The Quantum Week: FCA Picks Up the Qubits, Anthropic Picks Up the Banks, and Stablecoins Shed Fourteen Billion
Regulators are experimenting with quantum machine learning to hunt fraudsters, AI labs are carving up the banking enterprise market, and the stablecoin supply just posted its sharpest contraction since 2022. Welcome to the week technology stopped being adjacent to finance and became the point of it.
If you blinked this week, you missed roughly a decade of financial technology compression. The UK's top conduct regulator formally entered the quantum era. Anthropic and OpenAI squared off over the right to power the next generation of banking software. Hong Kong dropped a tokenisation framework with enough ambition to rattle London and Singapore. And stablecoins, the backbone of modern on-chain finance, shed more than $14 billion in a single month, the largest contraction since the 2022 bear market.
If the question was "is finance about to become a software problem," this week was the answer: it already is, and the software just got weirder.
The FCA Goes Quantum
Start with the headline that genuinely surprised a lot of practitioners. On November 5, the UK's Financial Conduct Authority (FCA) formally announced the launch of a Quantum Experimentation Group, with its first real-world use case focused on applying quantum machine learning to financial crime detection, according to the FCA's own press release and coverage from The Quantum Insider, Fintech Futures, and FS Tech.
What the FCA actually said
The regulator isn't just running a thought experiment. The Quantum Experimentation Group will bring together UK universities, quantum hardware providers, and regulated firms to test whether quantum-enhanced machine learning models can meaningfully outperform classical approaches at detecting anomalies in transaction data, the raw material of fraud and anti-money-laundering detection, according to Regulation Tomorrow and A&O Shearman's FinReg bulletin.
Why this matters: AML has been stuck in a well-documented productivity trap for years. Suspicious Activity Report (SAR) filings keep rising, false positive rates remain stubbornly high, and compliance teams burn through resources chasing noise. If quantum machine learning can genuinely compress that signal-to-noise problem, and that's still an open question, the economics of financial crime compliance change dramatically.
The context nobody should ignore
The FCA move slots into a bigger regulatory shift. CMS's recent analysis and the World Economic Forum's work on quantum-secure financial sector regulation both point to the same conclusion: regulators are no longer treating quantum as a 2030+ problem. The combination of post-quantum cryptography migration deadlines and quantum ML experimentation means financial firms need a quantum strategy in 2027, not 2029.
Expect other regulators, notably MAS (), the SEC, and BaFin, to publish their own quantum experimentation initiatives in the coming quarters.
Anthropic vs OpenAI: The Banking Core Is Up for Grabs
While regulators fiddled with qubits, the AI labs continued their push into the banking enterprise. PYMNTS reports that Anthropic and OpenAI are now in an open race to capture what the headline calls "banking's services core", the system-of-work that supports operations, servicing, risk, and research across every major bank.
What's new this week
Multiple sources confirm Anthropic has launched a dedicated enterprise finance unit, deploying a suite of bank-focused AI agents built in partnership with JPMorgan and Goldman Sachs, per coverage from PYMNTS, Pulse.bot, and The Paypers. The move follows Anthropic's reported $15 billion joint venture framework earlier this year.
OpenAI, meanwhile, is making its own moves. American Banker reports BBVA and Goldman Sachs have both participated in OpenAI's new venture, while eMarketer separately reports that OpenAI is expanding into UK banks even as Anthropic restricts access to its newest frontier model, codenamed internally as Mythos. The emerging picture: Anthropic is going deep with a smaller set of flagship banks, while OpenAI is going wide across the global banking landscape.
For CIOs, this is no longer an evaluation project. It's a procurement decision. The platforms being picked today will define the shape of banking software for the next decade.
Why bank leaders should care
An analysis highlighted by Banking Exchange reveals a parallel trend: banks are quietly reducing their reliance on any single AI model vendor by building their own orchestration layers, so they can swap models underneath. That's the smart defensive play. Whichever AI lab wins the banking market, no bank wants to be hostage to a single vendor's pricing power or roadmap.
HKMA Drops the DART Framework
On November 2, the Hong Kong Monetary Authority (HKMA) used the opening of Hong Kong FinTech Week 2026 to unveil what it's calling the DART framework, covering Data, AI, Regulation, and Tokenisation, per The Standard and the HKMA's own press release.
