Accumulate shares of Amazon (AMZN) on short-term pullbacks ahead of earnings in seven weeks to capture upside from its cloud expansion and multi-billion-dollar investments in OpenAI and Anthropic.

Establish a position in Unity Software (U) as a key beneficiary of automated game creation following its direct integration into OpenAI's GPT-6 Astra.

Buy dips in Micron Technology (MU) to profit from surging memory consumption driven by broader artificial intelligence model deployment.

Invest in the AI power infrastructure theme by buying Vistra Corp (VST) for steady grid exposure, while waiting for price pullbacks before adding to Bloom Energy (BE) after its recent rally to $280.

Consider SoftBank Group (SFTBY) and SK Telecom (SKM) as high-upside public proxy plays to gain equity exposure to OpenAI and Anthropic ahead of anticipated future IPOs.

Amazon (AMZN)

• The hosts significantly increased their position in Amazon, viewing it as one of the most reliable and direct beneficiaries of the current AI scaling cycle.
• Amazon offers three distinct avenues to capture upside from frontier AI development:
  • Equity Value: Amazon has an initial $15 billion invested in OpenAI with commitments up to $35 billion, alongside an estimated 8% to 9% stake in Anthropic, positioning it for substantial cash/valuation windfalls if these companies IPO at projected $2 trillion to $2.5 trillion valuations.
  • Cloud Infrastructure: OpenAI expanded its AWS agreement by $100 billion over eight years, which includes 2 gigawatts of Trainium compute capacity.
  • Distribution: AWS operates as the exclusive third-party cloud distributor for OpenAI Frontier models while collaborating on custom chips with Qualcomm and runtime software for Amazon Bedrock.
• Internal AI efficiencies across Amazon’s logistics, digital advertising, and retail business lines are expected to expand operating margins significantly over the coming quarters.
• Short-term headwinds—such as oil prices climbing above $100 per barrel and news surrounding a third-party cargo plane accident—are viewed as temporary distractions from long-term fundamentals heading into earnings in roughly seven weeks.

Takeaways

• Accumulate shares of AMZN on short-term macro weakness to capture multi-layered upside from cloud infrastructure spend, equity stakes in top AI startups, and AI-driven internal margin expansion.

Unity Software (U)

• Initiated a new moderate position in Unity following the launch of OpenAI's GPT-6 Astra.
• Astra’s agentic capabilities use the Unity Editor directly to modify scenes, generate assets, and write code, turning the engine into critical infrastructure for AI-assisted game development.
• OpenAI used Unity as a primary case study for Astra alongside Playco, demonstrating 50% fewer manual fixes compared to prior model generations.
• At an $18 billion market capitalization, the company represents a relatively low-cost acquisition candidate ($30 billion to $35 billion) for frontier AI labs looking for defensive moats in interactive environments.

Takeaways

• Establish a position in U as a "picks and shovels" software play benefiting from the democratization and acceleration of AI-generated games and 3D environments.

SoftBank Group (SFTBY)

• SoftBank is under evaluation as one of the purest publicly traded proxies for OpenAI, holding an estimated 13% equity stake.
• With a total market capitalization of roughly $230 billion and major assets including Arm Holdings (ARM), SoftBank offers asymmetric upside if OpenAI lists at a multi-trillion-dollar valuation.
• Investor sentiment has historically discounted SoftBank due to debt and general skepticism toward OpenAI leadership, creating a potential valuation arbitrage opportunity.

Takeaways

• Research and monitor SFTBY as a high-leverage vehicle to gain exposure to OpenAI’s equity upside ahead of a potential public listing.

Bloom Energy (BE)

• Bloom Energy has surged from $170 to $280 (a ~70% gain) following momentum in data center energy demand and its recent S&P inclusion.
• The power constraint theme remains one of the cleanest and least disputed bottlenecks in AI infrastructure, as data centers require massive on-site generation.

Takeaways

• Maintain exposure to the AI power theme, but look to add to BE specifically on pullbacks rather than chasing extended rallies.

Vistra Corp (VST)

• Added as a conservative, pure-play investment on grid electricity generation and industrial power demand.
• Expected data center expansions are projected to significantly increase regional electricity consumption (potentially doubling demand in markets like Texas).
• Trades at a reasonable price-to-earnings multiple, offering a lower-volatility way to play long-term AI power needs without taking speculative technology risk.

Takeaways

• Use VST as a stable, long-term holding to capture rising baseline power demand driven by hyperscale AI compute clusters.

SK Telecom (SKM)

• Highlighted as an under-the-radar public equity proxy for Anthropic.
• Approximately 20% of SK Telecom's enterprise value is estimated to be tied to its Anthropic equity holdings.

Takeaways

• Consider SKM as an alternative international equity play to gain discounted exposure to Anthropic ahead of its anticipated IPO.

Take-Two Interactive (TTWO)

• Long position held on the thesis that generative AI will lower development costs and accelerate content expansion for massive franchises like Grand Theft Auto 6 (GTA 6).
• While generative AI allows individuals to create simple games, top-tier publishers hold a strong moat around premium intellectual property, large multiplayer networks, and live-service operations.

Takeaways

• Hold premium video game publishers like TTWO, as AI toolsets will reduce margin pressures on multi-year development cycles while preserving the value of major intellectual property.

Micron Technology (MU) & AI Memory Infrastructure

• Bought the dip in Micron, noting that memory markets are responding positively to new model announcements.
• The concept of Jevons Paradox is playing out: as AI models become more compute- and memory-efficient per task, the total cost drops, resulting in 100x more aggregate usage and driving higher overall demand for memory and hardware.

Takeaways

• Buy dips on core hardware and memory providers like MU, as increased model efficiency accelerates real-world deployment and overall hardware consumption.