[BidClub_]
The Shkreli Pill · · 26 min

Q3 Is Going to Shock the Market. I'm Extremely Bearish on Bitcoin & Space Stocks | Martin Shkreli

Martin Shkreli

EquitiesSemisBiotechInvesting
YouTube ↗
TL;DR
  • Shkreli expects NVIDIA’s Q3 guidance to deliver a market-wide shock. His thesis is that removing Singapore revenue he regards as China sales, alongside circular financing that he says only Jensen Huang supported, could force a sharp reset. “The bank is closed, my friend.” He says he would sell, while leaving open the possibility of Huang returning.
  • He is bearish on Bitcoin and space stocks, with a telecom-related exception. He supplies no Bitcoin valuation or catalyst, while describing SpaceX as “a Neo cloud selling GPUs, renting GPUs.” His broader AI-infrastructure stance remains fluid: “I change my mind about this daily, if not even more frequently than that.”
  • Lilly’s valuation requires GLP revenue to triple and hold in Shkreli’s model. With shares near $1,500, an 8% discount rate produces roughly $1,200 even assuming GLPs last indefinitely. He owns none and concedes tripling is possible, but questions saturation, product substitution and generics: “GLP revenue has to triple and stay there.”
  • Strong clinical results can still disappoint an expectant stock. Shkreli calls CG Oncology’s NMIBC recurrence-free-survival result “Really nice data,” citing a 0.32 hazard ratio, yet shares retreat from about $100 that morning to $80. His explanation is tentative: “It sounds like it was highly expected.”
  • CoreWeave’s credit pricing adds another concern beside its equity retreat. With shares back around $50, Shkreli reads his CDS screen as putting Argentina and CoreWeave at “similar default probability.” His reaction: “Yikes.”
  • His bearish calls coexist with selective long ideas and limited trading activity. “Not really focused on trading right now,” Shkreli says, describing his startup as his day job. He still sees compute and memory as an exciting trend, calls Meta “a good long,” and favors healthcare during cyclical stress.
Digest · the substance, structured for research

1. NVIDIA’s guidance is the market-wide risk

  • Shkreli’s Q3 warning rests on a possible “kitchen sink quarter”: NVIDIA might remove Singapore revenue he regards as China sales and abandon circular financing. Without Huang selling into China and supporting those deals, he expects weak forward guidance to hit the stock dramatically. “The bank is closed, my friend.”
  • His specific management scenario has Colette resetting expectations so sharply that NVIDIA and the broader market fall, creating pressure to bring Huang back. He imagines the response: “No, no, no, no, no, let's get Jensen back.”
  • In his account of the political dispute, Trump and Lutnick support Huang while Bessent and Rubio oppose him. He considers a return possible, but doubts it happens before the next day’s results.
  • The live tape repeatedly challenges the bearish stance: NVIDIA recovers toward flat around $162 before falling below $156, down roughly 4.2%. SPY later drops 2%. Shkreli keeps returning to guidance as the decisive catalyst.

2. AI infrastructure remains a moving target

  • Challenged about earlier bullishness, Shkreli responds, “Who said I was bullish on AI infra?” His qualification matters: “I change my mind about this daily, if not even more frequently than that.” He nevertheless calls compute and memory an exciting trend, while describing photonics as competitive.
  • His hardware view still favors GPUs, with TPUs and other parallel ASICs also participating as “GPU-like systolic arrays.” He accepts that Gemini does not run mostly on NVIDIA and says he is spending less time on optics and photonics.
  • CoreWeave’s rebound has reversed, leaving shares around $50 again; he remains negative on Nebius and IREN. Later, reviewing newly available CDS data, he places Argentina and CoreWeave at “similar default probability.” “Yikes” is his entire immediate assessment.

3. Lilly’s growth still struggles to justify its valuation

  • Shkreli owns no Lilly despite shares approaching $1,500. He cites quarterly GLP revenue around $17.1 billion versus $16.2 billion previously, and company revenue growth of 36%. Q4 will “probably beat my number,” he says, while still anticipating slower growth.
  • His model treats $65 billion of revenue as “basically in the bag.” For 2027, he weighs outcomes around $80 billion–$90 billion, describing growth cases of approximately 18% and 30%. The question is how long that expansion can persist.
  • He initially thinks Retatrutide is already filed, then corrects to a Q1 filing. He asks whether using a PRV could permit approval later that year and Q4 revenue. “Retatrutide's gonna be a monster, huh?” remains a question, alongside uncertainty about substitution for Mounjaro.
  • Even assuming GLPs last forever, an 8% discount rate gives him roughly $1,200 per share. His aggressive modeling includes Retatrutide at $62 billion and Mounjaro at $80 billion, but he struggles with $150 billion in GLP sales. “GLP revenue has to triple and stay there.” He concedes that is possible; saturation and eventual generics remain unresolved.

4. Biotech enthusiasm comes with clinical and regulatory limits

  • Looking at CG Oncology, Shkreli highlights a 0.32 hazard ratio for recurrence-free survival in NMIBC and describes an intravesical gene therapy/oncolytic immunotherapy. “Really nice data,” he says, but the stock’s retreat from about $100 to $80 suggests to him that strong results were already expected.
  • He likes Sarepta while retaining the qualifier “Still speculative.” The uncertainty is fundamental: “you never know if a drug works or not.”
  • He predicts “Capper” will get a CRL and points to Summit, which he says also got a CRL after the FDA told them they were up. He calls that “par for the course.”

5. Bitcoin and space draw bearish calls with selective exceptions

  • “I'm kinda bearish on Bitcoin,” Shkreli says, then calls himself “pretty bearish on space, other than telecom-related.” He offers little additional Bitcoin reasoning. His SpaceX framing instead emphasizes its exposure to renting GPUs as a neocloud.
  • Shkreli says his startup occupies most of his attention rather than trading. His exceptions to the bearish tone include Meta, which “seems like a good long,” money markets as “always a good play,” and healthcare during cyclical stress. He gives no indication of a corresponding Meta position.
Full transcript