Anthropic's Fable Banned, Will Microstrategy Blow Up & Structuring A Portfolio in 2026

Summary

• Jordy Alexander (Seleni) calls the MicroStrategy panic “way overblown” even with STRC at 85 cents and a participant citing less than eight months of dividend coverage. Jordy says Saylor is still raising about $200M a week, keeping half in cash and buying Bitcoin with the rest. Jason lays out selling MSTR down from roughly 1.1–1.15 mNAV toward 0.7 or cutting the preferred dividend; Santi adds selling $5B of Bitcoin. Jason says Bitcoin is likely last because Saylor is too important to it. Santi still says a Monte Carlo simulation eventually blows MicroStrategy up.
• Anthropic released Fable around June 9, and the US government moved to block its export three or four days later. Santi says repeated warnings about job losses create fodder for negative narratives; Jordy says the AI backlash will intensify around the US midterms, when anti-data-center politics will be useful. The panel frames AI as having a storytelling and PR problem, with Dario, Sam, Dez and Elon spending significant time in the media.
• The investable read-through is open-source models plus smart routing, not privacy maximalism. Santi cites Case’s reversal—“I was wrong. Z.ai is on another level”—and GLM-5.2 running Hermes through AskVenice. Jordy says being six months behind is tolerable if governments censor the extra six months of frontier progress. Rob counters that governments may restrict open models too if they treat them like nuclear weapons; Jason says normal tasks will favor cheap local or crypto-rail models and that tokenization of AI may be more promising than privacy rails.
• Hyperliquid’s revenue has stayed just under $1B a year, while its multiple increasingly reflects DAT buying rather than trading growth. Santi says they remain long but have scaled out. Jason says HYPE looked fully valued around $62–$63 and that the harder underwriting question is whether it can take share from CME and ICE. Trade.xyz is the growing RWA venue, but its revenue is not accruing to token holders, creating a potential conflict among its shareholders, Hyperliquid stakeholders and fee payers.
• Jordy says his firm has no outside capital yet and is trading its own assets; he is flat-to-waiting on a SpaceX short. He calls SpaceX a low-float shitcoin facing months of unlocks and is watching OTC brokers quote locked shares, which he models like locked tokens. Santi says his venture book has done only three crypto deals since 10/10—Canton, Catalyst with SEOA and one other—versus roughly 15–20 AI and biotech frontier deals, plus defense and rare earths.
• On Bitcoin, Jordy is buying the dip and Zcash in the $400s, but expects a choppy path: maybe $50K, probably not $40K except possibly a one-day wick. He wants Saylor off CNBC and thinks Bitcoin needs two or three new anchor figures such as Paul Tudor Jones, Ken Griffin, Ray Dalio or Elon Musk. Santi’s own threshold is below $50K. Jordy also says Elon and Sam Altman may be more bullish on crypto than the crypto-native people on the podcast.
• Jason’s darkest call is that low-float IPOs are targeting passive Nasdaq/Russell and 401(k) flows. Family offices are FOMOing into single-asset SPVs, money velocity is high, and Jason says the first money will be the dumbest money and that these structures will bleed once unlocks begin.
• Coinbase’s rollout is a broad everything-exchange and money-management push. The panel discusses pre-IPO perps, prediction markets, combos, more access to Coinbase International and Deribit for US users, a developer platform, and consumer-finance products. Jason finds the exchange offering broadly expected, the developer changes unclear, and the consumer-finance work the most interesting. The guest frames Coinbase as more trusted and institutional than Robinhood, while Santi is glad it is trying to build a crypto-rail money-management super-app.

Deep dive

1. AI’s real problem is inequality—and health is the way back

• Jordy’s macro read is that AI backlash is “not even close to an all-time high” because it will intensify around the US midterms, when politicians have incentives to oppose data centers. He says he thinks Sundar was booed or faced a walkout during a Stanford commencement speech. The deeper theme is inequality: “it’s less about AI per se than about people feeling left out.” Even Jordy, who sees frontier deals firsthand, says “even I feel left out.”
• His proposed solution doubles as a trade: labs may need to convince the public that AI makes people healthier. He is “very long Eli Lilly” and would not be surprised if OpenAI or Anthropic built an internal biotech operation or bought a biotech company.
• Santi’s supporting case is that spending on GLP-1s and health—despite the circularity in parts of AI spending—is already a huge and potentially larger category. He is bullish on health and biotech.
• The health interlude is Santi’s, not Jordy’s: Santi says he moved from 26% to 14% body fat while doing 15,000–20,000 steps a day, lifting three times a week and eating clean. He also takes four or five orally bioavailable peptides, including BPC, NAD and SLU-PP-332, but explicitly says the result was not just the peptides.

