
Tech • IA • Crypto
A new AI-focused investment firm is betting that compute infrastructure, cost discipline, and physical verification systems will reshape both technology markets and legacy platforms like eBay.
Recent financial disclosures show eBay spent $2.4 billion on marketing while adding only 1 million net new users, growing from 134 million to 135 million. That implies roughly $2,400 per new user, raising questions about whether much of that spending is effectively reacquiring existing users rather than driving real growth. The figures suggest significant operational inefficiency in a platform already widely known to consumers.
The proposed interest in combining GameStop and eBay is less about product innovation and more about cost restructuring. The core thesis argues that up to $2 billion in expenses—particularly marketing—could be reduced and reallocated. Under this view, savings alone could offset financing costs, making the deal viable even before operational improvements.
GameStop’s 1,600 retail locations are positioned as a potential advantage in e-commerce, particularly for physical verification of goods. This capability could address one of the biggest challenges in resale markets: fraud. By verifying items in-store, the combined entity could create a trust layer that purely digital platforms struggle to replicate.
The collectibles and used goods segment is described as structurally resistant to Amazon’s logistics model. Unlike standardized products, items such as rare cards or vintage goods cannot be easily processed through automated warehouses. This creates a niche where marketplaces like eBay retain a competitive edge.
The rise of AI-driven shopping agents is expected to amplify the importance of authenticity verification. While AI can search and compare listings efficiently, it cannot independently confirm whether a physical item is genuine. A system combining digital discovery with trusted physical verification could become essential infrastructure for “agentic commerce.”
During the NFT boom, platforms like Discord handled over $10 billion in transaction volume without needing physical verification, thanks to blockchain-based ownership. Attempts to extend that model to physical goods failed due to the lack of verification systems, reinforcing the need for hybrid digital-physical marketplaces.
A newly launched firm, AMP (Public Benefit Corporation), is targeting AI’s biggest bottleneck: compute capacity. The company aggregates and leases compute resources, aiming to improve utilization rates that are currently far below optimal levels. Some large clusters reportedly operate at around 11% efficiency, representing billions in underused infrastructure.
Within eight weeks, AMP secured over $1.3 billion in commitments for its first fund and plans to raise several billion more. The firm combines venture investing with infrastructure provisioning, supplying compute to portfolio companies at cost to accelerate development.
AMP frames AI development as analogous to industrial-era energy markets, where fragmented and inefficient resource use slowed progress. By pooling compute into a shared “grid,” the firm aims to increase efficiency and enable smaller teams to compete with large tech incumbents.
The firm is backing projects such as AI-driven materials discovery, including efforts to develop high-temperature superconductors. Using automated labs that combine AI prediction, robotic synthesis, and rapid verification, these teams have reportedly achieved more experimental results in 90 days than in the prior decade in some areas.
AMP is structured as a public benefit corporation, allowing it to prioritize ecosystem-wide gains over short-term shareholder returns. The model is intended to address positive externalities in venture capital and infrastructure, where private markets often underinvest without coordinated incentives.
Despite rapid advances, global adoption of AI tools remains limited. Large portions of the world are still unfamiliar with or not actively using systems like ChatGPT, suggesting that current developments represent an early phase rather than a speculative peak.
The convergence of AI infrastructure, cost restructuring, and hybrid digital-physical commerce models is reshaping both emerging startups and legacy platforms, with compute access and trust systems emerging as critical competitive factors.