A high-conviction sell recommendation was issued for SELLAS Life Sciences (SLS), with the analyst stating the company's drug "does not work" and urging investors to exit the position immediately. Investors are advised to avoid quantum computing stocks like IonQ (IONQ) and D-Wave (QBTS) for AI applications, as the technology is considered a "dead end" for these workloads. A bearish sentiment was expressed for Viking Therapeutics (VKTX), with the analyst suggesting its positive run is likely over. Despite strong fundamentals, Eli Lilly (LLY) may have limited near-term upside from its $1,000 price, with valuation models suggesting a target of around $1,100. The key long-term theme is identifying the successor to NVIDIA (NVDA), with photonic computing seen as a more promising field than quantum.

Eli Lilly (LLY)

• The stock was mentioned as being at an "unbelievable" $1,000 a share.
• The discussion centered on the "GLP wars" between Novo Nordisk (NVO) and Eli Lilly (LLY), with the speaker noting that Novo "looks dead on arrival" while Lilly "looks fantastic."
• Munjaro consensus sales estimates have reportedly risen from $18-20 billion to $30 billion.
• A key future product is Orforglipron (brand name likely Orpho), an oral GLP-1 pill.
  • Bank of America is reportedly bullish on the Orpho launch in Q2.
  • Visible Alpha projects Orpho sales of $19 billion in 2034.
  • The speaker notes that a successful, cheap oral pill could be a "GLP gateway drug that every single person in America takes."
• Risks:
  • Pharma companies can be "a victim of its own success" due to the patent cliff. Once patents expire, revenue from blockbuster drugs can disappear.
  • The company must successfully reinvest profits from hit drugs like Munjaro to ensure long-term growth. The speaker points to Pfizer (PFE) as an example of a company that failed to reinvest its COVID profits effectively.
• A model mentioned in the podcast values LLY at $985 per share, but after adjusting for a smoother patent cliff landing for Munjaro and Zepbound, the valuation rises to $1,100 per share. The speaker concludes, "It's hard to see too much upside here."

Takeaways

• Eli Lilly has seen incredible success with its GLP-1 drugs, driving the stock to $1,000.
• The upcoming oral pill, Orpho, is seen as a major future catalyst with massive market potential, though it may cannibalize sales of their injectable drugs.
• Despite the success, the speaker expresses caution about significant further upside from the current price, citing the inherent risk of patent expirations in the pharmaceutical industry and the need for management to reinvest profits wisely. A valuation of $1,100 is mentioned as a potential target.

NVIDIA (NVDA)

• NVIDIA is described as the dominant force in computing, with a market cap approaching $5 trillion.
• Its success is driven by the massive demand for AI, which is currently 98-99% of computer workloads. The core task for AI is matrix multiplication (MatMul), which NVIDIA's GPUs are exceptionally good at.
• The central investment theme of the podcast is the search for the "next NVIDIA" – a company that can create a new, more efficient type of computer for AI workloads.
• The speaker believes NVIDIA's dominance is "not going away anytime soon" and that NVIDIA itself is his "biggest worry," as they will not slow down innovation.

Takeaways

• NVIDIA is the undisputed king of the AI hardware market, a $5 trillion behemoth.
• The investment opportunity is not necessarily in betting against NVIDIA, but in identifying the next paradigm of computing that could either complement or eventually challenge its GPU architecture.
• Any company that can create a significantly faster way to perform MatMul has the potential for "thousand X" returns, similar to early investments in NVIDIA.

Quantum Computing (IONQ, QBTS)

• The speaker is extremely bearish on quantum computing for AI and general-purpose workloads.
• He argues that people are only interested in quantum stocks like D-Wave (QBTS) and IonQ (IONQ) because they are publicly traded "lotto tickets" for the "next NVIDIA," not because the technology is viable for AI.
• Technical limitations for AI:
  • No Speedup: Quantum computers offer no significant speed improvement for the dense matrix multiplications (MatMul) used in AI.
  • Slow Clock Speeds: Quantum computers operate at 1 megahertz, compared to 2 gigahertz for GPUs.
  • Few Cores: They have essentially one core, compared to 15,000+ in a high-end GPU.
• The only area where quantum computing has a "massive" advantage is in decryption (using Shor's algorithm), but this represents a tiny fraction of global computing needs.
• The speaker concludes that quantum computing is a "dead end" for the problems the market cares about and that basing an investment on the hope of future undiscovered algorithms is a poor thesis.

