IBM Ventures Backs a Company Selling Quantum Algorithms on Ordinary GPUs, as Quantum Money Moves Down the Stack

by Vikram Desai · September 3, 2026

Funding & Deals Reporter, Frontier Tech Wire

Two deals announced on 2 September point at the same thing from opposite ends of the stack: investors are paying for what quantum computing needs in order to work, rather than for quantum computers.

The larger is a strategic investment by IBM Ventures in BQP, disclosed in a press release carried on PR Newswire with an 08:56 ET dateline. The size of IBM's cheque was not disclosed. The release states the round brings BQP's total funding to $8m, with Venn10 Capital joining and existing investor Monta Vista Capital participating; BQP's 2025 seed round was backed by New York State's venture arm alongside other institutional investors.

Selling quantum algorithms without selling a quantum computer

BQP's pitch is unusual enough to spell out. Its BQPhy platform takes solver techniques developed for quantum hardware and runs them on the CPUs and GPUs a customer already owns. The commercial premise is the company's own diagnosis, stated in the release and not independently verified: that classical physics solvers leave roughly 85% of a GPU's floating-point capacity idle, because the sparse iterative methods underneath simulation are bound by memory access rather than compute.

The product in production is called QuantumNOW and runs on customer infrastructure today. A second solver, QuantumMAX, is described as quantum-native and still in development, designed for heterogeneous execution across CPU, GPU and QPU. In other words, the company sells classical software now against the option of a quantum back end later.

Founder and chief executive Abhishek Chopra put the positioning this way in the release: "An engineer with a deadline doesn't care where the answer came from. They care that it arrived in time and that the physics is right." He added: "We spent those years earning our way into engineering workflows, so when the quantum machines are ready, nothing about how those teams work has to change." Emily Fontaine, global head of IBM Ventures, said in the same release: "What stood out with BQP is that users don't have to change how they work to get quantum-accelerated results."

Growth figures with no denominator

The release reports that since the 2025 seed round, contracted and committed revenue has grown eightfold, the customer base has tripled and platform users have risen twentyfold. Readers should note what is absent: no absolute revenue figure, no customer count and no user count are given, so the multiples cannot be sized. An eightfold increase on a small base is a different proposition from an eightfold increase on a large one, and the release does not permit the distinction. BQP is private, so there are no filed statements against which to check any of it.

The company does cite specific technical results, all of them self-reported. Under a Small Business Innovation Research award with the US Space Force and SpaceWERX, BQPhy is said to propagate a catalogue of more than 3,000 space objects in under a second, roughly 250 times faster than the open-source Orekit solver. Separate work on electric-vehicle battery thermal management is funded by India's Ministry of Heavy Industries with the Indian School of Business as partner. Both the 85% and the 250x are company-supplied figures; Frontier Tech Wire has seen no independent benchmarking of either.

A four-person bet on superconductors

The second deal is far smaller and further upstream. Meissner, a Toronto materials startup, raised $2.6m in US dollars - about $3.6m Canadian - in a pre-seed round reported by The Quantum Insider on 2 September. The company is four people. BDC Capital's Thrive Venture Fund anchored the institutional side, joined by a long list of individual backers: Andrew Talpash, Anthony Lacavera, Christian Weedbrook, Daniel Debow, Dennis Bennie, Eliot Pence, Greg Twinney, and Michael and Richard Hyatt.

Founder and chief executive Olivia Leng, who studied materials science chemistry at the University of Toronto, is building what the company calls a discovery engine: machine learning and quantum simulation used to nominate metal-based compounds, followed by laboratory validation. The stated goal is superconductors that operate at higher temperatures and quench less readily. Testing of the leading candidates began this month at the University of Waterloo's Quantum-Nano Fabrication and Characterization Facility, which means nothing has yet been confirmed in the lab.

Leng described superconductors to The Quantum Insider as "definitely the picks and shovels to unlocking high-growth, high-tech industries" - a framing that fits the BQP deal just as neatly. Neither company sells a quantum computer. One sells the software layer that would sit above one; the other, the material that would sit beneath one.

For readers tracking listed quantum names, the relevant point is where private capital is choosing to enter. Both rounds are bets that the input markets - engineering software, materials, fabrication - monetise before the machines do. That is a different thesis from the one embedded in most quantum equity stories, and it is being funded on a different timetable. This is reporting, not investment advice.