From Poker Pro To Trading Giant: Inside The Mind of Selini Capital

Summary

• Retail has left crypto altcoins, and Jordi says the flows prove it: ticket flow is down “massively,” diffused into equities (Micron, SanDisk), sports betting, and prediction markets. The tell was Grass — the team told investors it was making ~$50M cash and the token fell a third in a day. “What kind of shitcoin goes down by a third in a single day on good news?” The culprit is structure: Binance’s ~year lockups leave “a bunch of VCs that are down bad” selling whatever has liquidity until retail gives up.
• Selini is now trading the crypto version of equities — SpaceX was doing $2bn/day on Binance — across Binance, OKX, Lighter, and Bitget, hedged on the tradfi side. Jordi’s claim: “it’s not that hard to find alpha in SpaceX,” because the crypto playbook transfers wholesale — KOLs (Elon, CNBC) “bullshitting some story,” a “ridiculous amount of unlocks” in August, the next tranche, the earnings call two weeks later. “We’ve seen this in crypto. We know how this plays out.”
• Passive flows are the designated bag-holders and they’re predictable: the Nasdaq is about to pile money into newly listed, low-float SpaceX (not in the S&P 500, which Jordi puts at 13–14x Nasdaq’s size) — possibly the first time passive gets “exploited to such a degree.” Millennium’s $2.8bn, $3.7bn, and $2.7bn figures on index rebalancing are the same trade; crypto’s version was front-running Michael Saylor, “in essence the NASDAQ” of dumb money.
• Crypto has made a “choppy bottom” with zero momentum, and the AI-to-crypto rotation has to play out slowly — people must first exit SpaceX, memory stocks, and KOSPI winners. When the move comes it will be sudden and reflexive: “it’ll all happen for no reason… why are we up 10% again? Because yesterday we were up 10%.”
• October 10 cost Selini single-digit millions — arb teams “did spectacularly well,” but accounts running 3–4x leverage (to lift a 10–15% fully-collateralized return toward 60%) couldn’t rebalance fast enough and got liquidated. The permanent change: generally avoiding non-dollar collateral (USDE and wrapped SOL that “de-peg temporarily” get liquidated aggressively), more portfolio margin, more conservatism. Backtests never see these once-every-3-4-year events.
• Crypto-native microstructure edge survives the Jane Streets: perp funding is unknowable beyond the next window — “the formula just tells you the next window, not the one after that.” As perp open interest expands into SpaceX and trades 24/7, “this is our world… they’re in our land.” Jane Street steps into ETF-level games and will “probably shave off a few points.”
• The primitive of all finance, per Jordi: “you’re always exchanging risk for expected value” — de-riskers (founders, locked-share holders) give edge to secure the bag, while gamblers pay edge to buy variance (“they want that thousand X”). Practical corollary for retail: fees compound — a 5% edge leak “takes you from being profitable to unprofitable” — and platforms like Kalshi deliberately steer you into market orders.

Deep dive

1. The great filter: retail is gone and token structure broke it

• Jordi’s opening read on the cycle: “the great filter is underway” — lower-tier founders can’t get funded anymore, and the only money left is disciplined. The exception proving it: Prime Intellect, a Selini portco in crypto AI, announcing a ~$300M raise. Yet on the crypto-AI panel the host moderated, nobody would stand up for a specific project (Grass, Tao included) — everyone was “very bullish on AI and very bullish on crypto,” but “it was kind of hard to get people to join the two.”
• Grass as the diagnostic case: the team told an investor call it was making roughly $50M cash, and the token sold off a third — on good news. Founder Andre’s complaint, as relayed: as an equity he’d trade at a very high multiple; instead it sits around a 380 FDV despite the growth rate. “This doesn’t make any sense.”
• The mechanism Jordi blames: Binance forces ~year lockups to secure float, and a year later “a bunch of VCs that are down bad” have to sell whatever still has liquidity, dragging tokens down until retail quits. Ticket flows are down “massively” — diffused into single-name equity bets (Micron, SanDisk), sports betting, and prediction markets. “Retail don’t have any money.”
• On the AI-to-crypto rotation question: it has to play out slowly — people must exit SpaceX, memory stocks, and their KOSPI winners first. Crypto is a momentum game both directions and right now it’s “all sideways,” making a choppy bottom; the eventual move will be sudden and reflexive — “it’ll all happen for no reason… we’re up 10% because yesterday we were up 10%.” Investors are still just chasing FOMO; he doesn’t know when that changes.

