大约三年前，Goldman Sachs与Amazon合作后最初被视为早期领先者，但其基准测试将整个领域的预期重设：其发现某些投资组合问题需要数百万年才能计算，且至少需要800万个逻辑量子位，而当前设备少于100个。随后Goldman在大规模成本削减下解散了大部分量子团队。相比之下，JPMorgan保留了更大的研究阵容（超过50名科学家与数学家），并将研究重点放在优化、机器学习与密码学上，显示这家大行已从短期撤退转向长期布局。
JPMorgan引用了来自Quantinuum的Helios及2025年与Amazon共同展示的早期成果：可识别大规模不相关资产集合，以改善分散化与风险管理。该行表示真正可用的部署可能还要等待几年，取决于硬件的商业可行性。金融机构的整体格局是分歧的：UBS正在对约50名量化分析师进行技术提升，BBVA与Credit Agricole则在投资组合与信贷场景试行，许多机构也在为未来可能的加密破解进行密码升级。即便如此，领先者也承认，现有架构规模过小且杂讯过高，产业软硬件进展仍主要处于试验性阶段。
在金融业之外，Alphabet与IBM仍通过Willow与Heron等晶片主导平台推进，而BMW、Novo Nordisk、Roche和Exxon Mobil则在行业内进行场景化试点。市场数据显示泡沫循环：2024年Rigetti涨幅为1,450%，D-Wave为854%，IonQ为237%，随后回报降温，因实际限制逐渐明显且AI成为更集中关注。2026年HSBC报告在IBM Heron上，债券定价预测可提升高达34%，但样本集较小且机制不明。当前主流趋势因此从投机型热情转向以基准验证为核心、跨多年期的执行路线，大型玩家准备进入长周期落地。
Three years after Goldman Sachs first positioned itself as an early mover with Amazon, the bank’s benchmark reset expectations in the field: it found some portfolio problems would need millions of years to compute and at least 8 million logical qubits, while today’s devices have fewer than 100. Goldman then dissolved most of its quantum group under broader cost cuts. By contrast, JPMorgan maintained a much larger effort (more than 50 scientists and mathematicians) and now frames quantum work around optimization, machine learning, and cryptography, illustrating how one major bank moved from short-term retreat to long-horizon positioning.
JPMorgan cites early, narrow wins from Quantinuum’s Helios and a 2025 demo with Amazon that identifies large sets of uncorrelated assets for better diversification and risk management. The bank says useful deployments are probably a few years away, contingent on commercially viable hardware. In finance the pattern is mixed: UBS is upskilling about 50 quant analysts, BBVA and Credit Agricole are testing portfolio and credit applications, and many institutions are upgrading cryptography in response to future-breaking risks. Yet even leaders admit current architectures are too small and too noisy; industry-wide, software/hardware progress remains mostly experimental.
Outside finance, Alphabet and IBM still lead platform momentum through chips like Willow and Heron, while BMW, Novo Nordisk, Roche, and Exxon Mobil pursue sector-specific pilots. Market data shows the hype cycle: in 2024 Rigetti rose 1,450%, D-Wave 854%, and IonQ 237%, then returns cooled as practical limits became clearer and AI absorbed attention. In 2026 HSBC reported up to 34% improvement in bond-price forecasting on IBM Heron, but with a smaller dataset and unclear mechanism. The prevailing trend is therefore a shift from speculative exuberance to benchmark-driven, multi-year execution, with major players now preparing for long implementation horizons.