#FinanceDaily: “Sock-puppet”; $CAR 🔥; $LCID/ $UBER; Vega/Fertitta/$CZR

MarioNawfal avatar
Mario Nawfal
@MarioNawfalApr 23, 2026

The Spaces covered a fast-moving mix of markets, AI, energy, and legal-industry AI risks. Markets opened bid across risk and commodities (WTI near $90, Brent ~$99–100, Bitcoin >$78K) amid an indefinite land ceasefire but ongoing maritime tension in the Strait of Hormuz. The panel unpacked SpaceX/xAI’s option-style deal with Cursor (partnership at $10B or acquisition at $60B) as a talent-plus-platform play, debating whether a development environment merits that valuation and whether it’s partly an IPO optics boost. They assessed Robinhood Ventures’ $75M OpenAI stake and the broader trend of retail exposure to late-stage private rounds, warning of IPO timing, lockups, and asymmetric risk to retail. Kevin Wash’s confirmation hearing drew focus for a potential shift to trimmed-mean inflation—seen as expanding scope for earlier rate cuts—and the politics around Powell’s exit. Energy dominated mid-show: Strait disruptions, spot vs paper oil prices, supply-chain lags (jet fuel, polymers, copper via sulfuric acid to Chile), and the puzzling resilience of European equities. Views diverged on US insulation and supply elasticity. A major legal AI failure (Sullivan & Cromwell submitting an AI-hallucinated brief) catalyzed a broader critique of overreliance on AI in law and code, billing opacity, and rising security risks—alongside reports of banks quietly hardening defenses.

Session overview and participants

  • Host: David (Speaker 1)
  • Eric (Speaker 2)
  • Brandon/Brendan (Speaker 3)
  • Eladio (also spelled Eligio in transcript) (Speaker 4)
  • Dave (David) Nicoski/Koski (market technician; spellings varied in-session) (Speaker 5)
  • Paul (Speaker 6)
  • Justin (Speaker 7)
  • Dwayne (Speaker 8)
  • Cyrus (Speaker 9)
  • Robert (joined late in the AI/legal segment; Speaker 4 in that portion)

The session covered: a live market/geopolitical setup (oil, crypto, ceasefire dynamics and the Strait of Hormuz), a deep dive on SpaceX/xAI’s optioned deal with Cursor, retail access to private AI via Robinhood, IPO risks and lockups, Kevin Walsh’s confirmation hearing and inflation metrics, energy supply chain risk, Europe vs US market resilience, metals (copper, aluminum) and munitions inputs (tungsten), and an AI legal fiasco at Sullivan & Cromwell alongside broader concerns about AI misuse in law and software engineering.

Market setup and geopolitics at open

  • Tape and macro:
    • Futures up; WTI ~$90, Brent ~$99–$100.
    • Crypto bid: Bitcoin noted above $78,000; ETH ~$2,400 (as said on-air).
    • Ceasefire described as “indefinite/undefined,” no real visibility on peace talks; Strait of Hormuz effectively blocked/embargoed, with sporadic seizures and shots fired at ships.
    • Prior-day stocks faded early gains amid conflicting reports (e.g., speculation around JD Vance and others traveling to Pakistan for talks).
  • Political headlines: administration churn; Labor Secretary departure; Kash Patel filing a ~$250M defamation suit; Kevin Walsh’s testimony (with chatter about “sock puppets” in the hearing).
  • Tech tape headlines flagged (little time spent on most): Google announced two new AI chips (Nvidia competition narrative), SOX index at all-time highs, Uber’s >11.5% stake in Lucid (potential backstop), Avis (CAR) short-squeeze and manipulation chatter, Deutsche Telekom + T‑Mobile combination speculation (antitrust questions), and an AI misstep in a major legal case (detailed later).

SpaceX/xAI – Cursor: structure, valuation, and strategy

What was announced (as discussed)

  • Brandon: The SpaceX/xAI–Cursor deal “came out of left field.” The structure reportedly offers two paths: a full acquisition option priced around $60B, or a partnership at ~$10B. Cursor is essentially a developer environment/harness integrating multiple LLMs (Anthropic’s Claude, OpenAI, and others), with strong enterprise traction and bulk access deals to models. Cursor aspired to build its own state‑of‑the‑art model but lacked sufficient GPUs and funding; xAI has capital but has lagged top-tier talent and model performance (vs OpenAI/Anthropic). The option structure lets xAI benefit from shared training/knowledge and only pay up if Cursor reaches SOTA; it could also add an IPO “bump” for xAI.

