A strategy team at a US fintech sketching the next product on a whiteboard usually circles two things. What can we remove from the current flow? What do we add back so the result still works? Those are the operative questions behind disintermediation and reintermediation as a working practice. According to Bain research, US embedded finance flows will move about $7 trillion in 2026, and a meaningful share of the operator advantage in the segment comes from getting the remove-and-add decision right. This guide walks through how the cycle actually works inside US financial operators.
The cycle has been observed in many industries but the US financial services version has its own characteristics. Regulatory accountability stays with chartered entities, which constrains how aggressively disintermediation can play out. Trust and compliance functions are valuable enough that reintermediation almost always happens. The mechanics matter for operators trying to participate productively in the cycle.
How operators decide what to remove
Disintermediation decisions inside US fintech operators usually start with a value test. The team identifies an intermediary in the current flow and asks whether the value the intermediary delivers exceeds the fee it charges. If the answer is no, the intermediary is a candidate for removal. If the answer is yes, the team usually leaves the intermediary in place and looks elsewhere.
The test is harder than it looks. Some intermediaries deliver value that does not appear in any direct fee, like compliance protection that prevents future losses. Others deliver value that compounds slowly over time, like reputation that gives the operator credibility with regulators. The mature operators build models that capture indirect and delayed value, then run the test more carefully than the operators that look only at line-item fees.
The Federal Reserve’s payment systems framework shapes the analysis. Operators that disintermediate around correspondent banks for cross-border payments still need to satisfy Bank Secrecy Act obligations. Operators that disintermediate around acquiring banks for merchant settlement still need to handle disputes and chargebacks. The new flow has to deliver those functions or it carries unaddressed risk.
How operators decide what to add back
Reintermediation decisions follow the disintermediation. Once an intermediary is removed, the team asks which functions of that intermediary need to be replaced and who can deliver them at lower cost. The answer is usually a software platform that picks up two or three functions at once and a regulated entity that retains accountability for the rest.
The standard pattern in US fintech is to disintermediate one or two visible intermediaries and reintermediate with a software platform, a compliance vendor, and an underlying chartered entity. The configuration delivers most of the cost savings from the disintermediation while preserving the trust and compliance functions that regulators expect. The Banking-as-a-Service segment, projected by Fortune Business Insights at about $8.15 billion in the US in 2026, is the most visible piece of the reintermediation infrastructure.
The Genius Act, signed in July 2025, gave operators a specific tool for cross-border reintermediation. Stablecoin issuers, custodians, and on-chain analytics providers now have explicit roles under the law, which makes them easier to incorporate into a reintermediated flow. Visa’s stablecoin program reached a $4.5 billion annualized run rate by January 2026. Operators that disintermediated correspondent banks in 2024 are now reintermediating with stablecoin orchestrators that satisfy the Genius Act framework.
How the cycle is implemented in product
The product implementation usually proceeds in three stages. The first stage replaces the intermediary in a small share of transactions, often a single product line or geography. The second stage measures the outcomes against the operator’s targets for cost, speed, compliance, and customer experience. The third stage either expands the change to the full product surface or rolls back the change if the outcomes did not meet the targets.
The measurement layer is where most operators fail. A change that looks like a clear cost reduction in pilot can fail at scale because of operational issues that did not surface in the pilot. Operators that staged the rollout carefully and measured continuously caught the issues early. Operators that committed to a full rollout based on a small pilot often had to reverse course and lost customer trust in the process.
The compliance frame requires its own implementation discipline. The Office of the Comptroller of the Currency’s third-party risk management framework expects operators to document why each intermediary was removed and what replaced its compliance role. The Federal Deposit Insurance Corporation’s BaaS guidance expects sponsor banks to evaluate each change against their accountability obligations. TechBullion’s payments coverage documents how operators have handled the documentation discipline.
How the cycle shapes US fintech competition
The cycle is one of the central competitive dynamics in US fintech in 2026. Operators that ran the cycle successfully in payments now run it in lending, wealth, and insurance. Each segment has its own intermediary structure, its own regulatory frame, and its own reintermediation candidates. The operators that move from segment to segment with consistent discipline have outgrown peers that stayed in a single segment.
The cycle also reshapes the intermediary tier itself. Intermediaries that lose customer share to disintermediation tend to consolidate. The US BaaS sponsor bank count contracted from about 175 in 2023 to roughly 110 in early 2026, partly because disintermediation pressure squeezed the sponsors that could not modernize their compliance and technology stacks. The survivors are larger, more capable, and more strategically important to the operators that reintermediate around them.
The Consumer Financial Protection Bureau’s open banking rule under Section 1033 has shifted the cycle in favor of consumers and small businesses. Customers can now move their data between intermediaries, which weakens lock-in and lets the cycle play out faster. Operators that designed for data portability from the start are better positioned to participate in further rounds than operators that depended on lock-in to retain customers.
What to watch in the next twelve months
Three trends will shape disintermediation and reintermediation in US digital finance over the year ahead. The first is the expansion of stablecoin disintermediation. The Genius Act framework and Visa’s growing stablecoin program have given operators a clearer toolkit for displacing correspondent banks. The reintermediation tier built around stablecoin issuance and custody will grow in step.
The second is the rise of direct charters as a disintermediation strategy. A few US fintechs have applied for or received their own charters, which lets them bypass the BaaS layer entirely. The strategy has merit but the reintermediation costs are higher than they look. A direct chartered fintech still needs compliance, fraud, identity, and processor partners, which means most of the BaaS ecosystem reappears in a slightly different form.
The third is the role of AI in the cycle. Generative AI tools have begun disintermediating certain analytical, compliance, and customer service tasks. The reintermediation tier appearing in response includes AI-oversight vendors, model auditing services, and regulatory translation tools. The cycle will continue. The operators that read each round correctly will keep extracting value from the cycle. The ones that misread will lose share to peers that read it better. The discipline is not new but the pace is. US fintech operators that ran the cycle once a decade are now running it on a multi-year basis, and the gap between operators with strong cycle discipline and those without is widening.
