Investors should consider Micron Technology (MU) as a primary "memory trade" play, given its massive $38 billion profit projection and critical role in supplying memory for autonomous vehicles and Anthropic. Broadcom (AVGO) remains a high-conviction infrastructure pick as it partners with OpenAI to develop custom "Jalapeno" chips and continues to manufacture Google’s TPUs. For exposure to the intersection of AI and energy, monitor the advanced nuclear sector and companies like Valar Atomics, which address the massive power demands of frontier AI labs. Alphabet (GOOGL) offers a unique media angle through its $75 million A24 partnership, validating its generative AI tools for professional Hollywood production. Meta Platforms (META) is a key bet on consumer AI hardware, specifically through its sleek smart glasses designed to replace the phone as the primary AI interface.

OpenAI (Private)

• OpenAI has announced the development of its own custom AI chip, internally nicknamed "Jalapeno."
• The chip is an ASIC (Application-Specific Integrated Circuit) designed specifically for AI inference (supplying tokens quickly) rather than general-purpose computing like a standard GPU.
• AI-Driven Design: The hardware and software stack were reportedly designed by OpenAI’s own models (Codex and ChatGPT), accelerating the design-to-prototype timeline to just nine months (compared to the industry standard of 1-2 years).
• Vertical Integration: By owning the hardware, OpenAI aims to eliminate memory bandwidth bottlenecks found in NVIDIA chips, potentially reaching 100% utilization and significantly lowering the cost of running models.
• Strategic Partnerships: OpenAI is collaborating with Broadcom (AVGO) and MediaTek to bring these chips to market by the end of the year.

Takeaways

• Efficiency Gains: Investors should watch how this vertical integration affects OpenAI's burn rate. If successful, it reduces their reliance on expensive third-party hardware.
• Competitive Threat: This move signals a shift where AI labs become hardware competitors to established players like NVIDIA, following the path of Google (TPUs) and Amazon (Trainium).

Micron Technology (MU)

• Micron recently reported stellar earnings, posting $41 billion in quarterly revenue, which the analysts noted was roughly 15% more than NVIDIA made in the same quarter a year prior.
• The company projected $55 billion in revenue for the next quarter with $38 billion in profit.
• Market Position: Micron is the #1 American supplier of memory chips and #3 globally.
• The "Memory Trade": The discussion highlighted that AI models require massive amounts of memory for inference and data storage. As autonomous vehicles (Teslas require 5-8x more memory than standard cars) and robotics scale, memory demand is expected to grow exponentially.
• Anthropic Partnership: Micron has signed a deal to be the primary memory provider for Anthropic and has also invested in Anthropic’s Series H funding round.

Takeaways

• Bullish Sentiment: The analysts argue the "memory trade" has not topped; rather, demand is just beginning due to physical AI, robo-taxis, and humanoid robots.
• Diversification: Micron offers exposure to the AI theme beyond just processing power, acting as a critical infrastructure provider for any entity building AI hardware.

Broadcom (AVGO)

• Broadcom is highlighted as a key partner for OpenAI’s new chip venture.
• The company already manufactures Google’s TPUs (Tensor Processing Units).
• Broadcom’s CEO has claimed their current AI chips are on par with NVIDIA’s Blackwell architecture.

Takeaways

• Infrastructure Play: Broadcom remains a central player in the custom silicon (ASIC) market, making it a primary beneficiary as more tech giants (OpenAI, Google) move away from off-the-shelf GPUs.

Meta Platforms (META)

• Model Development: Meta is reportedly working on a "Mythos-class" model, though analysts expressed skepticism regarding its 9-month timeline, suggesting it may already be behind Chinese open-source labs.
• Hardware: Meta recently revealed new smart glasses (partnering with Kylie Jenner for the AI voice).
• Product Strategy: Unlike high-end AR headsets, these glasses focus on a sleek, "normal" look with microphones, speakers, and cameras, acting as an intermediary AI interface.

Takeaways

• Consumer AI: Meta is betting on wearable hardware as the primary way users will interact with AI daily, rather than just through a phone app.
• Open vs. Closed: There is a noted tension in Meta's strategy as they shift between open-source and closed-source models to stay competitive.

Alphabet / Google (GOOGL)

• A24 Partnership: Google DeepMind has entered a $75 million deal with film studio A24.
• Integration: A24 will use Google’s video (VO), music, and image (Nano) generation models to aid storytelling.
• IP Protections: The deal specifically prohibits Google from training its models on A24’s copyrighted content, focusing instead on tool creation for creators.

Takeaways

• Media Transformation: This partnership validates Google’s generative AI tools in a professional Hollywood setting, potentially setting the standard for how AI is used in high-end film production.

Valar Atomics (Private)

• The company recently achieved a milestone by running a nuclear reactor for 24 hours straight.
• Energy Theme: The analysts are highly bullish on the intersection of AI and energy, noting that frontier AI labs will require massive amounts of power that companies like Valar Atomics aim to provide.

Takeaways

• Energy Infrastructure: For investors looking at the "long game" of AI, the energy sector (specifically modular or advanced nuclear) is identified as a critical bottleneck and opportunity.

Anthropic (Private)

• Mentioned as a major player in the "coding model" space.
• Claude Tag: A new "multiplayer" feature for Slack that allows teams to tag the AI into workflows like a human colleague.
• Internal Efficiency: Anthropic reported that 65% of their own internal code is now generated by their AI tools.

Takeaways

• Workforce Automation: The shift from "chatbot" to "autonomous agent" is accelerating. Anthropic’s integration into existing work tools (Slack) represents a move toward AI becoming a functional member of a corporate workforce.