Mentioned in This Article
1. Mizuho Trims Its Circle Target and Keeps an Underperform Call
Mizuho lowered its price target on Circle Internet Group, the issuer of the USDC stablecoin, to $43 from $45 in a research note dated Monday, October 26, 2026, and kept its Underperform rating. The bank argues that AI agents acting for consumers could pull $4 billion to $9 billion out of USDC's market capitalization and cut Circle's reserve revenue by 10% to 20% over the medium term.
The new target sits roughly 46% below where the shares traded when the note appeared. Circle stock fell about 7% on Monday as investors absorbed the argument, which is unusual because it concerns neither a rival token nor a regulator nor a downturn in digital asset prices. Mizuho's concern is that software now being built to manage personal finances may simply not choose USDC when it decides where a user's cash should sit.
A target cut of $2 is small in itself. What drew attention was the reasoning attached to it. Mizuho kept its bearish stance and, instead of leaning on familiar worries, introduced a new channel through which demand for the stablecoin could erode. That is why the note is being read as a statement about how automated financial assistants could reshape where money is parked, with Circle as the first listed company to be tested against the idea.
Circle's USDC is one of the largest dollar-pegged stablecoins, used for trading, payments and settlement across blockchains, and the company trades publicly, which makes its fortunes easier for investors to track than those of most issuers. Any shift in how consumers decide to hold dollars therefore shows up quickly in its share price and in analyst models.
2. What Did Mizuho Find About AI Agents and USDC?
Mizuho tested how AI personal assistant engines respond when asked where to hold money. According to the research, assistants including one called Muse steered users away from USDC and toward bank products insured by the Federal Deposit Insurance Corporation (FDIC) that pay higher yields. The firm treats this as evidence that an agent optimizing for return and safety could consistently pick a bank account over a dollar-pegged token.
The detail that stands out is that the agents are described as making these recommendations even when users said they preferred digital assets. In other words, a stated lean toward crypto did not override the agent's ranking of the options. If that behavior holds as the tools improve, a user who delegates the decision may end up with an allocation different from the one they would have made by hand.
This is a finding from a controlled exercise, not a measurement of real flows. Mizuho has not reported that USDC balances are falling because of AI assistants. Its figures are modeled estimates of what could happen if such agents gain wide adoption, and they should be read as a scenario for the medium term rather than a description of current demand.
3. How Large Is the Revenue Risk?
Mizuho puts the possible reduction in USDC market capitalization at $4 billion to $9 billion. Applied to Circle's income, the bank estimates that a shrinking pool of tokens could cut reserve revenue by 10% to 20%. The widely repeated 20% figure is therefore the top of a range, not a central forecast, and the lower end of the same range is half that size.
The scale matters because of the way Circle earns money. Each USDC in circulation is backed by reserve assets that Circle holds, and the interest those assets generate is retained by the company. Fewer tokens outstanding means a smaller reserve pool and less interest income. A decline in circulation therefore passes almost directly into the income statement, with few other revenue lines to cushion it.
That mechanical link explains why a demand forecast can move the stock so sharply. Investors are not being asked to believe that USDC is unsafe. They are being asked to weigh how many people will choose to hold it, and what that number does to the single most important source of income for the business.
4. Why Does Reserve Income Drive Circle's Business?
A stablecoin issuer takes in dollars from holders, invests them in short-dated, high-quality assets and keeps the yield, while token holders typically receive no interest from the issuer itself. This structure makes the business sensitive to two variables: the total amount of the token in circulation and the level of short-term interest rates. Mizuho's note concentrates on the first of those.
It also explains the competitive logic of the argument. A bank deposit that is FDIC insured and pays a higher rate competes for the same idle dollars as a stablecoin. Until now, the comparison was left mostly to the user, who weighed convenience, speed of transfer and use in crypto markets against the interest foregone. An AI agent given a mandate to maximize return on cash could weigh those factors differently and more consistently, which is the behavior Mizuho says it observed.
Because the issuer does not share the yield with holders, a rational agent has little reason to prefer a token that pays nothing over an insured account that pays something, unless the user needs the token for a specific purpose. The bank's warning rests on that asymmetry.
5. How Did the Market React?
Circle shares closed down about 7% on Monday after the note circulated, extending the pressure on a stock that Mizuho has treated skeptically for some time. The firm's Underperform rating means it expects the shares to lag the broader market, and the revised $43 target implies a decline of about 46% from recent trading levels.
The reaction shows how closely investors watch any analysis that touches the growth of USDC. The token's circulation is the foundation of Circle's earnings, so any credible suggestion that it could shrink is quickly priced in. At the same time, the move reflects a forecast and not a reported result, and the market may adjust again as more data arrives.
Mizuho's trimmed target is also a reminder that sell-side views on Circle have shifted over time, with the same firm having changed its stance on the stock in the past. Readers should treat any single target as one analyst's view at one date.
The coming weeks will show whether other research desks adopt the same framework. If more analysts begin modeling the behavior of consumer assistants, the question of how AI agents allocate cash could become a standard part of valuation work on stablecoin issuers.
6. Does the Warning Concern USDC's Safety?
No. The estimate is a forecast about demand, not a warning about USDC's backing or redemption. Nothing in Mizuho's argument says that tokens could not be exchanged for dollars. The claim is narrower: that fewer people, or fewer automated agents acting for them, may choose to hold USDC in the first place.
For existing holders, the practical consequences are limited. A smaller market capitalization would not by itself change the reserve ratio or the ability to redeem. For shareholders, by contrast, the same development would reduce the pool on which Circle earns interest, which is a financial question about the company and not a question about the asset.
The distinction is worth keeping in mind when headlines summarize the story as a risk of a 20% revenue loss. The figure describes a possible effect on the company's income under a particular adoption scenario, and the underlying range runs from 10% to 20%.
7. Why Agent Behavior Matters Beyond Circle
The note touches a broader question for digital assets: what happens when decisions about where to hold money are delegated to software. Today most stablecoin demand comes from traders, payment users and people seeking dollar exposure outside the banking system. Those users choose deliberately. If a growing share of cash management is handled by assistants that compare yields and protections, the criteria that decide who wins deposits may change.
Such agents tend to compare options against explicit objectives, such as return, insurance coverage and liquidity. Tokens that offer clear advantages for a task, for example settling a transaction on a blockchain, could still be chosen when the task demands it. Tokens held mainly as a store of dollar value would face direct comparison with insured bank products, and that is the segment where Mizuho sees the pressure.
Mizuho's research does not claim that this is already happening at scale. It identifies a direction of travel and attaches a dollar range to it, leaving it to later data to show whether the behavior of consumer assistants resembles what the firm saw in testing.
8. What's Next for Circle and USDC?
The next scheduled test arrives on November 4, when Circle is due to report third-quarter results. That release should show USDC balances and yield data, allowing investors to see whether circulation is growing, flat or shrinking, and how reserve income is responding to interest rates.
Analysts will look for any sign that the AI thesis is appearing in actual numbers. A stable or rising circulation figure would weaken the argument, at least for the near term, while evidence of outflows would lend it support. Management commentary on how it expects AI-driven payments and assistants to affect demand could also shape how the note is judged.
Until then, Mizuho's $4 billion to $9 billion estimate remains a modeled scenario. Its significance lies in the question it raises for the stablecoin sector: whether the customers of the future are people choosing a token, or software choosing on their behalf.