Investors should monitor eBay (EBAY) as a contrarian value play, focusing on potential cost-cutting of its inefficient $2.4 billion marketing budget to immediately boost cash flow. A strategic merger with GameStop (GME) could create a dominant "physical verification" moat for high-value collectibles, utilizing GME’s 1,600 retail locations to eliminate fraud in the used goods market. GameStop (GME) remains a unique asset play, currently valued primarily for its $9 billion cash pile, which management may use as currency for large-scale acquisitions. In the AI sector, prioritize "picks and shovels" infrastructure by investing in companies with secured access to NVIDIA H100/B200 chips, as compute remains the primary bottleneck for development. Look toward the "hard sciences" frontier by backing firms like Periodic Labs that use AI and robotics to accelerate the discovery of new materials and superconductors.

eBay (EBAY)

Anjney Midha presents a contrarian bull case for eBay, specifically focusing on its potential merger with GameStop (GME). He argues that the market is misinterpreting the strategic value of the deal.

• Operational "Fat": Midha points out that eBay spent $2.4 billion on marketing in fiscal 2025 to acquire only 1 million new users. He views this as a massive inefficiency ($2,400 per new user) that a new management team could cut to immediately boost cash flow.
• The "Physical Verification" Moat: The core thesis is that Amazon cannot easily compete in the "used and collectibles" market (e.g., rare trading cards, vintage pens) because high-value unique items cannot be processed through standard automated warehouses.
• GameStop Synergy: Midha suggests that GameStop’s 1,600 retail locations could serve as physical verification hubs for eBay transactions. This would solve the "fraud" problem in high-end collectibles.
• Agentic Commerce: In an era of AI agents (like Claude), Midha believes physical verification is the missing link. AI can find rare items, but it cannot physically verify them; a retail footprint provides the "stamp of authenticity" required for AI-driven commerce.

Takeaways

• Efficiency Play: Look for potential value extraction if management pivots from aggressive (and inefficient) marketing to cost-cutting.
• Niche Dominance: eBay remains a leader in "non-standard" goods where Amazon’s logistics model fails.
• Strategic M&A: Watch for developments regarding Ryan Cohen and GameStop; the "bull case" relies on leveraging physical stores to backstop digital marketplace trust.

GameStop (GME)

The discussion centers on GameStop's evolution from a struggling retailer into a holding company with significant cash reserves.

• Cash Position: GameStop is sitting on approximately $9 billion in cash, while the market is valuing the actual business operations (brand and retail) at only about $1 billion.
• The "Meme" Premium: Midha suggests that Ryan Cohen may be looking to use GameStop’s stock as "currency" to acquire larger targets like eBay, though the market has not yet provided the "pop" in share price needed to make a 50/50 cash-stock deal seamless.

Takeaways

• Asset Play: GameStop is currently valued largely for its cash pile rather than its retail performance.
• Transformation Risk: The investment outcome depends entirely on Ryan Cohen’s ability to execute a large-scale acquisition and integrate physical retail into a modern e-commerce/AI framework.

AI Infrastructure & Compute (AMP PBC)

Midha discusses his new fund, AMP, which operates as a Public Benefit Corporation (PBC) focused on the "bottleneck" of the AI era: compute power.

• Compute as a Utility: Midha compares the current AI era to the industrial revolution of 1885. Just as coal was the scarce input then, compute (GPUs) is the scarce input now.
• Utilization Inefficiency: He notes that even major players like Elon Musk have massive compute clusters (H100s) running at low utilization (e.g., 11% MFU).
• Business Model: AMP buys compute in bulk and provides it "at cost" to portfolio companies (like Anthropic) to unblock research, while taking equity in those companies.
• The "Bitter Lesson": A belief that scaling compute and data is the primary driver of AI progress, rather than just clever algorithms.

Takeaways

• Infrastructure over Applications: While many VCs are "paralyzed" by the fast-moving application layer, the "picks and shovels" (compute and energy) remain the most stable investment theme.
• Strategic Bottlenecks: Investors should look for companies that have secured long-term access to high-end chips (NVIDIA H100s/B200s), as this is the primary constraint on AI development.

Frontier Science & Materials (Periodic Labs)

Midha highlights Periodic Labs as a prime example of how AI is being applied to physical sciences.

• AI-Driven Discovery: Using AI to predict new materials, such as high-temperature superconductors, and using robots to synthesize and verify them in a closed loop.
• Verification Speed: The firm has achieved more material verifications in 90 days than the field typically sees in a decade by using AI to automate the scientific method.

Takeaways

• Beyond Chatbots: The next major investment frontier is AI applied to "hard sciences"—materials science, biology, and energy.
• Execution as the Moat: In these fields, the "moat" isn't just the software, but the physical laboratory infrastructure and the speed of the verification loop.

Investment Themes: "The Great Renaissance"

Midha identifies several macro themes for the next 5–10 years:

• Agentic Commerce: The shift from humans browsing websites to AI agents (like Claude or GPT) buying goods on behalf of users. This requires new infrastructure for trust and verification.
• Trust as a Moat: In a world where AI can generate content and code instantly, Midha argues that Trust, Community, and Culture are the only enduring moats.
• Not a Bubble: Despite high valuations, Midha argues we are not in a bubble because the "diffusion" of AI has barely reached the general global economy. Most of the world is still only using basic tools like ChatGPT, leaving massive room for growth.