Hype squared or breakthrough: Is it time to capitalize on quantum AI

Quantum computers have the potential to revolutionize medicine, materials science, the management of complex systems like cellular or power grids, and last but not least, cryptography. According to Quantum Insider, venture capitalists invested 50% more in the technology in the second quarter of 2025 than in the same period a year earlier. Perhaps in part because many companies have announced that they are beginning to explore the use of quantum computers in the super-hot topic of artificial intelligence. What happens when you combine these two technologies? Is this really a future incredible breakthrough or just hype squared?

As I've written before, quantum computers are a fundamentally new type of computing device that use very strange laws of quantum mechanics to compute.

A classic computer chip consists of a huge number of microtransistors that can take two values, "on" (1) or "off" (0), an elementary unit of information or bit. A quantum computer uses atoms or subatomic particles (such as photons, for example) that can create quantum mechanical bonds and form what are called qubits. Thanks to the quantum phenomenon of superposition, a qubit can have not two values, but many in the range from 0 to 1. It is often compared to a coin tossed into the air, which, before it falls, is neither "heads" nor "tails", but only some probability of occurrence of one or another result. And as the number of qubits in the system increases, the amount of information contained in them grows exponentially.

"A classical computer would take millions of years to decompose a 2048-bit number into prime multipliers. Cubits can perform this calculation in just a few minutes," Microsoft explains.

But there is a problem - the matter is that qubits are extremely fragile, they easily lose quantum bonds, and this leads to errors. In fact, the best minds of mankind are now working on reducing the number of errors and the possibility to combine more and more qubits into stable systems.

IBM unveiled its 1,121-cubit Condor quantum chip in 2023, Atom Computing a year later its AC1000 system, which is claimed to have"more than 1,200 qubits."

But it's not just about quantity. Reducing errors is a priority, which is why, for example, Google claims that the 105-qubit Willow chip developed at the very end of 2024 is "best-in-class."

Microsoft responded in January 2025 with the Majorana chip, which has only 8 special "topological" qubits, but the company claims they are better error-proofed and better controlled than ever before, and so the way is open to integrating up to 1 million qubits on a single chip.

Reading about quantum computers, it's hard not to notice that there are at least two approaches, and they are seriously different. Large companies such as Google, Microsoft and IBM talk about them mostly in the future tense: IBM's roadmap, for example, goes beyond 2033.

But there are quite a few startups that claim to already be using quantum computers and are successfully selling commercial solutions based on quantum AI to customers. The main thing is that investors seem to believe them.

In 2024, investors will invest nearly $2 billion in quantum technology startups around the world, a 50% increase from 2023, McKinsey wrote in a report.

The trend has continued this year, with Quantum Insider reporting that private equity investment in quantum startups rose to $1.16 billion in the second quarter, a plus 50% year-over-year increase.

However, these are mere trifles compared to the financing of another hot topic - artificial intelligence. Crunchbase estimates that $114 billion was invested in it in the first quarter of this year and another $82 billion in the second quarter.

If quantum computers are much faster than conventional computers, and if AI, as OpenAI CEO Sam Altman argues, is all the better the more computing power it has, the idea of trying to combine the advantages of these developments seems logical.

"I don't think there are two other topics in the tech world that are discussed more often than quantum computing and AI. Whether it's research teams in large corporations, scientists turned startup entrepreneurs, public and private investors, CEOs, educated commentators, experienced industry analysts, stock market advertisers and short-term traders, or newly minted social media experts, these areas have captured the imagination of many people around the world. Sure, we can look at them individually, but what is their potential when these capabilities are combined, or at least used next door?" - asks Bob Sutor, former IBM employee and author of Dancing with Qubits.

Startups are responsive to the conjuncture.

SandboxAQ started as a promising AI-based quantum computing division of Alphabet, Google's parent company. In 2022, the company spun off into a separate startup with former Google CEO Eric Schmidt as chairman. "As the world enters the third quantum revolution, AI + Quantum software will solve important business and scientific challenges," the company wrote.

In December 2024, Israeli startup Quantum Machines and US quantum computer company Rigetti revealed the successful calibration of a quantum computer using AI.

