• Consider the long SanDisk (SNDK) / short Intel (INTC) pair as a relative-value trade, but monitor upcoming Micron (MU) and Meta (META) commentary for evidence of sustained memory demand and enterprise AI adoption.
• Kodiak Sciences (KOD) is a highly speculative short ahead of Monday’s trial data; avoid treating it as a low-risk trade, since a positive surprise could cause substantial losses.
• Avoid broad AI-related shorts based solely on high valuations: the speaker expects IPO-driven enthusiasm could persist through 2028, and warns that timing can be costly.

Akamai (AKAM)

• Akamai announced a deal with Anthropic, which the speaker called good news, but “not that good.” The stock initially faded after the news, rallied, and then fell back during the session.
• The speaker said he went short and later exited. He watched levels around $125, $119, and $116, but said he did not really believe in chart levels.
• He described the trade as successful, while also regretting that he covered before the stock fell further.

Takeaways

• The discussion treats the Anthropic deal as a short-term trading catalyst, not proof of a lasting improvement in Akamai’s business.
• The speaker’s position changed quickly; the mentioned prices are live trading references, not price targets or an enduring recommendation.

SanDisk (SNDK), Intel (INTC), and AMD (AMD)

• The speaker favored a long SanDisk / short Intel pair, saying memory was still “king of AI” and dismissing the current enthusiasm for AI agents as played out.
• He also discussed a long AMD / short Intel pair. Later, he clarified that Keith was long AMD and short Lam Research—not the reverse.
• The speaker said AMD had strong momentum but might need a few days to cool off after a large run. He also noted that AMD’s CEO had dinner with President Trump, which he viewed as bullish.
• He emphasized that the outlook for CPUs depends partly on how AI agents develop and what companies report about enterprise demand. He said upcoming company commentary and results, including from Meta and Micron, could help clarify the trend.

Takeaways

• These were proposed or discussed relative-value trades, not straightforward calls that every semiconductor stock would rise or fall.
• The memory-over-agents thesis is the speaker’s view. Investors considering the theme should watch for evidence on actual AI-related memory and CPU demand rather than relying on enthusiasm or skepticism alone.

ASML (ASML)

• The speaker noted a short position in ASML as part of a pair-trading approach and described it as a possible mean-reversion trade.
• He said mean reversion can work but requires discipline.

Takeaways

• The short was presented as a trading strategy, not a claim that ASML’s business is weak.
• Mean-reversion positions can move against a trader if a strong trend continues; the transcript provides no target or holding period.

Lam Research (LRCX)

• Keith was described as short Lam Research while long AMD. The speaker also noted Lam sells equipment to companies seeking to manufacture semiconductors, including Elon Musk’s companies and Micron.
• In discussing semiconductor investing, the speaker stressed understanding the full supply chain, including chipmaking equipment and testing.

Takeaways

• The position was part of a pair trade, so it should not be read as a standalone bearish recommendation on Lam Research.
• The transcript offers no valuation analysis or price target for LRCX.

Broadcom (AVGO), KLA (KLAC), and Applied Materials (AMAT)

• The speaker suggested following Broadcom and mentioned shorting AVGO or KLA as possible examples of research-driven positions.
• Applied Materials and KLA came up while discussing semiconductor manufacturing steps and the companies that supply or test chipmaking equipment.
• No specific investment case, entry point, or target was given for these stocks.

Takeaways

• These names were raised as areas for further research, not as firm recommendations.
• The discussion supports examining the semiconductor equipment supply chain, but does not establish which suppliers are attractive at current prices.

Micron (MU), TSMC (TSM), SK Hynix, and NVIDIA (NVDA)

• Micron and SK Hynix were mentioned in the context of semiconductor demand and the memory supply chain. The speaker said Micron’s future results could help clarify the industry outlook.
• TSMC, SK Hynix, and NVIDIA also appeared in discussion of semiconductor suppliers and Keith’s trading activity; no clear directional view was provided for them.

Takeaways

• The speaker sees memory and broader semiconductor demand as important to the AI investment story.
• The transcript does not give enough company-specific analysis to support a buy or sell conclusion for these names.

Meta Platforms (META) and Microsoft (MSFT)

• The speaker observed that Meta was down after a large run while Microsoft was rallying.
• He also said company commentary, including from Meta, could help investors assess whether AI agents are translating into meaningful enterprise demand.

Takeaways

• The price moves were brief market observations, not a stated change in investment view.
• For the AI theme, the speaker’s comments point to watching adoption and business results rather than treating short-term price action as confirmation.