2. TradFi became crypto: SpaceX is the new majors

• Selini is trading “the crypto version of equities” across Binance, OKX, Lighter, even Bitget, hedged on the tradfi side — SpaceX alone was trading $2 billion a day on Binance, bigger than the Ethereum market. Jordi was never an equity guy (fixed income, commodities, FX), but “that’s where the game is being played. And honestly there’s alpha there.”
• The transfer of the playbook, in his own words: “It’s the same things we learned in crypto” — the KOLs (Elon, the CNBC talking heads) “bullshitting some story” to move retail; a “ridiculous amount of unlocks” in August, then the next tranche, then the earnings call two weeks later. “We’re tracking all this stuff. We know how this plays out.”
• The host’s addition: IPO allocations through Goldman (likely; “Oldman” as heard) are the Binance-launchpad equivalent — priced low, flipped by everyone. “We always would joke that crypto is going to become more like tradfi, and it has. And then tradfi has become like crypto.”

3. Passive flows hold the bag — and they’re mechanical

• Jordi’s sharpest structural call: the Nasdaq is about to pile money into a just-listed, low-float SpaceX — SpaceX isn’t in the S&P 500, which is “13, 14 times bigger than Nasdaq.” “These passive flows are always going to be the ones that in the end hold the bag, and these automatic flows are so easy to predict.”
• Millennium’s reported $2.8bn, $3.7bn, and $2.7bn figures on index rebalancing aren’t genius, in his view — “are they that smart that nobody else can figure out an index rebalance? I don’t think so” — just capital plus attention on mechanical dumb-money flows. Crypto ran the same trade on Michael Saylor, “in essence the NASDAQ” that passively buys: funds bought before him, sold after him, and had to exit when he stopped.
• The host’s bear metaphor, worth keeping: the bear market is a literal bear — “you don’t have to be the fastest person on the pitch, just faster than the slowest,” and the slowest are pension funds and index money. The catch is timing the pop: “you sell Intel at 110 yesterday, you wake up today it’s at $17… why did I sell?” Jordi’s heuristic: never maximize any single month or year — keep cash, keep smart beta, “compound over time.”

4. October 10: anatomy of a once-every-3-4-years day

• A mixed day, not a disaster: some arbitrage teams “did spectacularly well,” but under-capitalized accounts running 3–4x leverage — the difference between a 10–15% fully-collateralized return and ~60% — couldn’t rebalance fast enough, went one-sided, and got liquidated. Net damage: single-digit millions lost, “not a huge day for us, good or bad.” And backtests can’t save you: “it’s not happened before.”
• The street’s reflex: Jordi woke to 50 Telegram messages asking “are you guys okay?” — people assumed Wintermute or Selini had gone bust, because “that was the only explanation.” Wintermute has now “died” more than 20 times; Selini only two or three — “a lot more deaths to go.”
• The permanent changes: more unified liquidity via portfolio margin, and generally avoiding non-dollar collateral — 10/10 proved that USDE or wrapped SOL that “de-peg temporarily” gets you liquidated “pretty aggressively.” Maybe Bitcoin to some extent. “We use leverage, but we’ve gotten smarter on how we do it and a little more conservative.”

5. The primitive of finance: risk-for-EV, and who’s paying the edge

• Jordi’s first-principles frame: “You’re always exchanging risk for expected value… you are in essence going to make money if you consistently find someone that wants to reduce risk and will give EV for that.” Founders IPO-ing, locked shareholders, unlocks — “anytime there’s something structural, you can structure something,” including buying locked shares and hedging them. Selini’s own version: tight liquidity for exchange retainers, alongside tens of millions spent on infrastructure — microwave feeds from New York to Japan’s AWS data centers.
• The inverse trade defines crypto: gamblers give up EV to buy variance, not shed it — “they just want that thousand X and they’re willing to pay edge to have the variance.” The most successful crypto products sell exactly that, like the lottery. The host’s specimen: the Yankees’ 50/50 raffle — half to charity, half to one winner, “the worst lottery that’s ever existed in the history of the planet” — still draws $250–500k a stadium, because it’s fun. Hence Robinhood’s gamification: “the more you introduce dopamine into trading, the more edge you can extract” — and the host’s standing warning, “do not trade from your phone.”
• The microstructure edge vs the Jane Streets: perpetuals are great UX but “you never know what the funding rate’s going to be the next day… the formula just tells you the next window, not the one after that.” “This is our world… they’re in our land” — and as perp OI expands to include SpaceX, 24/7, Selini’s opportunity “expands massively” while Jane Street steps into ETF-level games and will “probably shave off a few points.”
• Jordi’s own Kalshi (likely; “Kashi” as heard) World Cup bets exposed the retail fleece: the UI steers you toward market orders and buries passive limit entry — getting filled at 14 cents instead of 15 takes “extra clicks” — deliberately segmenting the price-insensitive gambler, Coinbase versus Coinbase Pro. The compounding lesson: if a 5% edge leak “takes you from being profitable to unprofitable, it changes your entire trajectory” — like the 2021 traders paying $50–100k in Ethereum gas for $400k profits “you end up giving back years later anyway.”