Feb 20, 2026 · 24-Hour AI Briefing: Amazon Tops the Global Revenue Chart, Server CPUs Enter a Seller’s Market
In the last 24 hours, two headlines clicked together: Amazon has edged past Walmart to become the world’s top revenue company, while Intel and AMD’s 2026 server CPU capacity is reportedly close to sold out—with prices set to rise. Taken together, they point to the same macro shift: in the AI era, advantage is increasingly defined by infrastructure, supply certainty, and monetizable distribution, not just product execution.
1. Amazon surpasses Walmart to rank #1 globally by revenue; Walmart posted $713.2B vs Amazon’s $716.9B
Commentary:
Walmart has held the top spot for more than a decade on the back of an unmatched physical retail footprint, everyday-low-price discipline, and dominance in U.S. groceries. But Amazon’s revenue engine is no longer “retail-only.” Cloud, advertising, and third-party seller services have been compounding fast enough to steadily close the gap—and finally flip the ranking.
The key point is mix, not the ~$3.7B difference. Strip out AWS and Amazon’s 2025 revenue would have been about $588.0B. In other words, AWS is the decisive lever: cloud revenue grew ~20% YoY to roughly $128.7B in 2025, and that “infrastructure flywheel” is what pushed Amazon over the line in total revenue. This wasn’t Amazon simply selling more stuff—it was Amazon scaling higher-leverage services.
AI amplifies the divergence. Walmart’s AI focus is primarily operational: supply chain optimization, merchandising, fulfillment, pricing, and store productivity. Amazon can do all of that and sell AI directly as cloud compute and tooling to the world—then monetize demand through a powerful ads stack. The ranking change is the symptom; the real story is the difference in scalability and profit structure.
Do you think Amazon will stay ahead of Walmart through 2026?
2. Intel and AMD’s 2026 server CPU capacity is reportedly near sold out; both plan 10–15% price increases
Commentary:
This is what a supply-constrained market looks like—now extending beyond GPUs into CPUs. As AI agents move from experiments to large-scale production, demand for general-purpose compute is rising sharply. In many agent workflows, a large share of latency-sensitive work still sits on CPUs: orchestration, data prep, networking stacks, storage paths, virtualization, and security isolation. Meanwhile hyperscalers (AWS, Azure, Alibaba Cloud, etc.) are aggressively rolling out next-gen server platforms, using long-term agreements and early bookings to lock capacity well ahead of time.
A 10–15% hike is meaningful in a high-ASP server market. It lifts system TCO immediately and tends to propagate up the stack—motherboards, memory, networking—by resetting pricing expectations for the entire platform. The impact will also be uneven: hyperscalers usually have stronger bargaining power and can lock supply via long contracts, but even they may face a binary choice in tight periods—pay more or accept longer lead times. Smaller datacenters and enterprise private clouds are more likely to get squeezed on both price and availability.
Who do you think is next to raise prices?
Most important AI events from the past 72 hours: