FirstQFM reports quantum-powered forecasting results at ISC 2026, touting NVIDIA CUDA-Q build

FirstQFM, a startup building machine-learning foundation models for quantum computing, says it reached a new milestone for commercial quantum workflows at ISC 2026, the High Performance conference. In a market report published by Yahoo Finance, FirstQFM said its quantum-powered system outperformed leading classical forecasting approaches in a test presented at the conference.

The announcement links the work directly to NVIDIA technology. FirstQFM stated that its system was built using NVIDIA CUDA-Q, NVIDIA’s platform for writing and running quantum programs, and that the reported forecasting results were achieved with that quantum software stack.

FirstQFM’s framing was notable for what it emphasizes. Rather than focusing only on quantum circuit research, the company highlighted a practical performance comparison between a quantum-powered forecasting system and classical forecasting models, a category of tools used to predict future values from historical data. The implication is that quantum capabilities could be evaluated in applications with established benchmarking expectations.

For NVIDIA, the reference to CUDA-Q matters because it reflects a broader strategy in quantum computing. CUDA-Q is positioned as a developer-oriented framework that aims to make quantum programming more accessible for researchers and builders. By appearing in a public performance demonstration associated with forecasting, CUDA-Q is shown not just as a development environment, but as an enabling layer for an end-to-end solution a customer could consider using.

However, the published report does not provide key technical details that investors and researchers typically want when assessing claims of “outperformance.” The account does not disclose which classical models were used as comparators, which forecasting datasets or metrics were applied, the size of the test problems, or whether the results reflect a single benchmark run or repeated trials.

The report also does not describe the hardware configuration used for the quantum-powered runs, such as whether the system executed on a real quantum processor, a simulation environment, or a hybrid approach. Without those specifics, it is difficult to assess how broadly the results generalize beyond the particular experimental setup.

Market reaction to quantum announcements often depends on transparency. In this case, FirstQFM’s statement offers an outcome-level claim at a high-visibility venue, but leaves the methodological picture incomplete for outside validation. That may be appropriate for an initial demonstration, but it also means the strength of the evidence will be determined by what the company shares next, including benchmarking methodology and reproducibility details.

What to watch next is whether FirstQFM follows up with more granular information on the forecasting task, the experimental design, and the performance metrics it claims. For NVIDIA, continued evidence that CUDA-Q can support repeatable, application-level benchmarks could strengthen the case that quantum developers can move from prototyping toward production-oriented workflows. For the sector overall, public comparisons between quantum- and classical-driven forecasting are likely to be closely tracked, particularly if other teams publish comparable results with full methodological disclosure.