TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
FirstQFM reported that its quantum reservoir computing system, running on Rigetti hardware, outperformed the classical models included in a financial forecasting test across 43 anonymized assets and 600 out-of-sample periods. The company reported 52.02% directional accuracy and 112% higher cumulative gross portfolio profit than the strongest classical benchmark by profit; the release does not establish whether the results generalize beyond this evaluation.
FirstQFM said November 3 that its quantum reservoir computing system outperformed the classical forecasting models evaluated in a test spanning 43 anonymized financial assets and 600 out-of-sample periods. The company reported 52.02% directional accuracy and 112% higher cumulative gross portfolio profit than the strongest classical model by that measure, but the figures come from a company-reported evaluation and do not by themselves establish performance in live trading.
The system used quantum features generated on Rigetti Computing’s Cepheus-1-108Q hardware, according to FirstQFM’s announcement. Its foundation-model-based quantum reservoir computing system recorded 52.02% directional accuracy and a 51.76% macro-F1 score, which FirstQFM said were the highest among the models tested. Directional accuracy measures how often a forecast gets the direction right; the release does not provide enough detail to assess the practical significance of the margin over each competing model.
For an equally weighted portfolio covering all 43 assets, FirstQFM reported 112% higher cumulative gross profit and loss (P&L) than the classical benchmark with the highest P&L. Its annualized Sharpe ratio was 1.22, compared with 1.02 for the classical model with the highest Sharpe ratio. Those are distinct comparisons: the benchmark with the best profit was not necessarily the one with the best risk-adjusted return. The announcement describes gross results and does not say that they reflect trading costs.
FirstQFM also reported directional accuracy above 50% for 36 of 43 assets and positive gross P&L for 41. A separate five-asset portfolio, selected using what the company called a historical diversification methodology, had a reported Sharpe ratio of 1.73 and higher directional accuracy, macro-F1 and cumulative gross P&L than the full portfolio. The release does not provide the specific asset identities, and the figures remain results reported by the companies involved.
What the Forecasting Results Could Mean
If independently reproduced, results like these could help researchers assess whether quantum computing can add value to a practical forecasting task, rather than only demonstrate a laboratory calculation. The reported comparisons span forecast accuracy, classification performance and portfolio measures, offering more than one way to judge the system. Yet a higher backtest profit is not proof of reliable future returns, especially without information about transaction costs, market impact, exposure to risk and performance in live conditions.
The announcement also matters to quantum hardware developers. Rigetti CEO Subodh Kulkarni said empirical work can help probe current machines’ capabilities and limits and identify which hardware improvements matter. For investors and prospective enterprise customers, the results are an early signal to examine, not evidence that quantum forecasting is already commercially superior. The source material does not include an independent replication or customer trading record.
Products Worth Considering
Bose QuietComfort Headphones (2nd Gen), Wireless Over-Ear Noise-Cancelling Headphones, Hazelnut Taupe
Bose QuietComfort Ultra Headphones (2nd Gen), Wireless Bluetooth Noise-Cancelling Over-Ear Headphones, Black
Apple MacBook Air 15-inch with M5 chip, 16GB Memory, 512GB SSD, and Liquid Retina Display
As an affiliate, we earn on qualifying purchases.
How FirstQFM Set Up the Test
The evaluation used data supplied by SW7 Group, which anonymized the assets and forecasting task before sharing them with FirstQFM. The target was next-day forecasts. FirstQFM says the evaluation separated data chronologically into training, validation and test periods, and that model selection relied only on validation data. That setup is intended to keep test-period observations from guiding model selection, although the public announcement alone is not enough to verify every methodological choice.
The comparison set included classical machine-learning and statistical models, as well as time-series foundation models identified in the release as Chronos-2, Granite TTM-R2 and Toto 2.0. FirstQFM said its quantum features were produced on Rigetti hardware. The company released a beta version of its foundation-model-based QRC forecasting system in July 2026; it says the system is based on proprietary foundation-model technology. The release points to an accompanying technical blog for detailed methodology and results.
“Empirical work such as FirstQFM’s helps us probe the capabilities and limits of today’s quantum computers, while helping hardware teams understand which improvements matter most.”
— Subodh Kulkarni, CEO of Rigetti Computing
Questions About the Backtest Evidence
The announcement does not specify the dates covered by the 600 out-of-sample periods, the exact forecast and portfolio-construction procedures, or the fees and other trading costs included in the gross P&L. It also does not give the asset-level results in a form that would let readers assess how performance varied across the sample. The comparison basis for FirstQFM’s stated 112% profit advantage is the strongest classical benchmark by P&L, but the release does not provide the underlying dollar amounts.
It is not clear from the supplied material whether an independent party audited the analysis or whether the result has been reproduced using another dataset. The announcement also does not report live deployment results or show whether the model continues to perform under changing market conditions. The company’s results are therefore best described as reported benchmark findings, not a guarantee of trading performance or proof of general quantum advantage.
Next Steps for Independent Review
FirstQFM’s technical blog is the next source for readers seeking the full methodology and results, which the announcement says are available there. The key next developments would be further disclosure of the test window, costs, model configurations and portfolio rules, followed by independent replication on other data. Evidence from live or paper trading, reported with clear risk and cost measures, would help determine whether the benchmark advantage carries over beyond the evaluation.
No independent review, follow-up test or live-trading milestone is specified in the source material. Until more information is available, the confirmed development is the companies’ announcement of the evaluation and its reported metrics; the extent to which those results generalize remains open.
Key Questions
What did FirstQFM report?
FirstQFM said its quantum reservoir computing system beat the classical models tested on next-day forecasts across 43 anonymized assets and 600 out-of-sample periods. It reported 52.02% directional accuracy and higher cumulative gross portfolio P&L than the strongest classical benchmark by profit.
What does the 112% figure compare?
It is FirstQFM’s reported difference in cumulative gross portfolio P&L between its equally weighted 43-asset portfolio and the classical benchmark with the highest P&L. The announcement does not provide the underlying amounts or say that trading costs were deducted.
Was the test independently verified?
The supplied announcement does not identify an independent auditor or report an external replication. FirstQFM says SW7 Group anonymized the data and that model selection used validation data, but the published results are reported by the companies involved.
Does the result show the system will make money in live markets?
No. The announcement describes an evaluation and gross P&L, not verified live-trading returns. Fees, market impact, the test dates and performance on other datasets are among the details needed to judge how well the result might carry over.
Source: thequantuminsider.com
