• Consider NVIDIA (NVDA) as the strongest large-cap AI opportunity: the speaker cites roughly 70% forward growth, strong margins, and attractive relative valuation, though no price target is provided.
• For AI memory exposure, Micron (MU) is the speaker’s preference for U.S. investors over SK hynix, whose U.S.-accessible shares are said to carry an approximately 40% premium; Micron’s AI-demand outlook is promising, but the analysis is preliminary.
• Nebius (NBIS) is a high-conviction growth idea after its sell-off, supported by differentiated neocloud offerings and strong growth and margins; no price target is available.
• IREN may offer deeper value, but weigh its Texas project concentration and management criticism against the speaker’s bullish view; both neocloud stocks could remain volatile.

Nebius (NBIS)

• The speaker describes Nebius as a differentiated “neocloud” company, citing sovereign AI offerings, certifications, diversification, and leadership with experience at Yandex.
• The speaker says Nebius is still renting older H100 GPUs at premium prices, arguing that older hardware has not become worthless.
• Nebius is described as trading at 0.11 on the speaker’s enterprise-value-to-gross-profit-to-revenue-growth metric, with reported EBITDA margins around 50% and very strong expected growth.
• The speaker sees the sell-off as likely driven more by contagion and volatility than by a change in Nebius’s fundamentals.

Takeaways

• The speaker’s view is strongly bullish: they see Nebius as attractive at the cited valuation, based on growth, margins, and business differentiation.
• The transcript does not provide a price target. The speaker flags broad market volatility and notes that concerns about CoreWeave’s debt should not automatically be applied to every neocloud company.

IREN (IREN)

• IREN is described as a neocloud company with more power ownership and a repeatable development process, but less differentiation than Nebius.
• The speaker says IREN is heavily reliant on two large Texas developments, making its business more concentrated geographically.
• Management has faced more criticism than Nebius’s, according to the speaker.
• The speaker cites a valuation of 0.07 on their enterprise-value-to-gross-profit-to-revenue-growth metric and describes IREN’s growth and “Rule of 40” profile as unusually strong.

Takeaways

• The speaker sees IREN as potentially very cheap, but investors should weigh that view against the company’s Texas project concentration and the management criticism mentioned in the transcript.
• The speaker argues that IREN’s sharp decline appears inconsistent with its fundamentals, but offers no price target.

NVIDIA (NVDA)

• The speaker says NVIDIA fell about 6% over five days and has struggled to reach new highs, despite remaining the AI-chip leader.
• They cite a forward growth rate of about 70%, a Rule of 40 of 135, and an EBITDA margin of about 65%. The speaker also says demand is supply-constrained: NVIDIA could grow faster if it could make more GPUs.
• NVIDIA is valued at 0.32 on the speaker’s enterprise-value-to-gross-profit-to-revenue-growth metric, which they describe as the cheapest among the “Magnificent Seven” stocks they cover.
• The speaker argues that the market’s negative reaction reflects volatility and sentiment more than a deterioration in NVIDIA’s business.

Takeaways

• The speaker is bullish on NVIDIA, viewing its leadership, growth, and margins as attractive relative to its valuation.
• The transcript provides no price target. Its broader caution is that the market is volatile and, in the speaker’s view, not consistently trading on fundamentals.

Micron Technology (MU)

• The speaker says Micron fell about 15% despite reporting what they describe as stellar quarterly results.
• They cite demand for high-bandwidth memory (HBM), which is needed to help GPUs process workloads faster, as an important AI-growth driver.
• The speaker refers to research indicating that major contracts could support strong growth into the 2030s, while noting they had not yet completed their own full analysis.
• Micron is cited at 0.08 on the speaker’s valuation metric, with a “Rule of 40” near 180. The speaker considers it more attractive for many U.S. investors than SK hynix, which they say carries an approximately 40% premium when bought in the U.S.

Takeaways

• The speaker is bullish on Micron’s AI-memory opportunity and sees the sell-off as hard to reconcile with the company’s reported results and demand outlook.
• The speaker had not completed a full analysis, so the valuation and growth figures presented are their preliminary view, not a comprehensive assessment.

SK hynix

• The speaker compares SK hynix’s “Rule of 40” with Micron’s, saying the figures are similar.
• They say SK hynix may look more attractive to investors in Korea, but that U.S. investors may face an approximately 40% premium relative to Micron.

Takeaways

• The speaker’s preference depends partly on where an investor can buy the shares: they favor Micron for U.S. investors unwilling to pay the premium they cite for SK hynix.
• No price target or separate company-specific risk is discussed.

Advanced Micro Devices (AMD)

• The speaker acknowledges that some investors like AMD but says they do not understand its valuation.
• They describe AMD as twice as expensive as NVIDIA on their valuation comparison, despite NVIDIA being the faster-growing company.
• The speaker suggests some investors may be attracted to AMD’s smaller market capitalization and the possibility it could reach a trillion-dollar valuation.

Takeaways

• The speaker is bearish relative to NVIDIA, arguing that AMD’s valuation is difficult to justify against the AI leader’s growth.
• This is a relative-valuation opinion; the transcript does not provide a price target.

Broadcom (AVGO)

• The speaker says Broadcom is more expensive than NVIDIA on their valuation comparison and cites a value of 0.56 on their metric.
• They suggest investors may be drawn to Broadcom because its smaller market capitalization could appear to offer more room to double.

