The Two Biggest Issues In Crypto Today | Weekly Roundup

Summary

• Santiago Roel Santos de-risked to cash going into Q4 and is now redeploying — under 10% so far — into equities down 30-50%, explicitly not tokens. He holds Google (“it’s a monopoly… I don’t think they’re going to be unseated”), Western Union, and ServiceNow, and tracks a “crypto-enabled businesses” watchlist of Klarna, Robinhood, and Figure. His accumulation logic: nobody catches the exact bottom (few bought Solana at $8 despite the call), so “you buy at 20 and then 15 and 12 and 8” because “we’re a tweet away from the market ripping again.”
• Rob’s macro pushback: the market is still pricing some chance of a quick resolution. There is no direct US-Iran negotiation (messages via the Pakistanis), oil is near $120, there has been no equity capitulation yet, and a Hormuz supply shock would feed inflation and hurt rate-cut prospects — “if you’re trading the short-term market right now, I don’t think you can be in the market at all… you get paid four and a quarter to just sit on treasuries.”
• The “crypto VC mass extinction” is consolidation: hit rates are collapsing toward traditional venture just as the easy token exit disappears. Santi’s frame: “You could invest in anything at the early stages and the token would launch” — now a few funds raise huge and “everyone else is going to die.” Santi sees 5-10 crypto companies a month offering to sell themselves, and estimates “80% of the crypto data companies have tried to sell in the last year or two.”
• Venture’s playbook has flipped from hunting the new thing to piling tens of millions into runaway winners like Rain and Polymarket. Moats form and break faster than ever, so once Ramp-vs-Brex-style separation appears, you concentrate; Carta data shows the median of the top 10% of seed deals at ~$120-125M valuation, up ~2.5x in a year, while sensible crypto seeds price at $20-25M — with hot-topic exceptions (pre-revenue stablecoin rounds at $50-100M+).
• Crypto token secondaries are generally in the 80-90% discount-to-spot range — and spot is already below the last round or even the Series A — the widest Santi has ever seen. At a 90% discount he flips: “I’m now a buyer at that level of certain projects… price is the single determinant of returns. Historically tokens just don’t go to zero.” Jason disputes 90% marks for many of the names he surveyed, while Santi says liquid names can still find buyers around 60% discounts; past the top ~30 tokens, however, “there’s probably not a buyer for the vast majority. Today.”
• The Drift hack — ~$280M drained in 12 minutes from a Solana perps DEX with ~$500M TVL — was primarily an operational-security failure. A worthless token (CVT) was wash-traded into oracle-accepted collateral, 30+ pre-signed durable-nonce transactions were staged, and a security-council multisig had been lowered to 2-of-5 with no time lock. It followed the Resolv Labs ($50M USR) and Stakehouse incidents; Jason’s response: “I’m not keeping any money in DeFi right now.”
• The real vulnerability is people: “The vast majority of hacks happening today are social engineering,” and nobody audits team ops the way they audit code. Jason’s quant tallies ~$1.1B in 2025 DeFi losses ex-Bybit and $350M this quarter alone, while Aave stablecoin yields sit below treasuries — “you should be paying 30% APY, not 10 or 8” for looped composability risk. Jason used Mike’s tweet to frame the two biggest issues as transparency around security and tokens, which he said may be 12-18 months from being solved.
• Google’s quantum paper cut the qubits needed to break wallet cryptography 20x (10M → 500,000), pulling the threat timeline toward maybe 2029 — and Jason predicts a contentious Bitcoin hard fork this year. There’s no post-quantum Bitcoin address format yet (BIP 360 is only a proposal, and rushed code “might create a catastrophic vulnerability in the same move”), and coordinating a leaderless community evokes 2017’s block wars. “Once again Nick Carter was right.”

