This post is less a set of predictions than it is a working thesis towards the next chapter of fintech. We’ve outlined what we see as the foundations that underpin the generation before us, four particularly compelling opportunities, and a few assertions about the future of payments, money, and financial services.

I. Foundations

Macro developments support the thaw in early-stage fintech. Public markets are again rewarding innovation, not just short-term profit. Multiples have reset. And regulation is catching up, particularly around stablecoins.

Public Markets

Even at scale, some of the largest public fintechs ship new products and post growth metrics at the pace of a lean startup. And public markets investors reward it, which instills downstream confidence in the ecosystem. A few examples:

• Robinhood has moved well beyond its core trading app, now operating nine product lines with over $100M in revenue each. At the center of its roadmap is asset tokenization, a natural extension of its mission to democratize finance. That means making U.S. assets accessible abroad, bringing non-public assets like art or private companies on-chain, enabling 24/7 trading, and offering self-custody through blockchain-native infrastructure¹. Importantly for startups, tokenization will depend on new infrastructure for custody, settlement, valuation, and liquidity which creates wide-open opportunities

• Circle is evolving from a stablecoin issuer into a broader fintech platform. In just the last year, it launched programmable wallets (with over 19 million deployed), Circle Gas Station (abstracting gas fees for end-users), and a cross-chain transfer protocol (CCTP)²

• Affirm has endured through multiple cycles, despite its distinct boom during ZIRP days. Its resilience shows that BNPL has become a genuine alternative to relying on interchange and revolving balances to extend credit, a genuine crack to the cards hegemony

• Nubank still grows topline like a startup, even after a heroic push to profitability during the fintech winter. The company has aggressively expanded across Mexico and Colombia and new product lines from insurance to SME credit while maintaining the core model: impossibly low CAC, low cost-to-serve (now under $1), no branches, and a tightly integrated tech stack³

Importantly, the rise in value (perhaps with the exception of Robinhood and Coinbase) is not based on inflated multiple expansion as much as fundamentals. Consider EV/ NTM sales multiples over time for some iconic names:

Regulation

Global regulation is hardly fixed, but the posture is materially different than it was during the fintech winter. As the direction of travel seems more constructive, teams will find it easier to plan, ship, and scale with greater confidence. A few recent developments:

• Stablecoins (US + EU): The GENIUS Act, now passed with bipartisan support, creates a federal license for stablecoin issuers with 100% reserve backing and permits treatment as cash equivalents under GAAP. In the EU, MiCA has gone live as the first comprehensive regulatory framework for digital assets in a major economy, establishing unified standards for reserve backing, licensing, custody, and disclosures. It brings previously fragmented national rules into a single regime (…More on these regulations later)^6,7

• EUDI Wallet (EU): The European Digital Identity Wallet will allow citizens to store and share verified identity, credentials, and payment details across borders. For fintechs, the program creates a common KYC and authentication layer

• Project Guardian (Singapore): This MAS-led initiative brings together DBS, JPMorgan, and others to test tokenized deposits, on-chain FX, and programmable settlement logic. It’s the most feasible attempt from a government to bring real-world assets and compliance-grade logic to public infrastructure.

• RTP Networks (Global): India’s UPI and Brazil’s Pix set the benchmark, moving real-time payments from simple P2P transfers into B2B and credit use cases. Today, one in eleven adults worldwide uses one of the two.⁸ What began as national experiments is now global: FedNow in the U.S., Europe’s TIPS, and cross-border pilots in Southeast Asia and Africa are following the model. And these networks are expanding beyond instant settlement to add services like on-rail lending and Request-to-Pay (R2P).

Building on the Last Generation

Much of the innovation driving the current fintech wave wouldn’t be possible without the infrastructure built in the last one. A few key developments now serve as foundations for the next cycle:

• Open Banking led by foundational companies like Plaid unlocked permissioned access to real-time financial data. Plaid has since expanded into a broader platform of open banking enablement, from identity verification to payments, that empowers agents to read, reason over, and act on individual financial context

• Blockchain introduced programmable money that can move globally, 24/7, without intermediaries. That plumbing now enables agents to transact, settle, and store value with deterministic finality

• Mobile Apps reshaped distribution and now serve as the starting point to distribute the next generation of fintech apps from real time payments to AI wealth managers

• BaaS lowered the regulatory and technical barriers for embedding financial products. Banks like Column, Lead, and Cross River provided the core foundations that allowed fintechs to issue accounts, manage funds, and move money without becoming licensed institutions themselves, a model that now lets AI-native platforms embed financial services as a primitive rather than a prohibitive hurdle

These foundations preface the next chapter in fintech, including the three themes we’ve outlined next. 

