IonQ jumps as much as 15% on an Nvidia install and an error-correction test, then gives most of it back
IonQ said a Superion 256 machine will be the first quantum computer at Nvidia's quantum research center, and that its decoder kept pace across 408 logical qubits. That test ran on simulated circuits, the Nvidia deal has no disclosed dollar value, and the stock was up under 5% soon after the open.
IonQ shares gained as much as 15% before the opening bell on Wednesday, CNBC reported, after two announcements in two days. By 9:57 a.m. Eastern the stock was at $42.61, up 4.6% from Tuesday's close of $40.74, according to Nasdaq data, after touching $46.02 in the first minutes of trading. That leaves it about half its 52-week high of $84.64. Other quantum stocks, including Rigetti Computing and D-Wave, rose more than 4% before the open, CNBC said.
What IonQ announced
On Wednesday, IonQ said a Superion 256 system will be the first on-premises quantum computer at Nvidia's Accelerated Quantum Research Center. It will be wired directly to an Nvidia GB200 NVL72 system, and the two companies plan joint research on hybrid quantum and GPU software, with target uses in financial risk modeling, materials science and drug chemistry. Installation is scheduled for next year. IonQ said the Superion 256 is available to order now, with first customer deliveries expected in 2027.
On Tuesday evening, IonQ said it had built what it calls the industry's first end-to-end real-time error-correction decoder running on a single ordinary CPU. Quantum bits are fragile, and correcting their errors quickly enough that the machine does not have to stop and wait is one of the main hurdles to useful quantum computers.
What the releases do not say
Two details matter for anyone reading the headlines as a revenue story.
The big decoder numbers came from simulated circuits. IonQ's release says it evaluated the decoder "across complex benchmark circuits simulating up to 408 logical qubits," running more than 31.5 million operations, with the decoder adding as little as 0.02% to run time. That is a meaningful engineering result, and the research is posted on arXiv. But it is a test of the classical decoding software against simulated workloads, not a 408-logical-qubit quantum computer running today. IonQ itself describes the result as a foundation for its roadmap "beyond 256 physical qubits."
No money is attached to the Nvidia deal. Neither the price of the Superion 256 system nor any payment terms were disclosed. The release describes a research tenancy and joint program, not a purchase order. Nvidia was already named in IonQ's materials as a partner, and the two companies published joint research with Oak Ridge National Laboratory earlier this month. In that study, IonQ's own release noted, every circuit was simulated on an Nvidia GPU rather than run on quantum hardware.
Why the fade makes sense
A company worth about $16.2 billion at Wednesday's price, based on Nasdaq's market value figure, is valued on expectations of future contracts and revenue, and neither announcement adds a number to that line. What they do add is credibility for IonQ's technical roadmap and a showcase placement inside the largest AI chipmaker's research effort. We cannot know why individual traders sold, but the pattern fits a first reaction to the headlines followed by a second look at the details.
For traders, the next markers are concrete: the 2027 Superion delivery schedule, whether any NVAQC work turns into paid customer programs, and IonQ's next quarterly report. The company recently raised its full-year revenue outlook after its acquisition of chipmaker SkyWater, CNBC noted. For the rest of today's market, see our markets page.
Sources: IonQ; CNBC; Nasdaq. Share prices are from early trading and will change. This is market information, not investment advice.
Want your business to be the answer?
Get a full package of articles about your business, built so customers, Google and AI assistants can find you.