Amazon is spending like it’s 2010 again—because the next internet is being built right now

What people miss about Amazon is that it’s never been “an e-commerce company.” It’s a systems company that happens to sell you toothpaste, stream movies, and run a meaningful chunk of the internet.

Today, that systems obsession is showing up in two very different headlines: Amazon is making it cheaper to sell low-priced goods on its India marketplace, and Amazon Web Services (AWS) just dealt with a real-world reminder that cloud infrastructure is, in fact, infrastructure.

The fun part (for investors) is how those stories connect.

Amazon’s 2026 vibe: build, subsidize, repeat

On February 5, 2026, Amazon.com, Inc. (AMZN) reported results for the quarter ended December 31, 2025—and the numbers were a good snapshot of what the company is choosing to be right now.

Amazon said fourth-quarter 2025 net sales rose 14% year over year to $213.4 billion, while AWS grew faster: AWS sales were $35.6 billion, up 24% year over year. For full-year 2025, Amazon reported net sales of $716.9 billion (up 12%), and AWS sales of $128.7 billion (up 20%). Amazon also set expectations for a huge 2026 investment cycle, saying it expects about $200 billion in capital expenditures in 2026, largely tied to AI infrastructure.

That capex line matters because it frames almost everything else Amazon does: the company will happily spend aggressively when it thinks it can lock in a long runway.

Now look at India.

On March 1, 2026, Amazon disclosed it will stop charging seller referral fees in India for products priced under 1,000 rupees (about $11). Reports around the change indicate it starts in mid-March 2026. This is Amazon doing what it does best: making the marketplace easier to use, more predictable, and more compelling for sellers—especially in the messy, high-volume part of e-commerce where a few percentage points can decide whether a small merchant bothers listing at all.

This isn’t charity. It’s a land-grab for selection, price competitiveness, and seller loyalty in one of the most contested e-commerce arenas on earth.

AWS isn’t “the other business.” It’s the engine

Amazon’s consumer story gets most of the pop culture oxygen, but AWS is still the part of Amazon that can turn heavy spending into high-margin, repeatable revenue.

That’s why the second March 1, 2026 headline hit differently: AWS said power to a UAE data center was temporarily shut after objects struck the facility, causing sparks and a fire. The incident is notable not just because outages are annoying, but because it underlines a truth people forget when they talk about “the cloud” like it’s a vibe: cloud computing is physical—data centers, power, cooling, supply chains, geopolitics, the whole thing.

In 2026, every AI pitch eventually bottoms out in the same question: who has enough compute, close enough to the data, with enough reliability, to run the workloads that matter? Amazon is answering that question with spending.

Why this matters to everyday investors

Amazon’s story is less about “Will people keep buying stuff online?” and more about whether it can keep compounding advantages across commerce, logistics, advertising, and cloud.

The company is essentially running two flywheels at once:

• In retail marketplaces (like India), Amazon can pull levers—fees, shipping, selection—to attract sellers and keep customers.
• In AWS, Amazon can pour capital into chips and data centers so it can host the next generation of AI-heavy software.

If you own broad index funds like SPY, IVV, or VOO, you already own a meaningful slice of this storyline. The more interesting question is whether Amazon’s 2026 spending spree becomes one of those rare “it looked expensive, but it aged well” moments—or whether competition and operational complexity make it harder to turn all that investment into durable returns.

Amazon is betting it can do the hard, unglamorous work (data centers, logistics, fees) better than everyone else, and then let the internet come to it. That’s a very Amazon bet.