Valuation and build-vs-buy debate

  • Eric: $60B for a development platform is a steep price given today’s AI can help rapidly re‑engineer such platforms; he argues a capable team could replicate a Cursor‑like environment in 24–48 hours using AI to rewrite and re‑compose code, pointing to a prior incident where leaked Claude code was re‑implemented in another language by developers within a day. Eric would favor the $10B partnership path first; Elon is generally careful with capital. He raised open legal questions (e.g., if you partner, are you permitted to re‑engineer from scratch?) and deferred to counsel.
  • Brandon: Agrees the option is smart. Cursor’s enterprise foothold is meaningful; a key watch item is whether existing bulk‑use deals with Anthropic and others persist under xAI’s orbit.

Compute, RAM, and the two‑tier AI future

  • Eric: Hardware will bottleneck home/consumer AI experiences. VRAM/RAM constraints push a two‑tier AI structure: a ubiquitous “free/home” tier (today’s sweet spot ~24–32 GB VRAM) and a higher tier (128–256 GB RAM) that brings near top‑model capability locally. Expect continued RAM shortages and rising prices until new fabs come online.
  • Business model strain: Companies with diversified revenue (Google, Meta) can underwrite AI costs; OpenAI/Anthropic may face strain at the consumer level and will likely focus more heavily on enterprise to fund leading models.

Retail access to private AI and IPO risk (Robinhood Ventures Fund 1 → OpenAI)

  • Deal: Robinhood Ventures Fund 1 (publicly traded) invested ~$75M in OpenAI, its largest deal to date.
  • Brandon: Surprised they didn’t go bigger; retail exposure to private assets is happening regardless of debates on merits.
  • David: OpenAI has been reducing retail access in later private rounds but may include retail in an IPO; still, IPO pathways can shut suddenly (market window risk). Retail investors often embrace upside but may not appreciate the downside scenarios, creating litigation and political risk (e.g., congressional scrutiny) if timelines slip or first IPOs perform poorly and push others to delay.
  • Eladio: The late‑stage IPO phenomenon—after multiple private rounds—disadvantages retail versus earlier‑stage investors. Over‑regulation and governance burdens have pushed companies to raise more pre‑IPO capital. Mentions a controversial Delaware precedent around Elon Musk as an example of governance friction. Lockups matter: when earlier investors are free to sell, retail can be hit by supply. Recent bank IPO timing has been poor; despite that, after a correction he sees summer as a potentially better window for IPOs. Overall, participating immediately post‑IPO often isn’t in individuals’ best interests.
  • David (host): Recent AI‑adjacent IPOs have been volatile; more pre‑IPO investors than ever magnify lockup dynamics.

Kevin Walsh confirmation hearing: inflation metric shift and policy timeline

  • David: Key takeaway—Walsh expressed preference for a trimmed mean inflation measure (vs core PCE). That measure tends to run lower than core PCE, effectively widening room to cut rates without “moving the goalposts” explicitly. Walsh spoke well, but non‑disclosure of specific wealth holdings drew fire (Sen. Warren highlighted this). Questions were raised about communications with Donald Trump; David sees confirmation as likely, albeit clouded by the reported “criminal prosecution of Jay Powell” causing a procedural snag (as discussed in‑room), with the expectation it will eventually be resolved. He expects Walsh and Myron (ally) may push for cuts sooner than many expected pre‑hearing.
  • Eladio: Puts Walsh’s timeline in a broader political frame tied to Iran/Strait dynamics and midterms; the administration will seek a strong market into midterms, and counterparties (Iran) understand that timing pressure.

Strait of Hormuz, oil, and supply chains: scenarios and stress points

Physical vs paper oil, spot premia, and lag effects

  • Dave Nicoski:
    • European markets’ resilience is surprising given worsening logistics. He sees setups in European short ETFs (e.g., EFZ), citing double‑bottoms.
    • Lufthansa canceling ~20,000 short‑haul flights is a tell. Even if the Strait reopens, there’s a ~40–50 day lag before crude reaches refineries and becomes finished product; distribution adds further delay. If war ended today, the supply chain shock is only halfway through.
    • Distinguishes true demand destruction from “they want it but can’t get it.” LNG declines into Korea reflect supply constraints.
    • Cites physical prices far above screens: example of Sri Lanka reportedly paying ~$244/bbl for immediate delivery; UAE/Saudi commentary implying screens show “paper” not actionable spot availability. EU officials framed current supply loss as akin to the 1973 embargo plus COVID combined.

