finish.
The supercomputer demonstrates blazing speeds outperforming other supercomputers by over 10 quadrillion times when it comes to Gaussian boson sampling (GSB) problems, according to the team.
In another breakthrough recently, China has developed a superconducting quantum computer embedded with China’s largest quantum chip – the 504-Qubit Xiaohong. The quantum computer dubbed Tianyan-504 unveiled in Dec’ 24, is so far, the most advanced quantum system coming out of China.
Meanwhile, US-based Google has had its own breakthrough with the Willow chip, which made splashes for its computational performance and error correction capabilities which Google said opens avenues for large-scale quantum computer.
The cornerstone of scalable quantum computing is fault-tolerant systems that can self-correct their errors making the results reliable.
Microsoft, in collaboration with Atom Computing, is working on a project to build a high-fidelity “reliable quantum machine”, which is made with 24 entangled logical qubits from neural atoms. The systems are slated to be commercially available later this year.
“The quality of the qubits is important, but so is their architecture, connectivity, and networkability,” said Dr. Robert Sutor, author of Dancing with Qubits, and VP and practice lead at the Futurum Group. “My recommendation is to applaud innovation as we move toward Practical Quantum Advantage – the point where quantum and classical systems can significantly outperform classical systems alone on problems important to society and industry.”
In November, IBM unveiled the R2 version of Quantum Heron processor, another milestone in its ambitious Quantum Development Roadmap, which the company said is 50x faster than the previous version, and uses Qiskit quantum software that can execute some circuits up to 5000 two-Qubit gate operations, paving the path for what it calls “quantum utility” or quantum as a tool.
“Comparing the number of qubits between systems is easy, but how many do we really need? Dr. Sutor said. “I think 100,000 is a reasonable number, and it may be 10 to 100 times that. So getting excited about 100 or 1000 as a sign of progress is fine, but claiming that “quantum computing is here” is gross hyperbole.”
Cisco, led by Outshift, is also pouring in R&D investment with the goal of building a quantum internet.
“It’s really exciting,” Tim Szigeti, distinguished technical marketing engineer said at the Tech Field Day Extra at Cisco Live US 2024 in June. “If you think of where quantum is, what regular computing was in the ‘60s before people even thought of an internet. They weren’t really thinking of networking but we’re already thinking that because we’ve seen the evolution…We expect the development to go that much faster for our prior experience and knowledge with digital compute and networking.”
A palpable surge in interest has erupted around quantum computers as new innovations are splashed across headlines. Especially with potential applicability in AI, cybersecurity, military, life sciences and drug discovery, and the computers’ ability to outperform classical computers by a wide mark, quantum computing is believed to be the true partner-in-crime for GenAI, where it will supercharge AI’s ability to chew through stacks of data.
However, despite the breakthroughs and the optimism, quantum-AI-enhanced solutions are not a reality today. Forrester predicts in a 2024 report that despite significant strides of the past year, quantum computing continues to be an experimental technology that is yet to demonstrate any useful work that surpasses conventional computers.
“Quantum computing has been in the news a lot lately, but there are significant public misunderstandings of the power of the current systems,” Dr. Sutor said. “Do not get caught up in the irrational exuberance regarding where we are now, or quantum company stock prices,” he cautioned.