Kodiak Sciences (KOD)

• The speaker said he was short KOD ahead of trial data expected on Monday and was strongly skeptical of the company’s eye-disease drug.
• He argued that the drug had failed to show sufficient benefit versus established treatment, referring to Eylea and a biosimilar comparison. He also criticized the trial design and said that more frequent dosing would be unattractive to physicians and patients.
• He characterized the event as highly binary: he estimated a very small chance of a dramatic upside outcome—described as less than 1%—but said such an outcome could potentially more than double the stock. He also discussed downside scenarios that could materially reduce the portfolio’s gains.
• The speaker said the position could be large enough to take the portfolio close to its starting point if the outcome went badly. He debated whether to increase or reduce the position, making clear that the decision had not yet been finalized.

Takeaways

• This was the transcript’s clearest high-risk, event-driven trade. The speaker’s bearish view is his own and depends heavily on his interpretation of the clinical evidence.
• A short position can lose substantially if trial results surprise positively; the speaker explicitly acknowledged that risk. The discussion is not a low-risk investment thesis or a settled recommendation.

Moderna (MRNA)

• The speaker referred to trading Moderna on both the long and short sides and said volatility had created opportunities to trade in and out.
• He mentioned the stock around $190 and recalled missing a move near $208, but did not give a lasting bullish or bearish view.

Takeaways

• The comments describe active trading rather than a fundamental investment thesis.
• The transcript provides no price target or specific recommendation for MRNA.

Viking Therapeutics (VKTX)

• The speaker noted that VKTX had been difficult to buy around $35 and later traded near $39, before falling back toward $38.
• The comments focused on the price movement and trading activity; no company-specific thesis was offered.

Takeaways

• The cited prices are observations from the session, not targets.
• The transcript does not provide enough information to assess VKTX’s investment outlook.

Biohaven (BHVN)

• The speaker said, “I like Biohaven,” but did not explain why or provide a valuation view.

Takeaways

• This is a brief positive remark, not a developed investment thesis.
• Further company and valuation research would be needed before drawing an investment conclusion.

Datadog (DDOG)

• The speaker asked why Datadog was up and joked about the move, but gave no explanation or investment view.

Takeaways

• No actionable thesis was provided for DDOG.

IonQ (IONQ)

• The speaker called IonQ “kind of a scammy company” and said he had covered his short earlier that day, adding that the discussion might prompt him to short it again.
• He did not explain the company-specific basis for the criticism.

Takeaways

• The comments reflect a negative opinion but not a fully supported short thesis.
• The speaker had already changed his position, so this should not be treated as a current or firm trade recommendation.

Dell (DELL)

• The speaker floated Dell as a possible short, arguing that assembling computer systems may not provide much lasting value and that the stock had risen substantially.
• He suggested that this part of the business could be “computed away,” but offered no detailed financial analysis.

Takeaways

• This was a speculative short idea, not a firm recommendation.
• The thesis would need to be tested against Dell’s actual margins, customer relationships, and business mix.

Cerebras Systems

• The speaker mentioned Cerebras as a potential longer-term short idea, but did not provide supporting analysis or a ticker.

Takeaways

• Treat this as a research lead only. The transcript does not establish a valuation case or explain how investors could access the opportunity.

OpenAI, Anthropic, and the AI Investment Theme

• The speaker predicted that the anticipated OpenAI and Anthropic IPOs could fuel a prolonged market mania, possibly lasting until 2028. He said markets could keep rising or move sideways for the next 18 months.
• He also cautioned against rushing into a market-wide bearish trade just because valuations or sentiment seem excessive. Recalling that he was too early in earlier market downturns, he advised thinking through a thesis before putting on a position, especially a large one.
• Elsewhere, he was skeptical that AI agents were a durable driver of semiconductor demand, while acknowledging that enterprise adoption and future company commentary would matter.

Takeaways

• The transcript presents competing views: potential continued AI-driven enthusiasm, but uncertainty about whether agents will translate into durable demand.
• The speaker’s warning about being early is especially relevant to thematic shorts: a valid long-term concern may still lose money if timing is poor.

Canton Network

• The speaker said he had seen something about the Canton Network, describing it only as a crypto-related project. He offered no opinion on its prospects and mentioned no specific cryptocurrency or token.

Takeaways

• No investment thesis or actionable crypto recommendation was provided.

Overall Trading and Risk Lessons

• The speaker said he trades frequently and warned viewers not to assume a position will still be open the next day.
• He emphasized limiting losses, firing “non-performing” positions, and avoiding large positions based on an untested feeling.
• He specifically warned that shorting can create severe losses if a stock rises sharply, and said investors should not put an entire portfolio into one short.
• His comments about trading skill and market intuition were personal opinions; the transcript does not establish that the strategies are suitable for every investor.