Sheel Mohnot on tariffs, Circle's IPO, and why early-stage VCs should keep deploying through the chaos

Summary

Sheel Mohnot, general partner at Better Tomorrow Ventures, argues that competition strengthens industries while protectionism weakens them. Growing up in India before the 1990s liberalisation, he watched domestic car manufacturers sell 1950s-era vehicles at high prices with no incentive to improve. The Jones Act produces the same dynamic in US shipbuilding today, where vessels cost five times their international equivalent.

On Trump's tariffs announced in April 2025, Mohnot is direct. The administration framed these as "reciprocal" tariffs but the numbers tracked the US trade deficit rather than foreign tariff rates, revealing that this is industrial policy, not market reform. Commerce Secretary Lutnik's argument that tariffs will reshore manufacturing and replace income taxes makes the intent explicit. Mohnot rejects the softer "negotiating leverage" rationale as well.

The trade-deficit focus also ignores services. Switzerland runs a goods surplus with the US partly because its 8 million people make fine watches, but those same people are almost certainly Netflix subscribers. Digital services exports go uncounted in the tariff debate.

Mohnot draws a distinction between blanket tariffs and China-specific measures. He accepts the national security case for China action—DJI drones collecting US data, TikTok's data exposure—but notes these concerns have nothing to do with tariffs on allied partners.

His preferred model for competing with subsidised Chinese manufacturing mirrors the EV playbook. Government demand stimulus and subsidised loans beat import barriers. US taxpayers lent Tesla $500 million and Tesla competed on innovation. A similar drone-buying programme could have backstopped GoPro's manufacturing without distorting broader trade. The CHIPS Act and IRA point in this direction. A carrot approach beats a stick approach.

Circle's IPO remains the live test case now that market volatility has closed the broader window. Klarna pulled its filing and StubHub stalled. Polymarket estimated Circle's 2025 IPO probability at 86 percent at the time of recording.

Mohnot's read on the S1 filing is measured. The headline criticism is real—Coinbase captures roughly half of Circle's revenue through its USDC distribution agreement. But he has no strong conviction on IPO performance because public market pricing at launch tends to disconnect from fundamentals. Circle could easily have caught a stablecoin hype wave if it had been public during the Bridge or Stripe acquisition period.

The bull case is structural. Tether is the most profitable financial institution per employee in the world by some measures but remains opaque and faces persistent accusations of reserve mismanagement. USDC sits at a $60 billion market cap as the second-largest stablecoin and the dominant regulated Western alternative. First Digital USD is under $2 billion and PayPal USD is at $800 million. Circle has roughly 80 times the circulating supply of its nearest regulated Western competitor. For public market investors seeking stablecoin exposure through a transparent, regulated vehicle, there is no cleaner option.

Circle's narrative problem is real. CEO Jeremy Allaire is sharp—he took a company public during the dot-com era and worked at General Catalyst before founding Circle—but the company hasn't broken through in ways that make its story legible to generalist investors. That cuts both ways. It could trade as a meme stock on crypto sentiment, or it could be underappreciated until the stablecoin regulatory framework clarifies.

For early-stage investors, Mohnot's position is straightforward. Keep deploying. Better Tomorrow raised its fund on a multi-year cycle, so LP mark-to-market volatility today doesn't affect capital availability for the next few years. Every cycle since 2020, including 2022 when the denominator effect was supposed to freeze seed investing, shows that funds that stopped deploying missed the vintage.

He is sceptical of the claim from Semil Shah at Haystack that seed will become hot again with median entry prices up 50 percent in the next vintage. Seed valuations have risen in a near-straight line since 2021. Large multi-stage funds writing $2 million to $3 million seed checks tend to be poor seed investors anyway because backing the wrong company in a category costs them the option to write a $250 million check into the right one later. Portfolio companies are already receiving opportunistic outreach from investors trying to use market uncertainty as a re-entry point.

The later-stage picture differs. Growth and late-stage funds face an IRR clock, and a frozen IPO market compresses returns in ways that early-stage managers do not face on the same timeline.

On whether AI can materially improve loan underwriting, Mohnot sees genuine potential but a specific regulatory ceiling. The Equal Credit Opportunity Act and Fair Credit Reporting Act require lenders to issue adverse action notices explaining denials in human-readable terms. A black-box model that outputs a matrix of weights is not legally compliant. Without those constraints, AI-driven underwriting would likely be more efficient at targeting creditworthy borrowers. With them, step-change improvement is harder to achieve. Better Tomorrow has invested in a company focused on detecting bias in AI underwriting models to help lenders stay compliant. The near-term opportunity sits in tooling around the regulatory boundary rather than replacing the underwriting stack outright.

Insurance constraints are less binding. Root Insurance's approach of using phone sensors to measure driving behaviour outperformed Metromile's simpler mileage-based model. Mohnot sees continued room for AI to improve risk pricing in insurance.

The incumbent-threat meme is less potent in fintech than in consumer software partly because large financial institutions have a poor track record of shipping competing products. Stripe built a corporate card before Ramp existed, couldn't make it work, and ended up investing in Ramp and deprecating its own product. The robo-adviser category is the clearer cautionary tale. Wealthfront and Betterment disrupted fee structures but the 25-basis-point product was easy enough for Vanguard to replicate. Vanguard became the largest robo-adviser in the world. For founders, thin-margin commoditised fintech products face genuine platform risk while differentiated complex products generally do not.