II. AI Rewrites Cost Structures

The last true cost reset in financial services came from mobile. In the early 2000s, distribution still meant branches, call centers, infamous tents at college campus move-in days. Onboarding might require weeks of paperwork. Simple transactions, approvals, or deposits required a trip to the nearest branch. Mobile collapsed those economics.

AI has the opportunity to make that shift look quaint because it fundamentally compresses the costliest components like personalization, reasoning, compliance at near-zero marginal cost. First, it extends what software was already great at: automating routine tasks that are expensive, slow, and error-prone like parsing through documents, categorizing transactions, or routing policies. With LLMs or agentic workflows, these steps become even more instant, scalable, and context-driven.

But second, and more importantly, AI introduces core reasoning such that it can interpret messy inputs, carry forward consumer or enterprise context and memory, and pursue relevant next actions without reintroducing human intervention. So instead of designing an onboarding flow for the median customer, platforms can support edge cases at scale; or instead of scripting a help desk, a support product can surface context-driven, personalized answers and resolutions.

Wealth Management: Case study on unlocking new segments

Wealth management, as a case study, shows well how AI will collapse traditional cost structures. The industry today is less inherently exclusive as it is gated by unit economics. Managers who charge 1% of AUM, for example, just have a natural floor under which the math stops working. The costs weighing on the viability to serve lower ARPU customers include:

• Labor-intensive onboarding: KYC, risk tolerance intake, regulatory disclosures, investment policy statements. Every step requires compliance review and client back-and-forth

• Portfolio personalization and reporting: In theory, every portfolio should be as specific as the client context. Yet executing that pre-AI requires an understanding of the client that is unfeasible as a client base scales

• Support and servicing: Clients expect advisors to respond and react to how macro and micro events affect their specific portfolio, planning, and performance (which does not compress linearly as AUM decreases)

• Compliance and documentation overhead: Registered advisors are subject to audits, and disclosure requirements that add legal and administrative drag to every client served

With AI, it becomes easy to imagine a future where consumers with a few thousand in savings – or even better, loans to pay off – have access to the knowledge, expertise of a UBS private banker by collapsing cost centers and lowering the cost to serve like:

• Automating intake: LLMs can ingest tax docs, generate draft investment profiles, surface inconsistencies, and explain onboarding materials in natural language

• Genuine personalization at scale: Clients can ask questions about their portfolio and receive context-aware answers that reflect their actual holdings, goals, and preferences. Beyond “what’s my balance?”, a client could ask as specific of a question as “Should I accelerate my 529 contributions given my child’s age and current market conditions?”. With genuine personalization, every client interface can be a n=1

• Compliance (surprisingly): Compliance is usually a drag on innovation but might LLMs flip that. They can read massive regulatory texts, extract what matters, and turn it into concrete actions. They can flag violations, map rules to a client’s case, or pull up the right disclosure or form, without pulling humans back into the loop

• Portfolio construction: AI can design portfolios around each client’s exact circumstances like tax status, liquidity needs, risk tolerance, and time horizon. This level of personalization was once reserved for the wealthiest because it required so much manual input but AI collapses that cost. Ideally, that construction sits below smart contracts that can execute automatically like trading, rebalancing, enforcing allocation limits, or stopping activity if some compliance rules are breached

A Few of our Assertions:

• Highest to serve categories will be first: Opportunity will flow where the cost to serve is highest. That includes labor-intensive, compliance-sensitive, and high-variance workflows like tax planning, small business onboarding, KYC, and loan servicing. In markets like SMB finance or wealth management, where the cost floor has long outpaced the revenue ceiling for low-ARPU customers, AI reopens previously uneconomical segments

Personal context becomes the moat: Everyone will have access to fine-tuned LLMs. What differentiates is the ability to apply them against a deep understanding of the user: financial behavior, product usage, life events, idiosyncrasies (so less like CRM and more like memory architecture). The challenge is designing a product with sufficient (and secure) mousetraps to acquire that data, a lot of which is unstructured or conversational (e.g., how stable is the client’s income, what tradeoffs is an impact-minded client willing to make)

III. Payments without People

Projecting the future of agentic payments requires some humility because the foundations like protocols or policy frameworks are still developing. But even in this early state, one thing is fairly clear: the current financial stack is insufficient.

