Why Fundamentals Are Finally Winning and Where This DeFi Investor's Putting His Money | The Thesis

Investors are increasingly looking beyond meme coins toward projects with usage, revenue and credible value accrual, though the speakers do not rule out large gains from memes. Kyle describes a K-shaped market in which roughly 90% of investments fail to earn lasting backing while the strongest 10% rise to the top. Meme tokens can surge as buyers crowd in, but he sees little trading edge or fundamental basis for expecting them to keep growing. In his view, the shift toward fundamentals is visible in demand for protocols that generate revenue and return value to token holders.

Hyperliquid is Kyle’s clearest example of a token whose real activity has supported an “up only” chart: he points to its fees and revenue, alongside a 99% buyback, and says dollar-cost averaging would have worked. The speakers expect more projects to adopt value-returning practices as investors demand growth and better alignment. Kyle favors concentrated positions, contrasting 10–20% allocations he says Arthur at Defiance commonly held with funds that often cap positions at 5%. He joined Defiance Capital in 2025, after the 2022 3AC collapse, and says he does not know the full details of Defiance’s relationship with 3AC.

Kyle also entered the Ethena/Stablecoin X DAT trade after Evan’s tweet drew his attention to it and he saw it trading well below MNav. The speakers describe the structure as a way to give public-market investors ENA exposure while adding utility for Ethena’s digital dollars; native yield could help cover administrative fees. More broadly, they see DATs as a way to create value beyond simply buying an altcoin, by attracting traditional investors and adding business lines or utility. They regard the equity-funded, buy-more-Bitcoin model associated with Michael Saylor as one-dimensional and best suited to Bitcoin-like assets.

Derive is another example of fundamentals gaining traction: the speakers say it is taking market share from Deribit and credit founder Nick Forster’s long-standing work on on-chain options. They describe notional trading volume as up about fivefold year over year, with fees and revenue also rising sharply. One investor says experience across roughly 70–80 investments has made the founder and team more important than the initial idea. For them, “getting it” means combining determination, intellect and practical skills with the business sense to build and operate a real business.

The investor sees AI inference as the most promising opportunity in the AI stack, citing Jevons’ paradox and wanting long exposure to rising demand. They distinguish this broad thesis from any endorsement of the “AI money” argument for proof-of-work tokens such as Pearl, Knockchain and Tao. Their clearest expression is Venice’s VVV token: staking earns more VVV, while locking VVV creates DM, with each DM providing one dollar of Venice usage per day indefinitely. Venice revenue burns VVV, though the investor describes the buyback effect as small. They see usage growth, including charts shared by founder Eric Voorhees, as central to the case; Venice’s privacy offering may also gain value as frontier models restrict some responses. But the VVV relationship to Venice equity remains unclear: one speaker says an equity raise coincided with a price dip and left uncertain how value is divided among equity, VVV and DM. A fundamentals-focused speaker says the structure is not clean enough and does not currently hold VVV.

The investor also bought Orbeo at an $8 million market cap, prioritizing its founder’s track record over the initial idea. Orbeo began as an inference marketplace for selling unused capacity and has since added a launchpad. The investor imagines agents launched there buying inference through Orbeo, but says that broader plan is an inference, not something confirmed in a conversation with the founder. They describe the project as a small on-chain token with a novel product and strong founder, note its market cap was below $100 million and later below $80 million, and believe it could reach at least $500 million. Their highest-priority themes include inference, tokenized stocks, DATs and privacy; perp DEXes and stable points sit in the next tier. They favor product-market fit and growth linked to TradFi or other durable demand, while acknowledging privacy has a weaker outside-crypto growth case. They also cite Backpack’s long-term-oriented founders and say it has risen roughly four or five times from its bottom.

The speakers see signs of an inflection in fundamentals, while allowing that the market may still be early. Hyperliquid’s perpetual volumes are growing month over month and year over year, and tokenized stocks are receiving renewed attention from banks and BlackRock despite earlier precedents such as Mirror in 2020. The investor has mostly stopped angel investing, seeing enough opportunities in liquid markets and preferring to buy and hold projects with sustained growth rather than repeatedly time trades. They believe the bear-market bottom is behind them based on sentiment, price and adoption, but acknowledge a brief bull market could be followed by another downturn. At the time of recording, Bitcoin was around $84,000; a fall to about $58,000 would challenge that view. October looked pivotal to their outlook, and they warn that a sharp Bitcoin decline could pull the broader market down, even as altcoins might perform well with Bitcoin near $90,000–$100,000 or without a new all-time high.