Deep dive

1. Santi rotated to cash in Q4 — now nibbling equities, not tokens

• His setup: he de-risked heavily going into Q4 — “I just don’t think I’m getting paid enough to take this level of risk. Everything’s all-time high” — without foreseeing the Iran war; now less than 10% of that cash is redeployed, “nibbling at names I like that are down 30 to 50%”: Microsoft -24% YTD, Google -20%, Meta -17%. Equities only, no tokens, and “not jumping out of my seat” — he doesn’t think they’ve bottomed.
• Actual positions: Google (“I feel really comfortable owning for multi-year periods cuz it’s a monopoly”), Western Union, ServiceNow. His watchlist theme is crypto-enabled businesses — Klarna, Robinhood, Figure — to see how stablecoins and tokenization transform them, plus Better Homes for its facility with Sky.
• The accumulation logic, via Solana: Jason publicly called $8 “a generational bottom,” but “how many people do you think actually bought at eight? You still would have done well if you bought at 20 and then 15 and 12 and 8” — because “we’re a tweet away from the market ripping again.”

2. Rob’s macro pushback: no capitulation yet, treasuries pay you to wait

• Rob’s caution: “I don’t know if it’s as easy as Santi is making it out to be.” There’s no direct US-Iran engagement — messages pass “through the Pakistanis” — oil is back near $120, and last night’s press conference gave him “no confidence whatsoever” this ends soon. The market rushing back on any good news tells him people still expect quick resolution, which means capitulation hasn’t happened.
• His chain: Hormuz supply shock → inflation working through → “bad things for potential rate cuts” → macro challenged the rest of the year. “If you’re trading the short-term market right now, I don’t think you can be in the market at all… you get paid four and a quarter, four and a half to just sit on treasuries.” Long-horizon buying of secular-tailwind names is the one thing both agree you should be doing.
• For founders, the Lux memo (initially misattributed to Index) is the template: bake in far more conservatism — everything costs more, and capital markets “certainly aren’t as open in crypto.”

3. Venture’s new game: spot the runaway winner and pile in

• Rob’s lesson from a deliberately slow year (the fund took ~12 months to raise; ~14 months deploying): “It is easier than ever to be an entrepreneur… easier than ever to get disrupted really quickly. It’s harder than ever to get a moat” — so when someone pulls away, they pull away faster. His examples: Ramp leaving Brex behind, Rain’s month-on-month growth still accelerating at scale. Dragonfly has put tens of millions into a few proven names, Polymarket among them.
• The same dynamic runs through tradfi: one fund he spoke with went from $4B to $26B AUM in five years on a single early foundational-model position headed for a “trillion dollar plus IPO.” Rob’s twist — his job doesn’t change: to beat Lightspeed and Ribbit in stablecoins and neobanks, he still has to live with seed companies to know early which will win.
• They use OpenAI as a sentiment check: Santi notes its $122B round, while Jason points out that the timeline had been saying Claude was pulling away and OpenAI was dead. Santi’s conclusion is that “the timeline is more wrong than it’s ever been… it’s like a negative signal these days,” partly because “everyone’s a content creator” with Claude or ChatGPT whipping out posts.

4. The crypto VC mass extinction is consolidation, not collapse

• Santi’s core frame: “Crypto was easy on a relative basis for venture for a long time. You could invest in anything at the early stages and the token would launch.” Now failure rates are up, VC hit rates are falling toward traditional-venture norms, and with too much capital in a not-that-big industry, “it has to consolidate to the few that are [making money]… you’re going to see a few funds raise a bunch of money… and everyone else is going to die” — the Andreessen-raising-$15-18B pattern, at crypto scale.
• Santi from the trenches: five to ten crypto venture-backed companies a month offer to sell themselves to Inversion (“What am I buying here?”); “80% of the crypto data companies have tried to sell in the last year or two”; two of his portfolio companies closed shop the day of recording; acquirers buy the equity and deprecate the token, as with Tensor.
• The active-management corollary: if you weren’t in prediction markets, perps (“one perp really” — Hyperliquid) or stablecoins, “you should be looking at yourself.” Santi’s nuance: Lighter hasn’t performed like Hyperliquid, but it’s still worth $800M and plenty of investors are up a lot.

