After AI Steals the Talent, Only Six Tokens Are Left to Own
Delphi's Yan Liberman says crypto VCs make one deal a year because AI has sucked up all the smart people; the only things that can still make money are a handful of tokens with real revenue that don't depend on the crypto market.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
This rally has no obvious buyer
Yan says the market was caught off guard by what he calls "QE in the form of long-end bond curve control," which triggered a short squeeze, but he personally finds it hard to see new highs in the short term, because that requires a lot of new money, and the obvious historical buyer (Sailor) isn't buying now — he's cashmaxing and paying down preferred stock. He thinks leverage hasn't built up excessively, and that this sequence of slow grind, pause, alt rotation is actually healthy. Another structural change is that the alt market has gotten shallower, so capital concentrates more easily into a few assets, meaning you don't need a whole new cohort of buyers to move the market — it has become a "stock picker's market."
— Yan LibermanBuy the ones that look like hot coins but aren't as crowded
Yan admits there are only about 10 to 15 investable liquid tokens right now, and every crypto fund is already positioned in those 15. His conclusion: you should look for names that "look and smell like these coins, but aren't as crowded to trade." He's long-term bullish on Hyperliquid, but would be surprised if it went to 200 within four months; for short-term excess returns you have to look at less obvious names. Santi adds that with 15 investable coins, any liquid fund going on a podcast keeps naming the same five.
— Yan LibermanCrypto venture does one deal a year
Yan says only one of their newest venture fund's five investments is crypto, and over the past year they did just one crypto venture deal. The reason is that value capture and defensibility in crypto venture are now clearer, so you can screen out a large number of projects quickly, but the thesis that "crypto gets broadly adopted across many verticals" has been pushed to the sidelines, and AI is a major factor in drawing away talent, attention and capital. His one crypto investment is Tori Finance, a tokenized Turkish lira carry trade, dollar-hedged, with about 10.5% organic yield and no incentives.
— Yan LibermanThe six tokens share one thing: revenue that doesn't come from crypto
Yan's six public tokens are Hyperliquid, Zcash, Lighter, Venice, Grass and Aerodrome. He says he prefers projects with a "revenue story" that don't depend heavily on crypto activity, because that makes them more resilient and gives them a stronger TAM story. He goes deep on Grass: an estimated $70 million in revenue, $40 million in operating profit, doing residential IP network data scraping for frontier labs as customers, a token market cap of $330 million, unlocks finishing in October, about 30% of supply in the foundation, no equity, only the foundation — and if you valued it as a pure-equity non-crypto business, it could probably raise $1 to 2 billion in the private market.
— Yan LibermanCrypto's return requirements have changed
Yan says he no longer anchors on annualized return rates, but instead sets individual price targets, stop losses and invalidation conditions, and keeps a reserve of cash for a "total collapse" scenario. He admits he just likes volatility. Santi says he now won't do a crypto venture deal that can't at least 10x, that he can't see a clean 10x in Ethereum at a $300 billion market cap, and that the downside asymmetry is far larger than the upside. Yan responds that comfort comes from holding "productive assets" rather than "reflexive assets," and that the key to a productive asset is the durability of its earnings and how dependent it is on reflexive assets.
— Yan LibermanThe gaming vertical has been abandoned
Yan says gaming as a whole is a hard vertical to invest in, because games as a business have too short a life cycle — once you run the NPV, the long-term value story is hard to sustain. They tried Gunzilla, a big bet, and it didn't go well; he attributes that to management and spending issues, not game quality — the game itself is AAA level, on par with Call of Duty, PUBG and Fortnite. He says this may be their last attempt at AAA level, and that gaming as a broad category will be hard after this, with only smaller-scale games doing well. They now think the only truly investable vertical left is basically "financialized applications."
— Yan LibermanAerodrome is a classic catch-up trade
Yan says Aerodrome had planned to migrate to mainnet in mid-July and support multiple chains, but it was delayed, which hurt the price; the release is now set for the second half of September, with migration to ETH and ARC announced, and whether it migrates to Robin Hood unconfirmed. He sees this as a huge opportunity for TVL, volume and fees, and many people treat it as a catch-up trade. The prior problem was the relationship between emissions and revenue — emissions go to LPs, revenue goes to stakers, which drew criticism for offsetting revenue with emissions. Now they want to make emissions dynamic, tied one-to-one to trading volume, so price up means emissions down; at the same time they're adding multiple revenue lines, estimated to contribute an extra 40% of revenue with no emissions offset. Currently OI is not a large position relative to market cap.
— Yan LibermanThe investment committee now weighs founders most
Yan says if he could pick only one change, it's the weight given to founder due diligence. In early crypto it was obvious what needed to exist, so you invested more on the idea and defensibility, with lower founder weight. But once you move into other professional verticals, you can't really diligence the idea and the moat — the only repeatable, scalable thing is the founder side. They now ask, "What's your earliest memory, and then walk us through your life." He also mentions the "chip on the shoulder" from childhood trauma and the desperate desire to win as factors that are genuinely valuable, and gives Athena's Guy as an example, saying he will do whatever it takes to win.
— Yan LibermanIn their own words · checked verbatim
I personally have a hard time seeing new all-time highs anytime soon without some material new form of QE.
Yan Liberman2:02
I think the setup is there where there's appetite to look for compelling alts that that haven't really been bit up and and I think that's where the real opportunity is.
Yan Liberman5:06
Even our venture fund right now for the most recent one we have five investments and only one of the five is in crypto. Um and I think we've done one venture crypto deal in the last year I'd say.
Yan Liberman7:08
that business is doing 70 is expected to do 70 mil topline in revenue and and and based on kind of what they've suggested it's 40 million in operating profit.
Yan Liberman20:20
So it's it's basically my my take is that it's it's uh regulatory ARB on sports betting.
Yan Liberman34:36
the only repeatable one that you can that scales is the founder side.
Yan Liberman45:50
it's extremely hard for me to believe that we're going to see a trillion dollar business, not a commodity in crypto.
Yan Liberman53:55
Figures
| Grass operating profit | $40 million | 20:20 |
| Grass residential IP nodes | 8 million | 22:22 |
| Grass previous funding round valuation | $500 million | 22:22 |
| Aerodrome new revenue line contribution | an extra 40% of revenue | 38:41 |
| Number of emerging managers Delphi has engaged | about 500 | 42:46 |
| Number of emerging managers Delphi has invested in | 7 to 8 | 42:46 |
Glossary
- cashmaxing
- Refers to MicroStrategy's Sailor no longer buying bitcoin, instead holding cash and paying down preferred stock.
- reflexive assets
- Assets whose price decline causes fundamentals (such as trading fee revenue) to deteriorate in tandem, as opposed to productive assets.
- regulatory ARB
- Using differences between jurisdictions or regulatory classifications to gain an advantage, such as prediction markets being taxed on trading revenue rather than gambling revenue.
- emerging managers
- Fund managers raising their first or second fund; Delphi invests in them through a fund-to-fund approach.
How to listen
Crypto fund managers, token investors, and practitioners who want to understand how a frontline crypto VC allocates capital and evaluates founders.
The ad segment from 16:17 to 17:57 can be skipped.