• Avoid treating the prospective Anthropic IPO as a buy until the final offering price, share count, conversion terms, and compute commitments are clear; the reported growth does not settle questions about cash needs and returns.
• Approach CoreWeave and Nebius cautiously: investigate financing access, capacity utilization, and potential GPU overcapacity before investing.
• The EWZ selloff may look oversold, but the discussion offered no firm buy recommendation; political and currency risks remain high, and there was no company-specific case for PBR, NU, or MELI.

Anthropic (Private company; prospective IPO)

• The speaker reviewed Anthropic’s IPO filing and described rapid revenue growth: $41 million in 2023, $386 million the following year, and $5 billion last year. He speculated that revenue could reach $40–50 billion this year.
• The speaker said Anthropic was cash-flow positive for the first six months of 2026, but noted that compute-related investment may be off balance sheet.
• The filing reportedly showed $417 billion in aggregate minimum future obligations and arrangements involving vendors, leases, and a variable-interest entity. The speaker questioned where the company’s compute infrastructure is held. The transcript does not establish that the $417 billion is conventional debt.
• The speaker cited AIG, Bristol Myers, and DoorDash among companies providing customer testimonials. He also noted the company’s small board and large employee base.
• He discussed the founders’ ownership and different share classes, but said he was uncertain about the conversion ratios and the resulting share count. Treat his ownership and valuation calculations as unconfirmed.
• The speaker said he did not think the IPO was a good opportunity without greater confidence in management and the offering price. He emphasized that an investment depends on whether management can deploy capital effectively and generate returns.

Takeaways

• Anthropic’s reported growth and customer references are notable, but they do not resolve questions about the cost of compute, future obligations, or the economics of the business.
• Before evaluating any offering, examine the final share count, conversion terms, offering price, and how infrastructure commitments affect cash needs. The speaker’s cap-table math was explicitly uncertain.

AI cloud and GPU infrastructure

• The speaker discussed CoreWeave and Nebius in the context of AI-compute financing and Elon Musk’s warning that other cloud providers may have difficulty obtaining financing.
• He raised the possibility that Musk’s comments could unsettle financing for these businesses, but did not establish that this was Musk’s intent.
• He described the “neocloud” business model as flawed, while also saying that someone has to provide the infrastructure. He thought CoreWeave would survive, but questioned how much value would remain for equity holders.
• The episode referenced Ken Griffin’s warning about potential GPU overcapacity. The speaker agreed that the industry may need to reckon with this issue.
• Hynix was mentioned as trading higher. The transcript gives no investment rationale for the move.

Takeaways

• AI infrastructure companies may benefit from demand for compute, but their financing needs, capacity utilization, and ability to earn returns on expensive infrastructure are key issues raised in the discussion.
• Treat comments about financing stress and overcapacity as risks to investigate, not as confirmed outcomes or a reason by themselves to trade.

Brazil and EWZ

• The EWZ Brazil ETF was described as falling sharply amid election turmoil. The speaker cited a 26% decline, a price around $32 versus $42 the prior week, and what he called its largest single-day plunge. Figures were given during a volatile trading session.
• The speaker said he was not an expert on Brazil and was not trading the market. He later said the selloff seemed oversold, but also said he would not touch Brazilian assets because they were outside his area of expertise.
• He warned that developing markets can experience declines of 50% to 90%, citing currency and foreign-market volatility as risks.
• The transcript includes allegations of election interference; these are claims made during the conversation, not established facts.

Takeaways

• The discussion is a cautionary example of how political uncertainty and currency exposure can magnify volatility in emerging-market investments.
• The speaker did not make a clear buy recommendation on EWZ; his comments combine a possible oversold view with explicit reluctance to invest.

Petrobras (PBR)

• The Brazilian oil company was reported down 15% during the election-related selloff. No company-specific analysis or investment recommendation was provided.

Takeaways

• The move was discussed as part of broader Brazilian-market volatility, not as evidence about Petrobras’s underlying business or long-term value.

Nu Holdings (NU)

• The Brazilian neobank was reported down 22% during the same market turmoil. The speaker offered no company-specific thesis.

Takeaways

• The transcript provides context for a sharp market move, but not enough information to assess the company’s fundamentals or determine whether the decline represents an opportunity.

MercadoLibre (MELI)

• MercadoLibre, described as a Latin American e-commerce company, was reported down 12% during the Brazilian selloff. No specific investment view was given.

Takeaways

• The episode links the stock’s move to regional market turbulence, not to a change in the company’s business outlook.

Moderna (MRNA)

• The speaker said there had been patient deaths in a clinical trial and observed that Moderna initially fell only about 4%, later slipping further before recovering and trading higher.
• He characterized the trial data as “not that great” and told someone to stop shorting the stock, suggesting that they wait for the move to “decay.” No price target or timeline was given.

Takeaways

• The conversation highlights the potential for sharp, changing price reactions to clinical-trial news.
• The speaker’s comments are trading-oriented and do not establish whether the trial results materially change Moderna’s long-term prospects.

IonQ (IONQ)

• The speaker called IonQ “trash” and said it was overvalued at $30. No valuation method or price target was provided.

Takeaways

• This was a clearly bearish opinion, but the transcript offers no supporting company analysis. Consider it a view to investigate rather than a complete investment thesis.

Biotech clinical-trial mention

• The speaker discussed an ENPP3/CD3 bispecific for renal cell carcinoma and cited a 55% response rate, calling the result “pretty great.”
• The company name and ticker are unclear in the transcript, and no follow-up details about the trial, patient group, or durability of response were provided.

Takeaways

• The reported response rate may warrant further research, but it is not enough on its own to assess a drug candidate or identify an investable company.

Tesla (TSLA) and Bloom Energy (BE)

• Tesla was noted as up about 6% during the session, without further explanation.
• Bloom Energy was described as making a small comeback, up about 3%. No investment thesis was offered for either company.

Takeaways

• These were brief market updates, not specific recommendations.

13F holdings and stock research

• The speaker used institutional 13F filings to show example holdings, including Amazon, Alphabet (Google), CRH, TDS, Live Nation, Keysight, Somnigroup, Block, Flex, Capital One, Taiwan Semiconductor, Lam Research, Micron, Apple, American Express, Coca-Cola, Bank of America, Chevron, Occidental Petroleum, Chubb, Moody’s, and Kraft Heinz.
• He said that reviewing institutional holdings can generate ideas for further research, but warned against simply copying a fund’s positions. He noted that investors have different costs of capital and objectives.

Takeaways

• Use reported holdings as a research starting point, not as an automatic buy list. Check the company’s fundamentals, valuation, and fit with your own investment goals.