Takeaways

• The speaker favors NVIDIA’s valuation over Broadcom’s, given NVIDIA’s leadership and the speaker’s view that it is cheaper.
• No price target or specific company risk is mentioned.

Super Micro Computer (SMCI)

• The speaker describes Supermicro as consistently cheap, citing a valuation of about 0.1 on their metric, alongside what they characterize as fast growth.
• They say the company’s association with China may be one reason investors discount it.

Takeaways

• The speaker sees SMCI as a low-valuation AI-hardware opportunity, but the transcript identifies investor concern about its China association as a possible reason for that discount.
• No price target is provided.

Dell Technologies (DELL)

• The speaker says Dell fell about 22% in the week discussed, after having risen substantially beforehand.
• They argue that Dell’s servers remain important to AI infrastructure and view the decline as excessive.

Takeaways

• The speaker sees the sell-off as an opportunity to reconsider Dell’s role in AI-server infrastructure, but provides no valuation estimate or price target.
• The transcript does not discuss Dell-specific risks.

CoreWeave (CRWV)

• The speaker says some investors attribute neocloud-sector weakness to concerns about CoreWeave’s debt.
• They characterize CoreWeave’s debt situation as unusually high and specific to the company, and argue against applying that concern to all neoclouds.

Takeaways

• The speaker’s point is that CoreWeave’s financing concerns should be assessed on their own rather than treated as proof that every neocloud has the same problem.
• The transcript does not provide a valuation or investment recommendation for CoreWeave itself.

Broadcom, Credo, Marvell, Lumentum, and Astera Labs

• The speaker groups Credo (CRDO), Marvell (MRVL), Lumentum (LITE), and Astera Labs (ALAB) among AI hardware companies tied to potential future bottlenecks, including connectors, lasers, and photonics.
• They say these “next-wave” AI hardware stocks look expensive compared with current AI leaders such as NVIDIA and Micron.

Takeaways

• The speaker prefers the current AI leaders at the valuations discussed, rather than paying up for hardware companies tied to a later stage of AI expansion.
• No price targets or company-specific risks are mentioned for these names.

Taiwan Semiconductor Manufacturing Company (TSMC) and ASML

• The speaker says TSMC and ASML are too expensive for their preferences, without giving specific valuation figures.

Takeaways

• The speaker is cautious on these companies based on valuation, but the transcript does not provide further analysis or price targets.

Hims & Hers Health (HIMS)

• The speaker uses Hims & Hers as an example of a growth stock they have previously considered cheap.
• They say its share-price decline resembled the declines in neocloud stocks, despite Hims & Hers having little direct connection to the AI-related news being discussed.

Takeaways

• The comparison supports the speaker’s view that the sell-off across growth stocks may reflect broader contagion rather than company-specific developments.
• The transcript does not provide a new valuation estimate or price target for Hims & Hers.

Coca-Cola (KO), AT&T (T), Verizon (VZ), Visa (V), Mastercard (MA), and JPMorgan Chase (JPM)

• The speaker says older consumer, telecom, payments, and banking companies held up or rose during the sell-off, while AI-related and other growth stocks fell.
• They question the market’s willingness to value some established businesses as if their competitive advantages will persist indefinitely.
• For AT&T and Verizon, the speaker points to direct-to-satellite phone and internet services as possible future disruption to traditional connectivity providers, which they say have substantial debt and cell-tower assets.

Takeaways

• The speaker sees potential long-term disruption risk for traditional telecom providers from direct-to-satellite services, while questioning why these companies rose during the week.
• The transcript offers no price targets or specific recommendations for these companies.

SpaceX (Private company)

• SpaceX is mentioned as a direct-to-satellite disruptor and as having fallen during the same market sell-off.
• The speaker contrasts the decline in potential disruptors with gains in some established telecom companies.

Takeaways

• The discussion highlights direct-to-satellite connectivity as a potential competitive threat to traditional telecom services.
• SpaceX is private, and the transcript provides no valuation or investment route.

AI Infrastructure and Neoclouds

• The speaker argues that AI demand remains strong, particularly for computing capacity, tokens, GPUs, and memory. They say price per token has fallen, but total token demand and spending have not fallen in the same way.
• They describe the recent declines in AI-related stocks as likely reflecting panic, contagion, or possible margin calls rather than worsening fundamentals.
• The speaker also cautions against claims of a coming “Black Monday,” saying they do not know what will happen next and consider the market unusually volatile.

Takeaways

• The discussion presents a bullish long-term view of AI infrastructure demand, but emphasizes that short-term market moves may remain unpredictable.
• The speaker’s preferred approach, as expressed in the transcript, is to focus on what they consider relatively inexpensive current AI leaders rather than assuming every AI-related stock is attractive.

Direct-to-Satellite Connectivity

• The speaker says direct-to-satellite phone service is already working in emergency situations and that satellite internet could threaten traditional telecom businesses over the next year or two.
• They describe telecom companies with debt and cell-tower assets as potentially vulnerable to this disruption.

Takeaways

• Direct-to-satellite connectivity is presented as a possible investment theme and a competitive risk for traditional telecom providers.
• The transcript does not name a specific investment recommendation or provide a timeline beyond the speaker’s reference to the next year or two.