Intel is taking a longer view than most companies regarding what it will take–and how long it will take–to build commercially viable quantum computers, but a Gartner analyst said the company has at least a couple of things working to its advantage: It is taking a somewhat different approach to building qubits, and it is focusing on fundamentals, like improving qubit fidelity, rather than getting tied up in a race to be the one of the first to market.
Intel this week announced the limited release of its Tunnel Falls quantum processor, a 12-qubit chip that is being made available to the research community to evaluate and explore for their own needs, according to James Clarke, director of quantum hardware at Intel Labs. The announcement was Intel’s first major bit of quantum computing news since the company provided an update on its progress late last year.
Even though it is “not a commercial offering,” Clarke emphasized, the Tunnel Falls announcement is a pretty major milestone in an almost decade-long journey for Intel Labs to build a processor using silicon spin qubits. As for what defines this approach, Intel explained in its statement on the achievement, “In silicon spin qubits, information (the 0/1) is encoded in the spin (up/down) of a single electron. Each qubit device is essentially a single electron transistor, which allows Intel to fabricate it using a similar flow to that used in a standard complementary metal oxide semiconductor (CMOS) logic processing line.”
Ultimately, that means Intel will be able to leverage existing CMOS manufacturing processes to mature and produce its quantum processors at scale.
“Intel’s approach silicon spin qubit approach is uniquely differentiated and leverages Intel’s prowess in classical semiconductor leadership to begin addressing quantum challenges,” said Chirag Dekate, Ph.D., vice president and analyst, Quantum Technologies, AI Infrastructures and Supercomputing at Gartner, in an email to Fierce Electronics.
Dekate added, “The near-term quantum industry focus will continue to be on exploring diverse quantum system scaling approaches and most importantly building noisy intermediate scale quantum (NISQ) systems that enable development of quantum algorithms and derivative applications. Tunnel Falls is a key milestone in this broader technology evolution landscape. The Tunnel Falls 12 qubit silicon chip demonstrates the potential and viability of a silicon spin qubit based quantum approach.”
Intel is not the only company developing silicon spin qubit technology, but with its decades of experience in traditional transistor manufacturing, it could be in the best position to ensure eventual mass production of silicon spin qubit processors.
“We're building our quantum chip using all the tools that we have used to make our transistors. That's pretty unique,” Clarke said during a media and analyst briefing this week.
Clarke said the company plans to be a full-stack player in the quantum computing market, with plans to eventually build its own quantum computer. He acknowledged that other companies may be taking different paths to building commercially viable quantum computers–IBM and Google working on superconducting qubits, while Quantinuum and IonQ have approaches based on trapped ions–and that those companies may argue that the NISQ-era machines already have commercial viability. But he added Intel believes the best path forward is to leverage the work the semiconductor industry already has done. “The only hope, in my view, of hitting commercial relevance in the next 10 years is to ride the coattails of what the transistor and CMOS industry have been doing for the last 50 years,” he said.
Dekate added, “The quantum industry still needs to solve gnarly problems of scaling these systems while minimizing error rates and improving fidelity of operations to deliver reliable and resilient quantum computing systems. We are a few years removed from this. What is more important is to focus and improve fundamentals including coherence times, fidelity rates, error rates and usable qubits over time. Intel along with the rest of the quantum industry is focused on doing just that. Once these underlying semantics and fundamentals are proven, Intel’s silicon spin approach could be a promising way to build reliable larger scale quantum systems.”
During the Tunnel Falls briefing, Clarke declined to commit to a timeline for future Intel quantum milestones, again marking its approach as different and more cautious than many in the quantum space. While some quantum firms, particularly start-ups that are trying to find the quickest path to revenue, continue to hype milestones based on how many qubits their systems have, Dekate said Intel is taking a more measured approach that could prove prudent in a slow-developing market where innovators and researchers still have some thorny problems to solve before the technology fully matures.
“In an over-hyped marketplace that is about to go through a quantum winter of sorts, it is more important to under-promise and over-deliver,” Dekate wrote. “It is easy to suggest roadmaps, but delivering on the roadmap consistently is challenging. It is very important to not just focus on qubit volumes as a measure of technology maturity.”