Europe vs US, and substitution limits

  • Paul: Clarified the choke is raw inputs (crude → chemicals, polymers) that feed finished goods globally. Alternatives (wind/solar) can’t replace petrochemical feedstocks.
  • David (host): Flagged puzzling European equity resilience despite expected shortages; referenced reported real‑estate firesales in Dubai as another stress signal.
  • Eladio: The US is better insulated—largest producer, LNG capabilities boosted. Argues elasticity and new sources (e.g., Venezuela) can help cap oil >$100 and normalize within months once the Strait opens. Sees impacts landing harder abroad than in the US.
  • Dave Nicoski (rebuttal/nuance): The US consumes ~20M bpd and produces ~13.4M bpd; net energy exports exist but end‑use dependence remains. US relies on imported crude grades (sour from Canada, Venezuela) and refinery logistics; refined products and flows are globally fungible. Physical scarcity and stockpiling behavior abroad imply higher‑for‑longer dynamics.
  • Eric: Reinforces physical tightness—mid‑voyage cargoes being retendered to higher bidders; North America trade tensions (US‑Canada) complicate flows. Iran’s maritime disruption strategy aims to stretch to November; higher energy filters into broader inflation, complicating Walsh’s rate‑cut path.

Commodities and components: where pinch shows up

  • Dave Nicoski:
    • Autos: average ICE vehicle uses ~40–55 lbs of copper; EVs ~150–200 lbs. Data centers and industrial sites rely heavily on aluminum wiring (weight advantage), copper at larger gauges is heavy/expensive to pull.
    • A single missing component halts output (e.g., prior Ford “Blue Oval” emblem vs the more critical gearbox example); transmissions frequently imported (South Korea, Mexico, Canada, China), exposing assembly lines.
    • Mines reducing output: example of 2/3 employees sent home at an Australian mine; Chile’s state copper producer Codelco reported lowest copper output in 10 years, in part linked to sulfuric acid availability from the Middle East.

Ceasefire characterization and oil volatility

  • Cyrus: The ceasefire is “one‑sided”—Iran hasn’t acknowledged it and seized two tankers; expects a low‑intensity frozen conflict. Iran has regional escalation levers, but the US has overwhelming capability and is economically/energy‑wise more insulated; Asia and Europe bear faster/harder impacts. Oil volatility has spiked (oil vol index ~85, +15% day); Brent near ~$99. Elevated energy vol likely mirrors into equity vol. Longer term he remains bullish US equities and energy; sees Brent potentially easing to ~$75–$70 in ~6 months if hostilities truly cease, but pre‑war prices may be >12 months away given damage.
  • David (host): Raises the scenario risk that a land ceasefire persists while the Strait remains tense—keeping oil elevated and supply chains kinked. Asks whether that’s acceptable to the US politically and economically.

Metals and materials: copper, aluminum, tungsten

  • Dwayne:
    • Policy continuity: despite partisan rhetoric, resource policy shows continuity (IRA, nuclear expansion). Envisions an “Americas” resource bloc from North through Latin America shaping contracts over the next decade.
    • Copper guidance is mixed: BHP guided towards the top half of 1.9–2.0Mt (throughput gains at Escondida); Freeport‑McMoRan and Anglo American trimmed guidance; Teck Resources and Capstone increased. Despite some positive guidance, a structural deficit (~2–3Mt) persists. Price action is currently driven more by exogenous shocks than demand, and incentive pricing likely needs to remain elevated to bring new supply.
    • Chokepoints: sulfuric acid constraints (transiting Hormuz) hit Chilean production; broader shortage impacts likely emerge over 6–10 months depending on the conflict’s length.
    • Tungsten: critical to munitions and chips; cited as the top‑performing metal over 16 months (300% gain), with supply tightness so acute that certain systems reportedly have two‑week supply buffers.

Legal industry AI fiasco: Sullivan & Cromwell’s AI‑generated brief

The case and the error

  • David: In the liquidation of “The Prince Group” (Cambodia‑based scam conglomerate involved in industrial‑scale phishing/scams; reportedly held significant Bitcoin), Sullivan & Cromwell (S&C) disclosed that a brief submitted to the court contained AI‑hallucinated citations. S&C is a top‑tier “white shoe” firm, not traditionally a bankruptcy powerhouse, but had landed high‑profile insolvency mandates (FTX) via a partner with deep DOJ experience. S&C earned >$230M in the FTX matter (first 20 days >$7.5M), drawing criticism that crypto price appreciation—not lawyering—drove creditor recoveries.

