Can Neobanks Break Traditional Banking? | Roundup

Neobanks built on stablecoin rails were identified as an obvious winning category precisely because traditional banks will not build on stablecoin-first infrastructure, since doing so would directly undermine their own settlement advantages. Traditional banks are the only entities in the traditional financial system that can formally move money and settle, and speakers argued they will not willingly disrupt that position. Crypto is framed as directly disruptive to banks not only in income streams but in their strategic position at the base of the financial value chain. Morgan Stanley rolling out crypto trading on E-Trade was cited as a sign that incumbents are becoming much bigger players, even as their structural incentives prevent them from going further.

Revolut was described as the clear winner in fintech and crypto in Europe, with eight out of ten Europeans citing it as their favorite app, and it already allows users to buy crypto and send it directly to a wallet address. Speakers said there is still room for more companies like Revolut, including ones that allow activities beyond basic banking. The end state described for the DeFi mullet thesis is neobanks serving as a consumer front end with DeFi protocols on the back end, with neobanks predicted to become a major distribution channel for DEXes and borrow-lend protocols. Speakers said it will take years for a couple of neobanks to win and achieve scale, but that outcome represents the light at the end of the tunnel for broader DeFi adoption.

Large funds are currently being raised to do growth rounds for neobanks that have emerged as winners from a prior competitive cycle, with Brazilian neobanks cited as a model, growing from roughly a hundred million dollar valuation to ten billion. A portion of those same funds is being allocated to early bets on AI as the next race, though the AI fintech investment phase is described as still early and confusing with no clear winners. Not all stablecoin-based neobanks currently being funded will succeed, and speakers said only a few will win. Crypto products have historically suffered from poor design and weak user experience, partly because selling into a fragmented market did not require great products and because early token launches removed the incentive to build actual product value. The prediction is that crypto neobanks will close the product quality gap with Mercury, Brex, and Ramp within the next 12 to 24 months, though many current crypto teams are not expected to survive increased competitive pressure.

Stripe will push blockchain infrastructure including Tempo onto its customers from the top down, leaving partners little choice but to engage with the technology. A gap was identified for a more open and interoperable alternative to Tempo, which speakers described as still relatively closed. Lightspark, which launched with offerings to 65 countries, card access, and infrastructure running on a Bitcoin L2, is described as the closest thing to Tempo currently available. Base was named as a candidate to become the chain for AI agent activity if app distribution rather than technical differentiation determines chain selection.

On chain-level specialization, speakers noted that the general-purpose approach to blockchain go-to-market has failed and the current trend is to win one specific use case first. Solana has leaned into trader culture with upgrades oriented around becoming effectively a DEX chain, while borrow-lend is dominant on Ethereum. No blockchain has successfully targeted AI agent coordination as a specific niche despite many attempts. Most investments in AI and crypto infrastructure are not expected to work out well, with value accruing instead to applications and APIs using crypto rails. Bittensor was cited as one of the few AI blockchain infrastructure projects that has actually worked.

Brex and Ramp are among the fastest-growing companies in the US, and Ramp is now evolving into a full B2B banking service competing with traditional banks. Business banking for US customers still lacks adequate fintech solutions, particularly for cross-border and cross-jurisdictional needs. Traditional banks are seen as unlikely to deploy AI agents to deliver faster and more efficient services, whereas fintechs can. The broader prediction is that traditional financial services will shrink while fintechs including Stripe and agent-based services grow. The current unbundling moment in fintech is expected to be followed by significant consolidation in approximately three years.

A founder observation from Miami noted that mentioning both AI and crypto to a traditional business makes them very scared, but that using the term stablecoins instead of crypto is more effective. Speakers argued that DeFi was never really designed for individual retail users but rather as rails for companies to build on top of, and that the current crypto cycle may be the first where a meaningful number of real companies are in the scale-up phase. Most crypto CEOs are described as having no prior experience managing a business at that scale, which represents a distinct risk as competitive pressure increases.