2. The Fable ban: doomer messaging turns into a liability

• Jordy says Fable was released around June 9 as a public version of the model people had feared. Three or four days later, Lutnick and the US government moved to block its export; because Anthropic could not restrict access to US users, it shut the service down.
• Santi argues that Dario’s repeated warnings about job losses create fodder for negative interpretations of AI. The panel contrasts that with the more optimistic message from the Perplexity CEO on 20VC: AI makes new businesses possible, and entrepreneurship is rising as people build with fewer employees and less money.
• Jordy says the AI industry has a storytelling problem and that the ban was also a media and political failure. He argues that Dario, Sam, Dez and Elon are spending substantial marginal time on interviews because influencing public narratives may be one of the highest-return activities for a frontier-lab CEO.
• The discussion also notes the irony that pessimistic messaging can be useful for fundraising until it turns against the company with regulators and politicians.

3. Open source hits the national-security wall

• Jason says Fable reinforces the case for open-source models, especially as their performance improves. He is less interested in open source reaching the absolute frontier than in its being close enough for ordinary tasks, where a cheap local model or crypto-rail model may be preferable.
• Santi cites Case’s reversal: he had expected the gap between open and closed models to widen because of data, hardware and regulatory advantages, but now says, “I was wrong. Z.ai is on another level.” Case is running his Hermes agent on GLM-5.2 through AskVenice.
• Jordy’s counterpoint is that being six months behind is not necessarily important if governments censor those extra six months of progress before it becomes widely available. He agrees that open-source models are useful but doubts they will become frontier models.
• Rob’s pushback is the nuclear-bomb analogy. If governments really believe advanced models can cause national-security harm, they may try to restrict open models as well. He does not want crypto to recreate the argument that its primary use case is circumventing government controls. His dystopian scenario is simultaneous restrictions in the US, EU, China and elsewhere that halt the free flow of models.
• Santi compares the dynamic with Bitmain’s ASIC advantage: access to compute and capital may matter more than nominal model openness. He says new data centers may take roughly four years to approve. Jason is more bullish on tokenizing AI—compute or fundraising for future training runs—than on privacy rails. The panel puts AI around the third or fourth inning and crypto in its “first inning.”

4. Token spend enters its second phase: route, don’t max

• Jason divides the opportunity into three buckets: closed providers such as Anthropic, Google and OpenAI; open-source models such as MiniMax, DeepSeek, Meta’s models and possibly xAI; and routers such as OpenRouter and Venice.
• Blockworks initially ran AI spending uncapped to force adoption. When GPT-5 came out, Jason says the natural instinct was to send ordinary questions to the best model, even though perhaps only 5% of the company’s work should use the frontier model and most tasks could use older versions such as 4.6 or 4.8.
• Rob says “token maxing” is facing a backlash. The likely future is a middleware layer that learns which foundational model best fits each workflow, balancing quality, speed and cost.
• Jason describes Blockworks’ first phase: unlimited budgets, a shared Slack channel called AI Kitchen, and examples that let salespeople, engineers and analysts learn from one another. Ramp now provides better visibility, including the ability to see how much of the company’s token spend comes from each team. After six months of running wild, Blockworks is beginning to tighten budgets and measure ROI.

5. Jordy’s book: shorting SpaceX like a locked token

• Jordy says the firm has no outside capital at the moment and is focused on managing its own assets for at least five years before taking money from others. He describes a mix of high-Sharpe systematic strategies and more tactical trades.
• His SpaceX position was long into the run; he is now getting flat and waiting for what he calls an “inevitable short.” His framing is that SpaceX resembles a low-float shitcoin with month after month of unlocks. The tell is that crypto OTC brokers are receiving locked SpaceX shares, allowing him to apply the same discount and unlock modeling used for locked tokens.
• Santi says his venture book has done only three crypto deals since 10/10—Canton, Catalyst with SEOA and one other—against roughly 15–20 frontier deals across AI and biotech. He also likes defense and rare-earth bottlenecks tied to possible geopolitical escalation.
• Santi does not want to back “prediction market number 18” pre-traction. Jordy makes the parallel point that there are too many founders with pitch decks and says he would rather invest at the next stage, once traction is proven, even at a higher price.
• Santi says he is not being paid enough to be broadly long crypto. Backpack is the exception he mentions because he is already an investor and likes what the company has done.

6. Hyperliquid’s revenue ceiling and the Trade.xyz fault line

• Santi says Hyperliquid’s revenue has “never changed” in the relevant range: it remains just under $1B a year. He says they are still long but have scaled some exposure out. The multiple now feels more driven by DAT buying than by expansion in trading.
• Santi also questions the idea that Hyperliquid can simply turn on a fee switch or triple RWA fees. As with Uniswap, LPs expect their share and traders will not accept unlimited slippage.
• Trade.xyz is private, has become an important growing venue for RWAs and other products, and is reportedly raising a new round in the billions. But its revenue is not accruing to Hyperliquid token holders or other public participants. Jason says that creates a potential conflict among Trade.xyz shareholders, Hyperliquid stakeholders and the users paying fees; for now, “it’s kumbaya because it’s going up and the DATs are buying.”
• Jason says HYPE looked fully valued when it first reached roughly $62–$63. The relative-value argument has weakened, so the harder question is whether it can take meaningful market share from centralized venues such as CME and ICE.
• Jason also points to CFTC chair Mike Selig’s shift from talking about “DeFi” to “on-chain finance.” Jason interprets that as a preference for exchanges operating within KYC/AML and reporting rules rather than permissionless, non-KYC DeFi.
• Jason’s Solana view is that the asset is not a monetary instrument, has years of unlocks, and does not justify a $250B valuation on fees burned. The on-chain-Nasdaq thesis has also weakened because users may prefer a centralized platform that simply works.

