技术潜力集中于少数复杂任务，而非取代传统电脑。量子位元可同时处于 0 与 1，使机器能并行探索多种解答，特别适合化学、材料、药物发现、金融风险、投资定价、投资组合最佳化与诈欺侦测。Caltech 研究指出，有用量子电脑至少需要 1,000 个逻辑量子位元，而 OQC Genesis 目前有 16 个。OQC 最近完成 2.6 亿英镑融资，Chevron Technology Ventures 再度投资。
风险与限制同样明确。Q-Day 可能使现行大质数加密失效，威胁国安、医疗、金融与加密货币资料，并鼓励「先收割、后解密」。NIST 于 2024 年监督开发 3 种后量子演算法，2026 年又推进 9 种候选演算法。然而量子硬体仍受杂讯、热、磁场与错误校正限制；Gil Kalai 等怀疑容错能否足够可靠。
Quantum computing is moving from laboratories into commercial trials: industry observers estimate dozens of systems worldwide, while McKinsey forecasts about 5,000 by 2030. Google, IBM and other technology groups see useful machines around 2030, and Microsoft expects commercially viable problem-solving by 2029. Investor enthusiasm is rising too: Quantinuum’s Nasdaq IPO this month valued shares above $15bn, while IonQ’s stock has climbed more than 700 per cent since September 2024.
The technology’s promise is concentrated in a narrow set of complex tasks, not replacing conventional computers. Qubits can exist as both 0 and 1, allowing machines to explore multiple solutions in parallel, especially in chemistry, materials, drug discovery, financial risk, investment pricing, portfolio optimisation and fraud detection. Caltech research says a useful quantum computer needs at least 1,000 logical qubits, while OQC’s Genesis has 16. OQC recently raised £260mn, with Chevron Technology Ventures returning as an investor.
The risks and limits are equally clear. Q-Day could break current large-prime cryptography, threatening national security, health, financial and cryptocurrency data, and encouraging “harvest now, decrypt later” attacks. NIST oversaw three post-quantum algorithms in 2024 and advanced nine more candidates in 2026. Yet quantum hardware still faces noise, heat, magnetic-field disruption and error-correction barriers; sceptics such as Gil Kalai doubt whether fault tolerance can become robust enough.