Jefferies Analyst Brent Thill on CoreWeave Setback and the AI Trade

CoreWeave Setback and AI Trade Dangers

• ⚠️ The delay in CoreWeave's third-party data center is highlighted as an example of the dangers within the AI trade, where multiple components must align for success.
• 🧩 This setback is described as a delay rather than a lost transaction, impacting the company's ability to recoup and rebuild in Q1.
• 📈 Similar challenges are observed across the industry, including at Microsoft and Amazon, indicating that AI infrastructure buildout is not linear but characterized by lumpy deployments.

AI Infrastructure as a Long-Term Play

• 🚀 Despite short-term hiccups, the roadway for AI is still in infrastructure, with the current buildout being in years three to four of a projected 20-year cycle.
• ⏳ Many AI-related stocks are undergoing a "mini timeout" rather than a permanent setback, suggesting a need for a brief pause before potential reaceleration.
• 📊 The market has seen a sell-off in many AI-related names, including Oracle, indicating a broader need for the group to take a break after a significant run-up.

Government's Role in AI Development

• 🤝 The government is expected to be an active partner in AI development, not necessarily a backstop, to ensure US leadership against global competition, particularly China.
• 💡 This partnership aims to secure US leadership in AI, retain jobs, and maintain a competitive edge in a rapidly evolving technological landscape.
• 📈 While the US is currently leading in AI innovation, competition from China is seen as a positive factor that keeps the US on its toes.

AI's Impact on Software and Applications

• 🏢 Board members globally are showing unprecedented enthusiasm for AI adoption, with CEOs like the head of Walmart indicating flat headcount, signaling a need for other companies to embrace AI.
• 📉 A clear dichotomy exists between infrastructure stocks (Snowflake, Microsoft, Oracle) which have performed well, and application names (Adobe, Salesforce) which are currently struggling.
• ⏳ The time for application names to catch up is not immediate, with expectations for their resurgence around 2026-2027, after the infrastructure buildout is more complete and valuations have adjusted.