The infrastructure has been (understandably) built around users i.e., humans with IDs, phones, and login credentials. However, agents will increasingly initiate, route, and settle financial activity on their own. They’ll run payroll, rebalance treasury, pay for compute, license media, and trigger purchases without asking for permission each time. This shift breaks much of the infrastructure underpinning almost everything prior waves have built.

The Unlock

Transactional agentic systems can unlock a world of value to consumers and enterprises. Imagine: 

• A treasury agent that monitors cash balances across bank accounts, stablecoin wallets, and DeFi protocols to automatically sweep funds to maximize yield

• Cross-border contractor payouts that bypass costly banking intermediaries by using stablecoins and smart contracts to handle multi-currency disbursement, real-time FX conversion, and local compliance logic

• Royalties and licensing bots in support of the creator economy that monitor digital content usage (e.g., a podcast snippet used in an ad), determine how much the publisher owes, and trigger automated remittance based on a pre-agreed contract

• A marketplace settlement engine that detects when a job or service has completed (e.g., tutoring or rideshare), calculates what the customer owes, and releases funds

Why Existing Infrastructure is (Mostly) Insufficient

To illustrate how infrastructure breaks down, consider an example. A research agent tries to pay $1 to access a paywalled article that it will cite to answer a user query or draft a report. The transaction should be simple i.e., send funds and retrieve content. But it fails because:

• Identity checks assume a human buyer: Many sites require verification like two-factor SMS, CAPTCHAs, or a billing profile tied to a verifiable human. Agents have no phone number, no government identity, and no way to pass a checkout flow built for people

• Fraud systems look for human behavior: Payment fraud tools rely on patterns like mouse movement, IP address history, or saved browser sessions. An agent operating via API or script looks suspicious even if the intent is valid

• Settlement is slow and uncertain: Agents make decisions in real time, but settlement forces them to wait hours (or days) for “pending” to become “complete.” Without instant confirmation, they can’t reason about the next step or retry intelligently

• Microtransactions don’t clear: A lot of payment rails charge a fixed fee (~$0.30) plus a percentage. A $1 transaction loses 30%+ to fees. That kills use cases like agents paying per query

Two Philosophical Camps

So far, most innovation in response falls into two philosophical camps:

1. Bridge the existing stack. Give agents cards or virtual identities. Prompt the user to re-enter at key checkpoints (e.g., authorize a checkout). Essentially layer bots on top of the legacy system

2. Go native. Use stablecoins, wallets, and smart contracts. Make the money programmable. Avoid intermediaries entirely

There’s a strong appeal to starting fresh with agent- and stablecoin-native system. But that can obscure the very real strengths of existing infrastructure:

• Card networks offer global acceptance and merchant density. They don’t actually need to run on card rails e.g., Visa can settle in USDC

• Banks still provide compliance, custody, and regulated float all of which are useful even if the transaction is initiated by an agent instead of a person

• Some heuristics still work. Fraud and risk engines are built for human behavior, but they’ve been trained over years of data. In some corridors or transaction types, they’re still better than starting from zero

• Not all incumbents are asleep and should not be dismissed. Visa is launching agent-native APIs. Cloudflare is working on bot identity standards. Some of the best infrastructure may come from the legacy players adapting

Instead of a binary, we think of each layer deserves its own consideration i.e., which components need to be reimagined natively, and which can be abstracted or bridged? A few opportunities that jump out:

• Wallet architecture needs to evolve. Most U.S. wallets still ride on card rails and assume a human in the loop. Agentic wallets need to persist identity across sessions, verify intent between query and transaction, and route payments across cards, RTP, stablecoins, and ACH without manual input

• Micropayments do not work on old infrastructure. A $0.05 transaction cannot clear when the minimum fee is $0.30. Agents will trigger thousands of tiny payments for APIs, compute, and content. This needs stablecoin-native, instant, low-cost blockchains

• Verification has to change. CAPTCHAs and SMS codes block agents. The ecosystem will need agent-native ways to prove intent like cryptographic signatures, delegation rules, or policy-based constraints

• Trust cannot rely on KYC. Trust will need to be built on things like behavioral history, audit trails, or smart contract attestations rather than traditional identity checks like two-factor SMS

• Controls will be policy-based. Just like managers set limits on employee cards, users will need to set rules for agents like daily limits, transaction types, or guardrails that balance cost and risk