Takeaways

• Avoid investing in quantum computing stocks with the expectation that they will power the next wave of AI. The underlying technology is not suited for the task.
• The hype around quantum stocks is driven by a retail search for high-risk, high-reward plays, but the fundamental case is weak compared to other next-generation computing approaches.
• The one true use case, breaking encryption, is a niche market and not enough to justify the valuations or hype for a general-purpose computing revolution.

QCLS (QCLS)

• This is a publicly traded company that the speaker is personally involved with.
• He describes it as a "brand new company" or a "startup" that he is "pretty excited about."
• The company's goal is to build an optical/photonic computer. The speaker believes this is a long-term project that "could take a decade."

Takeaways

• QCLS is a high-risk, long-term bet on photonic computing, an alternative to traditional silicon and quantum computing.
• The speaker is directly involved, which is a vote of confidence but also means he is limited in what he can disclose. Investors should be aware of this affiliation.

Investment Theme: Next-Generation Compute

• The core thesis is that the world needs a faster, more efficient way to perform AI computations (MatMul) than what even NVIDIA's GPUs can offer, creating a race to build the next computing platform.
• Photonic/Optical Computing: This is the speaker's preferred bet.
  • Theory: It could potentially perform matrix multiplication in linear time (O(n)) or even constant time (O(1)), a massive leap over the polynomial time (O(n³)) of GPUs.
  • Challenges: Major engineering problems remain, including digital-to-analog conversion (DAC/ADC) and achieving non-linearity (which is essential for AI activation functions like ReLU).
  • Companies: QCLS (QCLS) is the public company mentioned. Private companies include Light Intelligence and Lightmatter.
• Analog Computing: A related field seeing a comeback.
  • Company: Unconventional, a private startup, was mentioned for raising $500 million at a $5 billion valuation, showing massive venture capital interest in this space.
• Biological Compute: A "wild card" idea using living cells.
  • Theory: The human brain computes with only 20 watts of power, versus 1,000 watts for a high-end GPU.
  • Companies: Private companies FinalSpark and Cortical are exploring this by growing neurons on a dish.
• Geopolitical Motivation: The speaker highlights that the world's most advanced chips are made by TSMC in Taiwan, which is "right next to China." This creates a major national security incentive for the U.S. to develop a new, domestic computing manufacturing base.

Takeaways

• The biggest long-term opportunity in technology may be in finding the successor to the GPU for AI workloads.
• While public market options are limited, the most promising fields appear to be photonic and analog computing, not quantum.
• Keep an eye on private companies like Unconventional, Light Intelligence, FinalSpark, and Cortical, as they may become public in the future if their technology proves viable. The high valuations in the private market indicate strong belief in this theme from sophisticated investors.

SELLAS Life Sciences (SLS)

• The speaker expressed an extremely bearish view on this stock.
• He stated, "I can't believe the stock's at five," and mentioned a 400% borrow rate, which indicates very high short-selling interest.
• He believes the company's drug "does not work" and that the company has not reported study data in years.

Takeaways

• The speaker issued a strong warning to investors, advising them to "sell everything on SLS" and "get the fuck out of it ASAP." This is a highly bearish, high-conviction call.

Other Stock Mentions

• Optical Interconnect Stocks (LITE, COHR, FN):
  • The speaker noted that Lumentum (LITE), Coherent (COHR), and Fabrinet (FN) were all down 5% to 10% on the day of the podcast.
  • Takeaway: These are publicly traded companies in the optics space that could be worth monitoring as part of the broader photonic computing theme, though they are currently focused on interconnects.
• Avivax (AVVX) and Sarepta (SRPT):
  • The speaker briefly mentioned that he "still like[s] them."
  • Takeaway: A continued bullish but brief mention. No new details were provided.
• Viking Therapeutics (VKTX):
  • The speaker stated, "I think Viking is kind of cooked, man."
  • Takeaway: A bearish sentiment on the stock, implying its positive run may be over.