The framework isn't a single rule; it's a packaged set of initiatives touching tokenised deposits, regulated stablecoins, agentic AI for supervisory work, and cross-border data interoperability. Separately, Hong Kong officials signalled plans for new regulatory requirements specifically targeting tokenisation, according to The Block, and the SCMP reports that the city's fintech strategy is progressing, though with uneven growth across subsectors.
What's clear: Hong Kong is positioning itself as the Asia-Pacific hub for regulated tokenisation, going toe-to-toe with Singapore's Project Guardian and the UK's Digital Securities Sandbox. Cointelegraph separately confirms HKMA plans to expand its tokenised bond infrastructure. Expect cross-jurisdictional competition on tokenisation frameworks to intensify through 2027.
The Stablecoin Reset
While the policy news made the headlines, the numbers told a quieter but important story. According to The Block, stablecoin supply dropped $14.2 billion in the most recent 30-day window, the largest contraction since the 2022 bear market. Crypto Briefing confirms the $14 billion monthly drop. GN Crypto and Crypto News report a $6.7 billion weekly decline as of this week, the steepest weekly drop of 2026.
What's driving it
A few forces are converging. Rising US short-term rates have made holding stablecoins less attractive versus Treasury money market funds. Regulatory clarity on US-issued stablecoins has actually accelerated redemptions as some issuers reposition reserves. And, perhaps most interestingly, Token Terminal data cited by Crypto News indicates that less than 10% of stablecoins are currently deposited in DeFi, meaning the ecosystem is more exposed than it was two years ago to pure payment and treasury flows rather than yield-seeking capital.
Medium's Coinmonks column points out the paradox: even as supply contracted, DeFi protocol revenue actually went up during the window. In other words, the stablecoin base shrank, but the activity-per-dollar in on-chain finance rose. That's a healthier ratio than a lot of 2021-era commentary gave the sector credit for being able to produce.
Zama Spin-off Launches Confidential Stablecoin Payments
Also this week, in a quietly significant move for privacy-preserving on-chain finance, Zama spin-off Raycash launched confidential stablecoin payments and private yield on Ethereum, according to The Block. The product uses fully homomorphic encryption (FHE) to let users transact in stablecoins without exposing balances or counterparties on-chain.
For enterprise adoption, this matters enormously. One of the biggest blockers to corporate treasuries using stablecoin rails has been the transparency problem: a treasurer who doesn't want competitors knowing their payable schedule can't publish payment flows to a public blockchain. Confidential-by-default stablecoin rails change that math. Expect to see enterprise-focused stablecoin issuers watching Raycash very closely through 2027.
Quantinuum Takes the Stage at Cantor Convergence
Rounding out the technology-frontier beat, quantum computing firm Quantinuum confirmed it will participate in the Cantor Convergence 4th Annual Crypto, FinTech, AI & Energy Infrastructure Conference, per PR Newswire and Quantum Zeitgeist. For investors watching the quantum-finance intersection, these conference appearances matter: they signal where quantum providers believe their next commercial revenue pools will come from. "Fintech" being in the conference title tells you everything about where the money is flowing.
The Pattern, And What To Do About It
Take all of these stories together and a clear pattern emerges. Infrastructure that used to be experimental is becoming regulated. Infrastructure that used to be regulated is becoming experimental. The FCA is running quantum pilots. Hong Kong is codifying tokenisation. Stablecoins are consolidating from speculative capital to transactional capital. And AI labs are negotiating directly with the C-suite of the world's largest banks over what the next decade of banking software looks like.
For bank leaders, three practical moves stand out. First, get your quantum strategy out of the back-office research pile and onto the CTO's active roadmap. Second, don't wait for the Anthropic-vs-OpenAI dust to settle, build an AI orchestration layer that lets you swap models without re-architecting. Third, don't dismiss the stablecoin contraction as a bearish signal; it's a maturation signal, and the enterprise use cases Raycash and others are enabling will drive the next adoption curve.
The next six months won't be about picking winners. They'll be about building optionality fast enough to not get stuck with a losing bet.
As of this week, financial technology isn't converging with finance. It's replacing it. Pay attention.