• Agent underwriting needs new models. If agents are spending or investing money, they need to be underwritten without scores that depend on human inputs

• For agents to move from simple payments into, for example holding private assets or managing diversified treasuries, they need assets represented programmatically i.e., as tokens. That unlocks a new layer of infrastructure, each piece a startup opportunity like: collateralization that let tokenized assets back loans or credit, marketplaces that match buyers and sellers across assets, sourcing and onboarding of off-chain assets like invoices or private equity, valuation and pricing services that bring transparency to illiquid markets, compliant and transfer platforms, and secondary liquidity venues to keep assets tradable

A Few of our Assertions

• The consumer credit card rewards structure does not persist in an agentic future. Rewards and interchange are predicated on human psychology like revolving debt, brand loyalty, and spending nudges. None of that applies to agents who will need new “top-of-wallet” incentives

• As agents act independently, apps won’t be for clicking so the UI becomes more of a control panel than a locus of transactions. It feels reminiscent of how luxury brands see flagship stores, which have become less where most purchases happen and more where the company builds brand and loyalty

IV. StableFI

A Tectonic Shift

Stablecoins, as digital tokens backed 1:1 by traditional currencies like the dollar that are designed to hold a stable value, make a similar leap. Just as the ledger abstracted money from physical settlement, stablecoins abstract the dollar (or euro, yen, etc.) from the banking system, effectively turning it into a programmable unit that can move without intermediaries or institutional gatekeepers (Circle aptly called this a “HTTP for money” in its S-1). They retain the properties of money, execute with the precision of software, and plug into systems that were never designed to interact with banks.

Enumerating the constraints it removes (or opportunities it unlocks) risks underselling the shift, the same way listing the benefits of double-entry bookkeeping would. But a few deserve to be made explicit:

• Settlement does not move in real-time. Traditional systems like netting, batching, end-of-day processing don’t match the cadence of modern liquidity. Even same-day ACH does not align with how capital wants to move

• The demand for commerce, savings, and liquidity is global, but access to banking infrastructure is not. A merchant in the Global South, for example, might want to sell goods online but cannot easily accept payments from buyers abroad because many platforms do not process local bank transfers. Stablecoins provide a universal alternative as a way to pay that works across borders without relying on whether a platform has integrated the local system

• Fintech’s first wave abstracted complexity but never removed it. Stripe still resolves to the card networks; or, banking-as-a-service still resolves to core processors. Stablecoins are inherently composable: they move inside smart contracts, across protocols, and between wallets without external reconciliation

• The financial stack is still organized around humans. Most money systems presume a human actor with a phone, ID, and intention. But the fastest-growing category of transactors will be non-human. They require a programmable settlement asset. Stablecoins are (by far) the best candidate

Foundations: A Killer Wedge, and the Legitimacy Layer

Cross-border payments have long been described as the wedge, but calling it a wedge undersells the scope as global cross-border flows exceed $194t in 2024, according to FXC, and the infrastructure supporting them still depends on architecture designed in the 1970s.¹²

Cross-border flows:
As an example, a mid-sized U.S. company wants to pay a contractor developer team in Moldova, a country with no direct correspondent relationship with the company’s U.S. bank.

1. The company wires dollars from its U.S. commercial bank to a domestic correspondent bank

2. That bank initiates a SWIFT message routed to a E.U. bank that does hold Moldovan clearing relationships

3. The E.U. intermediary performs the FX conversion at an institutional spread of 2%

4. While the payment clears, the local currency depreciates further on some local news, which further erodes the effective value at neither parties’ benefit

5. The funds are routed through Moldova’s local clearing system and finally deposited into employee accounts … minus local receiving fees, compliance costs, and unexpected delays

This process takes 3 – 5 business days, spans four institutions, and presumes legal presence, pre-funded nostro accounts, and intermediary trust. Along the way, value leaks: FX spreads, wire fees, and correspondent charges can strip 5–10% from the transfer before it reaches the recipient. With stablecoin infrastructure:

1. The company holds USDC in a self-custody wallet or through a treasury platform

2. A smart contract issues payments to employee wallets 

3. Employees convert to Moldovan leu using off-ramp platform with real-time quotes and transparent rates

4. Settlement completes in minutes with logic attached

Legitimacy

Legitimacy matters. A stablecoin is an abstraction of fiat value — a dollar, euro, or real that derives credibility from market-wide acceptance and convertibility. That acceptance is now being validated in both directions: from large Western institutions and from consumers and enterprises in emerging markets where institutional infrastructure is absent or untrusted.