Quantinuum in February this year announced a "breakthrough in generative quantum artificial intelligence with huge commercial potential". It is a Gen QAI platform for training generative AI based on data created by a quantum computer. The company hopes that the technology can be applied to applications ranging from developing new drugs to accurately predicting the behavior of financial markets and optimizing global logistics in real time.

From very recent examples, D-Wave in early August released "a set of offerings designed to help developers explore and develop innovations in quantum artificial intelligence."

It seems to be working, at least in terms of attracting investment. SandboxAQ received another $150 million in April (including from Nvidia and BNP Paribas) at a total valuation of $5.75 billion.

Quantinuum, according to Bloomberg, is in talks for another round "in the hundreds of millions of dollars" at a valuation of $10 billion.

Rigetti and D-Wave are publicly traded companies, with their shares up 1, 742% and 1, 462%, respectively, for the year. In addition, D-Wave announced a successful $400 million stock sale on July 1.

Everything, apparently, is going fine. But let's look at the second quarter results of the same D-Wave, a company that describes itself as nothing less than "a leader in quantum computing systems, software and services, and the world's first commercial supplier of quantum computers".

Revenue was just $3.1 million and net loss was $167.3 million.

In addition, a report on D-Wave from Kerrisdale Capital came out in April. This company are shorts who bet on stocks falling, but their plus is that they explain in detail every time why they decided to make such a bet.

According to their data, D-Wave had a share price in April of 57 times expected 2026 revenue, which is simply "an absurd valuation for a company that has never earned more than $9 million in annual recurring revenue, has no clear path to profitability and faces a stagnant customer base as its approach is increasingly perceived as a commercial dead end."

D-Wave applies the so-called "hybrid model", combining quantum and classical computing, but does not disclose their ratio to clients. According to former D-Wave employees interviewed by Kerrisdale Capital, the company's quantum technology does not actually create any additional value and is only good for promotional materials.

"Instead of a breakthrough tool, customers in need of optimization solutions have been sold an expensive marketing gimmick... D-Wave risks being a mere footnote in the history of the revolution of which it has declared itself the leader," the Kerrisdale analysts conclude.

Following the report, the Portnoy Law firm has launched an investigation into possible securities fraud and may file a class action lawsuit against D-Wave on behalf of investors.

On the other hand, in February, Jülich Research Center (Germany) announced that it is buying Advantage, D-Wave's latest quantum computer. "Connected to JUPITER, Europe's first supercomputer capable of more than a trillion operations per second, the system is expected to provide future breakthroughs in artificial intelligence and quantum optimization," the center said in a statement.

German scientists probably wouldn't buy a completely useless machine, would they? It's hard to say: after all, scientific research is one thing, but really useful commercial applications are quite another.

This example shows how difficult it is to navigate the world of quantum technologies. And this is still a public company, which at least publishes reports, let alone startups, where it is not at all clear what is really going on.

In mid-July, The Information published an investigation about SandboxAQ. According to the portal, the company's revenue amounted to only $10 million in the first quarter of 2025 (well below forecasts), while its expenses reach $200 million per year.

Despite the company's claims that it is already actively selling its technology to customers, 90% of the revenue came from research grants from Google co-founder Sergey Brin, who is a longtime friend of CEO Jack Hidary.

In addition, the company was investigating Hidari's expenses.

"Hidari had previously brought numerous young women who were not SandboxAQ employees to corporate events. According to current and former employees, he transported them on private planes paid for by the company, even when he himself was not on board," The Information writes.

However, the portal adds, again according to anonymous sources, that an audit of the expenditure did not seem to reveal any significant problems.

Joint use of quantum computers and AI is a promising idea, and Microsoft and IBM are exploring it. AI giant Nvidia is building a quantum research center in Boston and is negotiating investments in another quantum startup PsiQuantum, Reuters wrote in May. However, the practical application of quantum AI is still at least 7-10 years away, according to Bob Sutor.

"Until then, it is worth being cautious about statements in this area, where most often the buzzwords "quantum" and "AI" are simply glued together. You can get extra points if you mention "generative AI" as well! And this is not just sarcasm: forewarned is forearmed. Be sure to look for confirmation of the work and its scale from independent experts," he advises.

This article was AI-translated and verified by a human editor