Bernstein: Meta Muse Disrupts Financial Industry; Visa and Mastercard Could Emerge as Winners
nashnova research
Meta's AI agent Muse hit No. 1 on the U.S. App Store with 2.8 million downloads, dragging financial stocks down 5%–15% — yet Bernstein argues payment networks Visa and Mastercard could end up the biggest beneficiaries.
What is Muse actually doing for consumers?
Muse is Meta's consumer-grade AI agent — a digital assistant that can act on your behalf, not just answer questions. It topped the U.S. App Store immediately after launch, reaching 2.8 million downloads.
Consumers are using it to book reservations, order food, track flight prices, cancel subscriptions, compare insurance quotes, and contact customer service. This means → Muse operates across the full spectrum of daily financial friction.
Bernstein calls this a potential inflection point for consumer AI — a "consumer AI agent race" is now taking shape.
Why did banks, insurers, and brokerages sell off first?
Since Muse launched, stocks across insurance, brokerage, banking, and mortgage sectors have fallen 5%–15%.
The core fear: AI agents don't just advise — they act on behalf of consumers, gradually eroding the habit-based switching barriers financial firms have relied on for decades.
In plain terms = financial institutions have long profited from consumer inertia — people rarely shop around for a better insurance rate or move idle cash. An AI that does it for them attacks that profit engine directly.
How does the disruption hit each financial sub-sector?
Insurance: AI auto-compares quotes and helps switch providers → insurers lose the pricing power that comes from renewal inertia.
Brokerages: AI continuously hunts for higher-yield cash products → client funds move between accounts more easily → brokerages earn less from idle cash balances.
Banks: easier deposit migration could push up funding costs, pressuring net interest margins — the spread between what banks charge on loans and pay on deposits, their core profit driver.
Credit cards: AI picks the optimal card in real time based on rewards, rates, and promotions → the habit of keeping one "top-of-wallet" card may gradually fade.
Why would financial institutions refuse to cooperate?
Bernstein flags a core tension: without merchant and financial-provider participation, AI agents cannot scale — but participating means giving up the checkout gateway or the customer relationship, so firms lack incentive to fully open up.
Amazon has already blocked Muse, citing its failure to identify itself as an AI agent and its access to customer login credentials. Online insurance platform Insurify also blocked it, arguing automated quote-scraping may miss critical information.
This means → expect more agent-blocking, paid data-access agreements, and in-house AI agent development. JPMorgan has begun charging data aggregators for customer-data access — a template for future AI-platform fees.
Do consumers actually trust AI with their money?
A 2026 TD Bank survey shows 55% of U.S. consumers now use AI to help manage finances, up from just 10% a year earlier.
Yet while 62% believe AI can provide reliable information, only 18% would let AI independently make important financial decisions. In plain terms = most people's stance is "you can research for me, but don't sign on my behalf."
ACI Worldwide / YouGov found only 7% would let an AI assistant purchase without approval; Accenture's survey of 25,590 consumers found only 12% would grant AI full autonomy at payment. This means → AI agents are likely to remain "assistants," not financial-life takeover machines, in the near term.
Why might Visa and Mastercard actually win?
Markets widely view AI-agent commerce as a threat to Visa and Mastercard, but Bernstein argues the reality may be "exactly the opposite."
This reflects an infrastructure-layer logic: AI agents drive more digital transactions and increase demand for identity verification, risk management, and dispute resolution → payment networks, as "trust infrastructure," benefit rather than suffer. Visa and Mastercard hold over 7 billion payment credentials covering more than 100 million merchants.
Both have launched agent-facing payment solutions — Visa Intelligent Commerce and Mastercard Agent Pay — and Muse already uses single-use virtual cards. Bernstein rates both Visa and Mastercard Outperform, with price targets of $450 and $710 respectively.
Bernstein's bottom line: the three real constraints on AI-agent adoption are not model capability but consumer trust, data access, and regulation — whether these three gates open will determine how fast and how deeply the financial industry's business models are reshaped.
市场有风险，内容仅供研究参考，不构成投资建议。