As discussed earlier, the largest foundational regulations are the GENIUS Act and MiCA framework. To summarize:

Recently passed with bipartisan support, the GENIUS Act creates the first comprehensive federal framework for payment stablecoins in the U.S. Key provisions include:

• Fully reserved backing: Issuers must maintain 1:1 reserves in high-quality liquid assets — such as U.S. dollars, insured deposits, or short-term Treasuries — similar to how Circle and Tether work, but now standardized and regulated.

• Reserve transparency: Issuers must publish monthly reserve reports and redemption policies, certified by both executives and independent auditors.

• Priority in insolvency: Stablecoin holders receive first priority claims on reserves if an issuer fails, offering stronger legal protection than traditional unsecured creditors.

• No yield or interest: Issuers cannot pay yield to holders, underscoring the line between stablecoins and securities or savings products.

• AML and sanctions compliance: Issuers are subject to the Bank Secrecy Act and must implement standard anti-money laundering and sanctions screening.

• Note: Crucially, the Act allows qualifying stablecoins to be treated as cash equivalents under U.S. GAAP which unlocks the critical corporate treasuries use case

Similarly, passed in 2024, the Markets in Crypto Assets (MiCA) regulation establishes the EU’s first framework for crypto assets, including stablecoins.¹⁴ Key provisions include:

• Licensing and oversight: Issuers of fiat-backed stablecoins (so-called “e-money tokens”) must be licensed by an EU national authority and supervised by the European Banking Authority.

• Reserve and redemption requirements: Stablecoins must be backed 1:1 by liquid reserves, with daily redemption rights and clear procedures for user reimbursement.

• Issuance limits: Large stablecoins face volume caps if deemed systemically important (note: not found in the U.S. framework).

Unlike the U.S., MiCA explicitly addresses crypto assets beyond stablecoins, covering asset-referenced tokens, utility tokens, and crypto service providers. But together, GENIUS and MiCA provide both clarity and legitimacy.

These developments from the Global North have opened the path for institutions to issue stablecoins, hold reserves on-chain, integrate them into treasury operations, and rely on them for cross-border settlement. Bottoms up signals are equally powerful:

• USDT and USDC supply now exceeds $220B.¹⁵ Tether and Circle are among the largest holders of U.S. Treasuries globally.

• Platforms like Yellow Card, Littio, DolarApp, Belo, and Nala use them to deliver remittances, payroll, invoicing, and savings across Asia, Africa, and Latin America

• Usage is uncorrelated with crypto speculation. As it should be because stablecoins are not speculative assets

The StableFi Stack

A simple card swipe or Venmo payment obscures a tower of infrastructure: KYC, AML, fraud, issuers, acquirers, ledgers, reconciliations. This stack is dense because the financial asset is inert i.e., the rules that govern how it moves (who can access it, when it settles, what it's for) live outside the money, enforced by banks and payment processors. Stablecoins invert that because they allow developers to encode logic, access, and flow at the asset level. This allows us to imagine exciting use cases like:

• Lending run by code i.e., originated, collateralized, repaid, with automated margining and open liquidity

• Global commerce that settles instantly with embedded FX and tax logic

• Remittances that behave like APIs, not wire transfers

• Payment rails that do not assume a human with a card at the point of sale

But moving rules into the asset itself also raises new challenges. The most important open questions include:

• Risk: What constitutes fraud when transactions are irreversible and identity is optional?

• Identity: Can stake, history, or cryptographic reputation substitute for legal identity?
  Orchestration: How do we bridge stablecoins, local rails, FX endpoints, and compliance logic?

A Few of our Assertions

• Corridors still matter. Even if ledgers become universal, the plumbing beneath like compliance, off-ramps, liquidity, and trusted interfaces remains stubbornly local. Brazil-to-Mexico is not the same as US-to-Japan, because what clears legally, converts locally, and earns user trust varies market to market

• There will be many more stablecoins. The current landscape is vertically integrated. But issuance will decouple. Brale, for example, lets banks, corporates, or fintechs issue their own stablecoins with compliance, chain support, and attestation infrastructure built in. That unlocks corridor-specific, function-specific, or even internal-use stablecoins for loyalty, payouts, and treasury

V. Post-generic Fintech

Many of the largest winners of the last cohort of fintech companies were generic, by design. Stripe, for example, turned payments into an API. Or Plaid turned bank data into a universal access layer. Their value came from abstraction: one-size-fits-all infrastructure that moved offline workflows into digital form. But for large swaths of the economy, generic is a powerful foundation, but does not suffice. Financial needs are often specific enough that horizontal tools break down. The next wave will be built on context: platforms designed for the specific payment flows, frictions, and economics of each industry, community, or use case.