5. Seed is bifurcating: sensible pricing next to hot-topic froth

• Santi just did a seed at $20-25M that “would have been probably 40-50 a year or two ago” — sensible, if the team is smart. But Rob sees an AI-crypto crossover deal — “really kind of a seed, maybe they’re calling it an A” — getting done in the hundreds of millions against fundamentals worth “50, 60 million bucks,” plus pre-revenue stablecoin seeds at $50-100M and a pre-revenue second round above $100M happening right now.
• Carta data (all venture, AI-skewed, median not average): the median of the top 10% of seed deals is now ~$120-125M valuation, up roughly 2.5x in a year — “everyone’s piling into the things they think are going to win and ignoring everything else.” Rob himself just personally invested in a company’s first round at $4B.

6. Token secondaries at 80-90% discounts — and Santi flips to buyer

• Santi’s OTC survey of quality, venture-backed tokens (launched within a year, pre-first-unlock, teams and investors on one-year cliffs): discounts clustered in the 80-90% range to spot — and spot is already below the last round, sometimes below the Series A. “I have never seen that big of a discount” — historically 60% was where a buyer could hedge on perps and “almost categorically” not lose. Non-venture names have no bid at all. Crypto-equity secondaries are far tighter: a company raising at ~$4-4.5B trades secondaries around $2.5-2.7B, roughly 40% off.
• The flip, after a year of token bearishness: “At 90% discount I’ll buy that thing. At that point price is the single determinant of returns… historically tokens just don’t go to zero. I will flip — I’m now a buyer at that level of certain projects.”
• Jason’s dispute — he’s an investor in four of the ten names Santi quoted: “I’m almost certain I can sell them for a lot less of a discount than 90%…”; Santi adds that liquid tokens with a positive outlook can find buyers at 60% discounts because they can be hedged on perps. His hierarchy: maybe 5-10 tokens people really like, another 10-20 they’re constructive on at the right price, and “after you get past 30 tokens there’s probably not a buyer for the vast majority. Today. There will be in a bull market.”
• Jason’s blunter read: “People have written off tokens as an investable asset class right now… except for maybe 10 of them.” Santi’s rebuttal from the tape: ex-Bitcoin ($1.3T of a $2.3T total circulating market cap) there’s still more than $1T of liquid value and $107B in 24-hour volume — willing buyers exist.

7. The unresolved fight: is volume a proxy for what people actually want?

• Santi argues that most volume is market-maker and quantitative activity rather than a clean measure of what people like. Market makers participate where they expect soft flow, while other traders pursue volatility, basis, hedging, and cross-venue or cross-chain arbitrage. He notes that perp volume is down 60% since October.
• Jason pushes back on the funding-rate logic, raising base funding rates of roughly 10.95% and arguing that market makers can trade perps even without a directional view. Santi disagrees, and Jason moves on: “We’re not going to get to the bottom of it. I’m moving us forward.”

8. Anatomy of the Drift hack: ~$280M in 12 minutes

• The mechanics as Jason reconstructed them: the attacker created a worthless token (CVT), seeded a liquidity pool with a couple hundred dollars, and wash-traded it until price oracles treated it as real collateral. Then, using Solana’s durable nonce feature — transactions that can be pre-signed and held indefinitely rather than expiring in minutes — they staged 30+ withdrawals and parameter changes in advance and fired them in sequence, draining ~$270-285M from a perps DEX holding roughly half a billion in TVL.
• The enabling failure: weeks earlier, Drift’s security-council multisig had been lowered to 2-of-5 with no time lock. Santi’s read on the compromise: one signer “got hacked, essentially,” and somebody else blind-signed. The attacker then changed the admin key entirely and locked the team out — nothing could be frozen — after two audits had passed the protocol. Jason: “Why did you remove the time lock? It makes no sense.”
• The pattern that alarms him: Resolv Labs was hacked ~a week or two earlier ($50M in USR via a compromised AWS environment), and Stakehouse’s front end pointed to a wallet drainer two days prior. “I’m not keeping any money in DeFi right now. I’m just not. It’s just not worth the risk.”

