While everyone else rallied on AI hype, Amazon spent 2024 and 2025 building

Buying servers. Building data centers. Spending money like crazy.

But here's what Wall Street is starting to notice: all that spending is about to pay off.

Multiple big-name analysts are pointing to 2026 as Amazon's year. Not because of some vague promise. Because of three specific things happening at once.

Let me walk you through what they're seeing.

The Cloud Business 

Amazon Web Services (AWS), the cloud computing arm, is the real profit engine here.

Think about it this way: AWS brings in only 17% of Amazon's total sales. But it generates nearly 60% of the actual profit. 

Morgan Stanley says AWS could grow over 20% in 2026. That's a big jump from where it's been.

Why the acceleration? Two reasons.

First, companies are finally scaling up their AI projects. Not just testing. Actually running them. 

Amazon's partner Anthropic (the company behind Claude) is one example. These AI workloads need massive computing power. And they're willing to pay for it.

Second, Amazon has been capacity-constrained. They couldn't build data centers fast enough to meet demand. They had customers ready to buy, but nowhere to put them.

That's changing. All those billions spent on infrastructure are starting to come online. More chips. More server racks. More power capacity. 

When that happens, the revenue that's been sitting in the backlog gets released.

The $125 Billion Bet on AI

Amazon is spending $125 billion on infrastructure in 2025. 

Most of that is going into what they call "Project Rainier." The goal? 

Build data centers the size of small towns. We're talking gigawatt-scale facilities designed specifically for AI workloads.

Now, $125 billion sounds scary. And it is a lot of money. Some investors worry Amazon is burning cash for nothing.

But Oppenheimer broke down the math. Every gigawatt of new capacity generates about $3 billion in annual revenue. That's not a bad return if you can pull it off.

The question is execution. Can Amazon actually build this infrastructure fast enough to beat Microsoft and Google? That's the risk.

Trainium Chips

Here's something most investors miss about Amazon's AI strategy: they're building their own chips.

The chips are called Trainium. And they matter because right now, everyone running AI needs Nvidia's GPUs. Those chips are expensive. And often sold out.

Amazon decided to stop waiting in line.

Trainium chips are designed specifically for training AI models. They're not as powerful as Nvidia's top chips yet. But they're good enough for many workloads. 

Amazon can offer them to AWS customers at a much lower price.

Why does this matter for 2026? Two reasons.

First, profit margins. When Amazon uses Trainium instead of buying Nvidia chips, they keep more of each dollar customers spend on AWS. Better margins mean higher profits.

Second, competitive advantage. If you're a startup training an AI model, and Amazon can do it for 40% less than using Nvidia chips on Microsoft Azure, where are you going to go?

Amazon's not trying to beat Nvidia at their own game. They're just trying to give customers a cheaper option that's good enough. In cloud computing, "good enough and cheaper" wins a lot of business.

The Trainium rollout is ramping up in 2025. By 2026, it should be widely available across AWS data centers. That's when the margin improvement starts showing up in earnings.

Advertising: The Other Profit Machine

While everyone watches the cloud business, advertising is quietly becoming huge for Amazon.

It grew 19% in Q1 2025. And the margins on ads are much higher than selling you a toaster.

Amazon knows what you actually buy. Not just what you click on. What you put in your cart and purchase. That data is gold for advertisers.

They're also putting ads in Prime Video now. Think about how many people watch Thursday Night Football on Amazon. Every ad shown there is high-margin revenue.

Wedbush says this shift toward advertising is a key reason they're bullish on 2026. Amazon is going after Meta and Google's turf. And they have the customer data to compete.

What's the Stock Actually Worth?

So all this sounds great. But what does it mean for $AMZN?

Morgan Stanley set a "bull case" target of $350 per share. 

KeyBanc came in at $300

Morningstar says the fair value is $240, and the stock is currently undervalued.

The average across these firms points toward significant upside from current levels.

Why are they confident? As AWS margins improve and advertising revenue grows, Amazon's profit picture changes dramatically. The market tends to pay more for profitable growth than just raw sales.

Amazon is also getting more efficient with delivery. They call it "regionalization." Basically, they're making sure your package doesn't travel across the country when it could ship from a warehouse 50 miles away. 

Small thing. Big savings.

What Could Go Wrong?

This isn't a sure thing. Three big risks stand out.

First, antitrust. The FTC is taking Amazon to trial in October 2026. The case focuses on Prime subscriptions and whether Amazon is abusing its market power. Even if Amazon wins, the headlines could tank the stock for months.

Second, competition. Microsoft $MSFT ( ▼ 2.24% ) and Google $GOOGL ( ▲ 1.43% ) aren't sitting still. If Google's AI tools end up being better than Amazon's, AWS could lose customers fast. The cloud wars are brutal.

Third, returns on investment. If Amazon spends $125 billion and it doesn't generate the expected cash flow, free cash flow suffers. Investors would punish the stock hard.

The Bottom Line

Amazon appears to be shifting gears. From spending mode to harvesting mode.

AWS is about to accelerate. Advertising is printing money. The infrastructure investments are going live.

Is it guaranteed? No. But the setup is there.

For investors watching from the sidelines, 2026 could be the year the market finally rewards Amazon for all that patient capital investment. 

Either way, it's worth paying attention to what happens next.

Personally, I am holding Amazon and buying the dips.

My average price now is $222…and I think this stock is still somewhat dormant, but 2026 might be the year it gains momentum.

Disclaimer: This analysis is for educational purposes only and should not be considered investment advice. Always do your own research before making investment decisions.

Trader Insights Media tracks thousands of companies every week using rigorous financial analysis.

Here's what smart investors are watching right now:

Keep Reading