One useful lens is a two-by-two of relative transaction volume (TPV) and payment complexity to identify where horizontal tools break down and vertical solutions create the most leverage. High TPV does not always map to high revenue or headcount. Consider a solopreneur reseller: modest income and no FTE force, but massive payment throughput by virtue of reselling flow. Payment complexity comes from items like:

• Multi-player transactions: A simple payment, let’s say a buyer paying a supplier or a consumer paying a merchant, is typically straightforward. But in many domains, transactions engage multiple actors with different rules, timing, and entitlements. Healthcare payments may combine out-of-pocket expenses, HSA balances, and insurance reimbursements. Tuition often blends grants, parent contributions, and student loans. Construction or hospitality payouts can involve layers of subcontractors. Each added stakeholder multiplies the complexity and creates room for vertical platforms that can orchestrate the flow

• Rewards and incentive mechanics: Generic platforms usually default to blunt instruments like flat cash-back, universal points, broad perks. Vertical or context-driven platforms can go further by maintaining industry specific rewards that actually shape behavior in their domain. These domain-specific mechanics create stickiness and value in ways a 1.5% cash-back card never could

• Regulatory overlays: Industries like healthcare, education, and financial services carry their own rulebooks on top of standard payments compliance. HIPAA, Title IV, or banking charters shape not just how money moves but who can move it, when, and with what records. Vertical platforms win by baking those rules into product and effectively turning regulation from a barrier into a moat

• Timing and settlement mismatch: Payment flows often do not line up neatly with the underlying activity (advance deposits, milestone payouts, clawbacks, etc.). Contextual platforms with deeper visibility into the workflow can align liquidity with what is actually happening on the ground. For example, in freight and logistics, a platform can advance funds against a confirmed bill of lading rather than forcing a carrier to wait 60–90 days for invoice settlement. In construction, a platform can release milestone-based payouts once progress is verified, instead of tying subcontractors’ cash flow to generic net-90 terms.

A Few Examples

Several industries show how high volume and complexity create fertile ground for specialized platforms:

• Healthcare: Few industries have payment flows as tangled as healthcare. A single procedure can involve patients, insurers, employers, providers, and labs, each with their own reimbursement logic, timelines, and compliance obligations. That makes even routine payments a reconciliation problem. And while there are incumbents, the opportunity keeps expanding: new payment methods like HSAs and FSAs, new regulation such as the Medicare Prescription Payment Plan in the Inflation Reduction Act, and new care models like telehealth and at-home testing. Nitra, for example, addresses the practice spend side with products like healthcare-specific rewards (e.g., 5% back on medical supplies), procurement tools, bill pay, expense management, and a verticalized supplies marketplace

• Hotels & hospitality: Funds are often held months in advance, then released after check-in, with clawbacks for cancellations or no-shows. Add to that the need to split payouts between a franchise owner, local operator, and the brand itself, each with different contractual rules. Escrow, timing, and revenue-sharing turn hotel payments into a working-capital and reconciliation problem that generic platforms do not solve

• Immigrants & global freelancers: A feature of the last wave was demographic-specific fintech products (e.g., neobanking for X or Y cultural group). In hindsight, the winners were less those that tapped into cultural affinities as much as platforms where the demographic’s financial needs mapped cleanly onto distinct products. That lens fits immigrants and freelancers especially well who often straddle multiple financial systems: they need accounts in more than one country, fast and affordable remittances, and fair FX rates. Traditional banks make this slow and expensive; informal channels make it risky. Vertical platforms that bundle multi-currency accounts, local access, and trusted remittance rails can turn today’s patchwork of fees and delays into a single product. Companies like Comun and nSave show that demand is already there.

The post-generic platform confers several more holistic advantages. To name a few:

1. Better acquisition: Generic platforms often default to generic acquisition. A neobank trying to serve “everyone” might spend on low-conversion billboards off Madison Avenue because its product could, in theory, apply to any consumer or enterprise. Post-generic fintech takes the opposite approach: it meets customers where they actually operate like an industry trade show, a specialized SaaS marketplace, or a community forum. And because the product is built for their workflow, the features speak directly to them, not just to anyone. 