9. The attack vector is people, not code — start auditing the ops

• Santi’s thesis: “The vast majority of hacks happening today are social engineering” — Lazarus and North Korea’s actual playbook is getting you to click the wrong email or spoofing a front end, not out-mathing your smart contract. Teams get “lulled into a sense of security” by convenience-driven shortcuts like fewer signers. AI cuts both ways: attack surface is “much higher than it’s ever been,” but security should use it too.
• Jason’s proposal: audit firms audit code, but “nobody really audits the ops of teams” — MFA, password managers, what happens when someone sends you a picture of your kids. He cites Hasu’s checklist: circuit breakers on deposits and withdrawals, time locks on any change, security councils that can shut things down instantly. At DAS, demand pulled Blockworks’ token transparency framework toward rating RWA vaults’ smart-contract risk — “we’re not trying to compete with Moody’s.”
• Santi says he still keeps “a bunch of money in DeFi” and trusts particular teams with strong operational security, but he is selective about where that money goes. Rob says Dragonfly’s current VC funds have never put LP capital into vaults; an earlier liquid-fund strategy was spun out years ago.

10. DeFi yields don’t pay for hidden composability risk

• Jason’s math: Aave stablecoin yields sit below treasuries; Morpho pays 5-9%. But composability means “you’re only as secure as your weakest link” — he thinks Morpho had bad debt connected to the Resolv hack — so looped multi-protocol strategies “should be paying 30% APY, not 10 or 8.” He is more worried about capital preservation and impairment than about incremental yield.
• Santi says his own math has also often shown that he is not paid enough to take that level of risk, even though he still has some money in DeFi and sometimes earns better rates. He says this view is more representative of institutional capital; Dragonfly’s current VC funds have not pursued vault farming.
• Jason calls this a classic Taleb turkey. Rob’s response is that it is situation- and context-specific: some products pay for risk, while lending stablecoins on Aave may not. “It feels like this whole pod has been the death of crypto… maybe that’s a bottom signal.”
• Jason closes the loop with Mike’s tweet: the two biggest issues in crypto are transparency around security and transparency around tokens — “12 to 18 months away from solving both, but it’s clear capital won’t come back until we do. Investors won’t risk their entire principal for 4% yields.”

11. Quantum’s timeline just jumped toward 2029 — and a hard fork fight looms

• Two papers dropped within hours. Google’s quantum AI team cut the estimated resources to break the elliptic-curve signatures protecting Bitcoin and Ethereum wallets from ~10M physical qubits to 500,000 — a 20x reduction — with a key crackable in roughly 8-10 minutes on superconducting hardware, inside Bitcoin’s ~10-minute confirmation window, meaning a transaction could theoretically be intercepted mid-flight. A newer, less-vetted Caltech paper from a startup whose name Jason rendered uncertainly as “Aura Oratom Oratomic” claims 26,000 qubits on neutral-atom architecture, at ~10 days per key.
• Jason says the US government had already targeted phasing out current cryptography by 2035, the NSA had suggested perhaps 2030 for especially exposed systems, and Google’s estimate may bring the threat as early as 2029 — the acceleration is what has people scared.
• Rob explains why the upgrade cannot simply happen immediately: there is no switch to flip, Google itself would not be post-quantum until roughly 2029, a Bitcoin migration could take years, and self-custodied holders may not participate. Jason adds that there is no post-quantum Bitcoin address format yet — BIP 360 is a proposal, and rushing new code means “maybe you could fix the quantum problem, but you might create this catastrophic vulnerability in the same move.” Then you still need a hard fork with no CEO to order it, in a community whose hardcore resists — “once again Nick Carter was right.”
• Jason’s prediction, revised: “A Bitcoin hard fork will become very contentious… I think it’s this year” (he was off by a year), evoking 2017’s big-blockers-vs-small-blockers — though exchange-held BTC may make coordination faster. Open questions on the table: dormant coins and addresses may be more exposed; CZ says all crypto has to do is upgrade to post-quantum algorithms, while Brian Armstrong says he’s personally spending “a lot more time on this.”