Broader implications: billing, governance, and client risk

  • David: Argues AI use is pervasive among associates and unseen by clients; risk of overbilling for hours that AI compressed is non‑trivial. Clients paying $3,000/hour deserve transparency on how minutes are spent. He favors AI for efficiency but calls for penalties for misuse (not disbarment) and much greater candor and oversight.
  • Brandon: The failure was basic QC—five minutes of human verification would have caught fake citations. He also noted SBF’s early $200k investment in Cursor (liquidated at cost during FTX unwinding) would have been worth ~$2B at a $60B mark; SBF also invested in Anthropic, Solana, and Robinhood—underscoring the portfolio’s latent value had there been no fraud.
  • Justin: Warns of “workflop”—AI‑generated output creating more fix‑work than starting from scratch. He’s seen senior leaders outsource critical thinking to AI for nuanced communications, embracing speed/control over accuracy and ignoring hallucinations. This erodes internal challenge culture (AI rarely pushes back unless prompted), raising strategic risk.
  • Brandon (coding/security): In startup codebases, he frequently sees AI‑generated code with security weaknesses; cites Google’s statement that ~75% of new code is now AI‑written. Speed gains are offset by review/remediation costs; permission misconfigurations and insecure patterns are common.
  • Robert: Referenced a Japanese life insurer (name uncertain in‑session) suing OpenAI in Illinois for practicing law without a license after an individual, guided by AI, filed ~55 actions post‑settlement, compelling the insurer to spend ~$300k in legal fees responding. Case law around AI’s role in “practicing law” may be forming.
  • Eladio: Uses AI extensively but makes final decisions himself; calls AI a “compliance nightmare” that compliance departments won’t quickly catch up to.

Additional AI/security item: Anthropic “Mythos” and banks’ concerns

  • Brandon (closing): Banks, especially in Europe, are in talks with Anthropic to access a newer/secret model referred to as “Mythos.” There was an anecdote of limited access leaking and being used to build simple sites (quickly closed). He reiterated that banks are preparing behind closed doors for AI‑enabled hacks and broader security challenges.

Additional headlines noted but not substantively debated

  • Google unveiled two new AI chips, intensifying competition with Nvidia.
  • SOX (semis) index at all-time highs.
  • Uber reportedly holds >11.5% stake in Lucid, a potential stabilizer for the EV maker.
  • Avis (CAR) experienced a sharp squeeze; allegations of manipulation discussed briefly.
  • Speculation of a Deutsche Telekom and T‑Mobile tie‑up to create the world’s largest wireless carrier; potential FTC/antitrust review questions raised.

Key takeaways and watch items

  • xAI–Cursor: Monitor deal structure details (IP, partner restrictions, existing bulk model contracts) and whether the $10B partnership option is exercised while xAI builds internally. Talent acquisition vs build‑vs‑buy will determine speed to parity with incumbents.
  • Hardware constraints: Expect continued tightness in RAM/VRAM and a two‑tier consumer AI landscape. Watch fab expansion timelines and RAM pricing trends.
  • Retail access to private AI: Robinhood’s OpenAI stake is another proof point of retailization. IPO windows and lockups remain the principal risk vectors for retail outcomes.
  • Walsh/Fed: A shift to trimmed mean inflation could accelerate cuts. Political/legal noise (e.g., issues around Jay Powell referenced in‑session) may affect confirmation timing rather than outcome. Oil‑driven inflation complicates the path.
  • Strait of Hormuz: Even with a ceasefire on land, maritime disruption can sustain physical premia and supply lags. Track real‑world spot premia vs screens, tanker seizures, and flight cancellations/logistics signals.
  • Europe vs US: European equities’ resilience vs reported supply stress is a divergence to monitor; short‑side setups exist per Dave Nicoski’s charts.
  • Commodities: Copper remains structurally short despite mixed guidance; sulfuric acid chokepoints and tungsten constraints are non‑trivial for defense and manufacturing.
  • Legal and software governance: S&C’s AI brief failure is a high‑profile warning. Expect clients to demand AI/billing transparency and firms to implement stricter validation. In engineering, AI‑authored code at scale raises security and compliance risk; review discipline and secure coding standards are essential.
  • Financial sector and AI models: Banks’ interest in advanced models (e.g., “Mythos”) and their operational resilience plans (anti‑hack, data governance) are developing under the radar.

Closing

The session tied together an elevated macro/energy backdrop with acute policy and market structure questions (Fed metrics, IPO windows, retail access), a contested but potentially pivotal AI acquisition/option (xAI–Cursor), and a sober examination of AI’s real‑world failure modes in law and code. The near‑term hinge remains the Strait of Hormuz; policy timing (Walsh, midterms) and physical supply dynamics will shape inflation, rates, and risk appetite across the summer and into year‑end.