7. MicroStrategy: STRC at 85 cents and Saylor’s three-way triage

• STRC traded at roughly 85 cents during the recording. The transcript also includes a claim that there was less than eight months of dividend coverage. Rob says STRC is near the top of a tentative capital-structure waterfall: convertible senior notes first, followed by Strike, Strife, Stretch, Stride and common.
• Rob says the pricing implies a possible unwind and that realizable net asset value could be around half of the company’s current market capitalization, though he emphasizes that he needs to do the actual math. His conclusion is that the market is flashing serious warning signs.
• Jordy calls the comparison with Luna “way overblown.” He says Saylor is still raising about $200M a week, keeping half in cash and continuing to buy Bitcoin with the rest. Jordy does not expect him to sell Bitcoin soon.
• Jason lays out the triage: sell MSTR stock while it remains above mNAV, potentially pushing mNAV from roughly 1.1–1.15 toward 0.7; or kill the dividend and let the preferreds fall to 40–50 cents. Santi adds a third option: sell $5B of Bitcoin, which could help both the stock and STRC.
• Jason thinks Saylor is trying to protect himself and the preferreds while leaving Bitcoin for last because he is too important to Bitcoin. Santi remains much more skeptical, saying that if the strategy is simulated repeatedly, “eventually that thing blows up.”
• Santi’s counterargument is liquidity: there is enough capital in the world to buy distressed coins, Strategy securities or Bitcoin. He says $50B is no longer what it used to be, and Jason adds that Tether could also step in.

8. The Bitcoin bid: sub-$50K buyers and a new anchor

• Jordy says the firm is buying more Bitcoin on the dip and adding Zcash in the $400s as a hedge and potential monetary instrument. He expects a choppy path: perhaps $50K, but not $40K except possibly as a one-day wick.
• His bullish case is renewed concern about the dollar, similar to 2021, when gold and Bitcoin benefited from demand for neutral assets. He says gold has too much physical friction—“Dubai is getting bombed. You want to move your gold—how are you going to do that?”
• Jordy says quantum-computing fears were the main Bitcoin damage six months earlier, but that the only likely developments there now are positive ones such as agreement on a hard fork or another mitigation.
• Jordy thinks Saylor should stop doing so many CNBC interviews and privately meet Paul Tudor Jones and other large market participants. Bitcoin needs two or three anchor figures—Paul Tudor Jones, Ken Griffin, Ray Dalio or Elon Musk—to say it deserves a place in portfolios. He says Ricardo Salinas has about 90% of his net worth in Bitcoin but does not carry the same weight.
• Santi’s own threshold is below $50K: if Bitcoin gets there, he says it would deserve a place in his portfolio. Jordy also observes that Elon and Sam Altman may be more bullish on crypto than the crypto-native people on the podcast.

9. “Financial terrorism”: passive flows as exit liquidity

• Jordy says RIAs remain a major gatekeeper for American wealth. He thinks some may be uncomfortable after recommending assets near the $120K Bitcoin level, but he also says Bitcoin has not penetrated the wealth-management channel fully.
• Jordy adds that alternative investments are flooding RIA and wirehouse platforms, including private markets, private equity, funds and IPOs. Some large wirehouses are beginning to recommend small IBIT allocations.
• Jason says the strongest demand is concentrated in single-asset SPVs for neoclouds, robotics, frontier models and the next SpaceX. Family offices that had never created SPVs are now not only participating but considering launching their own.
• Jason’s darkest warning is that passive flows and 401(k)s are being targeted. Low-float companies can be pushed into Nasdaq and Russell exposure, forcing index-linked money to buy. He calls it “financial terrorism” and says the first money will be the dumbest money.
• His poker analogy is that, over enough hands, money flows from weaker players to smarter sharks. He says the velocity of money is high and is “100% sure” the low-float structures will bleed once unlocks begin.

10. Coinbase’s everything-exchange and money-management push

• The panel discusses Coinbase’s expansion across several categories. The exchange side includes pre-IPO perpetuals, prediction markets, combos and broader access for US users to Coinbase International and Deribit.
• Jason says the exchange breadth is broadly expected because Robinhood, Kraken, Bybit, Bitget, Binance and other platforms are pursuing similar one-stop trading products. He finds Coinbase’s developer-platform changes difficult to distinguish from existing features, though he sees value in making them easier for developers to productize.
• Jason considers the consumer-finance portion the most interesting and outside the box. Santi is glad Coinbase is attempting to become a broader money-management application and thinks the category should use crypto rails.
• The guest frames the branding difference as important: Coinbase feels more trusted and institutional, while Robinhood is associated with confetti, sound effects and zero-day-option speculation. The open question is whether Coinbase’s more serious brand can attract a broader wealth-management audience without losing its core users.