2. More monetization levers: Vertical players can usually monetize across more layers than their horizontal peers. That often starts with a SaaS subscription for industry-specific financial management or planning. On top of that comes a payments take rate, lending and credit products underwritten with domain-specific signals, and financial products that only make sense with industry context. Slash is a good example: by focusing on verticals like web3 companies, marketing agencies, and ticket resellers, it has been able to add products outsiders wouldn’t think to build. For agencies, that means virtual account creation and management across hundreds of clients. For crypto companies, it means a unified account to manage fiat and digital assets side-by-side, including instant swaps between the two.

3. Better foundation for AI: Vertical depth creates richer data than horizontal breadth. A healthcare platform doesn’t just see “payments” but also claims, reimbursements, and benefit use tied to procedures. A freight platform captures shipment confirmations, routes, and invoice timing. With this level of context, vertical platforms can build products that go beyond generic credit scores or expense trackers, offering decisions and recommendations grounded in the realities of each industry

A Few of our Assertions

• Opportunities cut across consumer as much as enterprise. The word “vertical” often connotes enterprise SaaS, but many of the most compelling opportunities are consumer-facing where financial needs are just as context-bound

• Vertical software is a powerful wedge.  Embedding into core workflows creates sticky distribution, but the bigger opportunity is financial. Once a platform controls the workflow, it naturally sits in the flow of money like vendor payments, employee spend, patient billing, contractor payouts. That position allows it to layer on payments, credit, insurance, and other financial products with higher margins and stronger retention than software alone.

VI. Geography of Opportunities

Fintech has always been disproportionately global. Unlike enterprise SaaS, which tends to cluster around concentrated buyer bases (Fortune 500, developer ecosystems, Bay Area tech), fintech demand is tied more to GDP than to enterprise budgets. When you sit in the flow of payments, credit, or working capital, your TAM grows with economic activity, not ACV.

That logic played out in the last wave of public and late-stage companies — from consumer platforms like Nubank, Revolut, Monzo, Klarna, Mynt, Paytm, Toss, Kaspi, and Ualá to cross-border processors like dLocal, Rapyd, and Airwallex, or infrastructure players like Thunes, Flutterwave, and Mambu.

The same wave also created new fintech talent hubs outside Silicon Valley and New York. Ecosystems like London, Stockholm, São Paulo, Bangalore, and Singapore now have prior outcomes that can seed the next generation. According to Harmonic, there are 122 ex-Revolut founders in the United Kingdom, 88 ex-Nubank founders in Brazil, 153 ex-Klarna founders in Sweden, and 231 ex-Paytm founders in India.¹⁷

These hubs also highlight how fintech adapts to local conditions — and why the next wave will again be shaped by geography. A few examples:

• Borderless FX infrastructure in Europe: Companies like Wise and Revolut emerged because the rest of European consumer life like messaging, e-commerce, mobility was cross-border, while payments remained national

• Real-time payments in Brazil and India: Pix and UPI are national, API-driven settlement layers with universal reach, zero marginal cost, and round-the-clock availability. Both systems now surpass card volume domestically — and have unlocked second-order innovation like NG.Cash, a consumer fintech platform with more than 3mm accounts natively built on top of Pix

• Global BNPL variants: The Buy-now-pay-later wave did not start in the U.S., and actually makes more structural sense in markets without high credit card penetration or rewards infrastructure. Klarna’s growth, first in Europe, for instance (in Sweden ‘klarna’ is colloquially a verb as in ‘to pay in installments’) has been built on a tight flywheel between merchant integrations, retail media and discovery, and proprietary underwriting 

• E-invoicing in Latin America: Regulators in LatAm mandated electronic invoicing, giving licensed startups permissioned access to tax-verified income and transaction data. That leveled the playing field with banks, who long held an information advantage, and let younger companies build higher-fidelity underwriting models on equal footing with incumbents

• Experimenting with new interfaces: In markets like Brazil, India, and China, fintech is increasingly delivered through messaging apps. India has integrated UPI into WhatsApp, while China proved the model with WeChat Pay. These interfaces bring payments and banking into the apps people already use, closing the gap between intent and transaction.

Collectively, these tectonic shifts across rails and applications usher in a new chapter. The only constant is change. Reach out to continue the conversation.