In 2019, Google said it had reached “quantum supremacy.”
IBM disputed the claim almost immediately, arguing that a properly tuned supercomputer could nearly match the result within days rather than the ten thousand years Google projected, according to CBS News.
The fight exposed the field’s real weakness. Nobody outside the company making the claim could actually verify it, and that argument has followed quantum computing ever since.
On Thursday, IBM ended up on the other side of that same argument. The company announced three separate quantum computing results, with the University of Chicago, Israel’s Qedma Quantum Computing, and Italy’s Algorithmiq, that it says clear the bar Google never fully reached, according to IBM’s announcement.
Each project paired a difficult computation with a built in way to prove the answer was correct, not just fast.
IBM, the century-old technology company now betting part of its future growth on quantum hardware, needed exactly this kind of result. Its 2019 objection to Google was never really about raw speed.
It was about trust, and that gap has shadowed the entire industry for seven years.
The trust problem finally gets an answer
This week’s announcements target that gap directly. With the University of Chicago, researchers encoded 70 logical qubits, a cluster of physical qubits bundled together so errors can be caught and corrected in real time, and completed a calculation in about 15 minutes, according to IBM.
The result matters less for its speed than for what came with it. The team built error detection into the computation itself, so the output arrives with its own proof attached instead of asking anyone to take IBM’s word for it.
That distinction is what separates this from 2019. Skeptics of quantum computing have never really doubted the machines could produce strange, hard to replicate numbers.
They doubted whether those numbers meant anything. A self verifying result closes that specific opening.
Two more partners point to the same conclusion
With Qedma, IBM used its Heron quantum processor alongside Qedma’s error mitigation software to model a two dimensional Floquet Ising system, a physics problem tied to magnetic behavior in materials.
The simulation reached 74 qubits, and classical methods tested independently by RIKEN and BlueQubit could not reliably reproduce the results at that scale, according to InvestorsHub. That inability is the actual finding, not a footnote.
Related: IBM CEO makes bold AI strategy claim
With Algorithmiq, IBM simulated a complex material relevant to catalysts and battery electrolytes, then published the underlying benchmark eight months ago through what it calls the Quantum Advantage Tracker.
No classical method has matched it since, according to The Quantum Insider. Algorithmiq also released its own classical simulation tool publicly, daring competitors and skeptics to beat the quantum result themselves rather than simply doubt it.
CEO Arvind Krishna said the same Algorithmiq work uncovered material behaviors researchers could not observe with classical computing, pointing to potential applications in better batteries, advanced materials, fusion energy, and improved medicines, according to CNBC.
IBM announced three verified quantum advantage demonstrations Thursday.
MANDEL NGAN / Getty Images
The market barely noticed
Despite three verified quantum results landing the same morning, IBM shares fell nearly 2%, closing at $221.96. The stock sits roughly a third below its 52 week high of $332.46, set in June.
The muted reaction is not about disbelief. Fourteen of sixteen analysts still rate IBM a buy, with an average price target near $244, implying meaningful upside from current levels, according to Investing.com.
Wall Street simply is not pricing quantum computing as a near term revenue driver yet, regardless of how credible the science looks.
That disconnect says more about IBM’s current story than about quantum computing. The company missed second quarter revenue expectations and cut its full year growth guidance, and investors are still weighing what that means for the rest of the business, according to Markets Daily.
More IBM:
- IBM stock surged again: The reason sounds like sci-fi
- IBM’s latest Wall Street call hides bigger shift
- IBM just answered a $5 billion cybersecurity question
A verified quantum experiment, however credible, does not fix a near term earnings problem, and traders priced it that way.
The scale of that earnings problem is worth spelling out. IBM shares dropped 25% on July 14, the stock’s worst single day decline on record, after the company disclosed that some customers had delayed capital spending, and shares have recovered only 2% since, according to CNBC.
Krishna has framed those delays as deferrals rather than lost business, noting that about 40% of the postponed deals closed within three to four weeks.
Quantum advantage is not quantum profit
What IBM proved this week is technical, not commercial. Quantum advantage in the strict sense demonstrated Thursday is not the same as a paying customer.
IBM and its rivals, including Google and smaller pure play firms like Rigetti and IonQ, still cannot point to a business application that only a quantum computer can solve at meaningful commercial scale.
Krishna is nonetheless putting a date on it. He told CNBC’s “Mad Money” that quantum computing should have “a measurable impact” on IBM’s revenue and earnings by 2028 or 2029, and said the company now sees it as “a trillion dollars of value” by the end of the 2030s.
Skeptics counter that commercialization remains distant given high error rates, hardware complexity, and scalability challenges still facing the technology, according to CNBC.
In May, IBM committed $1 billion of its own capital toward a standalone quantum chip foundry, matched by a $1 billion award from the U.S. Department of Commerce under the CHIPS program, as Alphabet and Rigetti Computing pursue their own paths to commercialization.
What changed is narrower but more durable than a single stock move. IBM spent seven years building the credibility that Google’s 2019 claim lacked, and this week it used that credibility as the actual product, ahead of any qubit count or processor name.
The next contest in quantum computing will not be decided by who is fastest, or even who is first. It will be decided by who researchers, rivals, and eventually investors actually believe, and whether that belief ever shows up in a revenue line rather than a research paper.