• CoreWeave (CRWV) is the strongest actionable watchlist idea: its expanded cloud offering and reported margin improvement are promising, but look for evidence of sustained customer demand in a crowded market before investing.
• NVIDIA (NVDA) remains strategically supported by its CUDA ecosystem, while early feedback suggests DGX Cloud Lepton has yet to prove it can outperform established cloud providers; monitor product adoption and partner relationships.
• Nebius (NBIS) received a favorable product-level comparison with Lepton, but the discussion provides too little financial or valuation information to support a buy decision.
• Treat Alphabet (GOOGL)’s Gemini benchmark results as preliminary until real-world adoption or business impact is demonstrated.

NVIDIA (NVDA)

• NVIDIA’s original plan for DGX Cloud Lepton was to aggregate compute and compete more directly with AWS, Azure, and Google Cloud.
• The product instead evolved into a software platform for managing GPU workloads across providers. Customers can still use their own hardware and maintain relationships with providers such as CoreWeave and Nebius.
• The hosts said early reviews and testing did not show Lepton outperforming direct use of some other cloud platforms. They viewed NVIDIA’s software and CUDA moat as important strategic advantages, but said NVIDIA still had work to do.
• The discussion suggested NVIDIA has found it strategically valuable to support many cloud providers that, in turn, buy NVIDIA products.

Takeaways

• The discussion points to NVIDIA’s software ecosystem—not only GPU sales—as a key part of its competitive position.
• Watch whether Lepton improves enough to attract customers without alienating cloud partners. The transcript portrays the platform’s early performance as a limitation, not a proven threat to existing providers.

CoreWeave (CRWV)

• CoreWeave launched CoreWeave Forge, expanding its cloud offering with CPUs, storage, and management tools. The hosts described the move as a way to compete more broadly with other “NeoCloud” providers.
• They characterized CoreWeave as a strong business with rapid margin expansion. They also said it had addressed depreciation concerns quickly, while the value it can generate per watt was increasing.
• The hosts described a crowded competitive field, citing roughly 300 NeoClouds. They said CoreWeave was aiming to reach a higher tier on the ClusterMax rankings.
• The discussion referenced $60 billion per gigawatt as a benchmark associated with comments from Elon Musk and Jensen Huang. It was not presented as a CoreWeave forecast or price target.

Takeaways

• CoreWeave’s expansion beyond GPU access could help it offer a more complete cloud platform, but it is competing in a crowded market.
• Monitor whether the company sustains margin expansion and turns its broader product offering into durable customer demand. The transcript’s positive comments are not a guarantee of future results.

Nebius (NBIS)

• Nebius was named among cloud providers that, according to the hosts, offered a better experience than NVIDIA Lepton in the testing discussed.
• The transcript provided no further detail on Nebius’s financial performance, valuation, or growth outlook.

Takeaways

• The discussion offers a limited, product-level positive comparison for Nebius, not a broader investment case.
• Further research would be needed to assess its business fundamentals and competitive position.

Instinct (Private company)

• Instinct was testing product recommendations through text messages. The hosts discussed a possible future business model involving retailer partnerships and commissions on purchases.
• The hosts suggested that highly relevant recommendations could be valuable, but said the examples circulating online included poorly matched or repetitive suggestions. They also questioned whether unsolicited product messages might feel pushy.
• They described Instinct as a small team of roughly 13 or 14 people and said the company was growing quickly and would likely need to raise capital. They also noted that its business was compute-intensive.
• A potential claim that recommendations might convert at seven times the rate of Instagram ads was discussed as a possible proof point for a future fundraising pitch—not as a measured result.
• The company had hired a former Snapchat executive to lead business efforts, which the hosts viewed as a potential boost to its commerce and partnerships strategy.

Takeaways

• The opportunity discussed is the potential for personalized AI recommendations to generate new shopping demand and revenue.
• Key risks raised in the conversation include recommendation quality, user trust, messages being filtered into promotions or spam folders, high compute needs, and the need to raise more capital.
• Instinct is private, so the discussion does not provide a direct public-market investment route.

Cognition and Factory (Private companies)

• The hosts discussed a public dispute involving Cognition and Factory, including disagreements related to an executive’s move and claims about conduct. The transcript does not establish the underlying facts.
• Cognition was described as focused on enterprise code generation and business automation. One participant compared it to Infosys; the hosts pushed back on the idea that it was simply trying to imitate Anthropic.
• The conversation suggested the companies may be pursuing different directions, but did not provide financial details or a clear comparison of their products.

Takeaways

• The discussion offers a possible enterprise AI theme—software that automates coding and business work—but little evidence for comparing the companies as investments.
• Both are private companies, and the transcript provides no valuation, revenue, or investment terms to support an actionable investment conclusion.

Alphabet (GOOGL)

• Google’s Gemini 4 Argon model was said to have posted impressive benchmark results.
• The hosts also noted uncertainty and controversy about how much Google employees were using the model internally for coding. They said its practical performance still needed to be assessed.

Takeaways

• The discussion provides a positive but preliminary signal on Gemini’s benchmark performance, not evidence of commercial impact.
• For an investment view, distinguish benchmark results from real-world adoption and business contribution; neither was established in the transcript.