When Amazon reports fourth-quarter earnings after Wednesday's market close, the numbers will tell only part of the story. Yes, analysts expect revenue of $211 billion and earnings per share of $1.98. But the metrics that matter most won't appear in the headline figures—they'll emerge from management's commentary on artificial intelligence investments that now define the company's strategic direction.

Amazon has committed to spending more than $100 billion on AI infrastructure over the next several years, joining Microsoft and Alphabet in a capital deployment race that has captivated and, at times, concerned investors. Tomorrow's earnings call will provide critical visibility into whether those investments are generating returns.

AWS: The Main Event

Amazon Web Services remains the company's profit engine and the primary vehicle for monetizing AI capabilities. Analysts project AWS revenue of approximately $35 billion for the quarter, representing growth north of 21% year-over-year—a meaningful acceleration from recent trends.

But the number investors really want is the backlog. Amazon disclosed a record $200 billion remaining performance obligation (RPO) for AWS at the end of Q3, representing contracted revenue not yet recognized. Growth in this metric—or any deceleration—will heavily influence how the market receives the report.

"AWS backlog growth is the single most important number in Amazon's earnings. It tells you where revenue is going over the next several years, not just the quarter being reported."

— Cloud Infrastructure Analyst, Goldman Sachs

The AI Capital Expenditure Question

If there's one issue keeping Amazon investors up at night, it's capital spending. The company has guided to capex approaching $100 billion for the current fiscal year—a figure that seemed astronomical until Microsoft and Alphabet announced similarly aggressive plans.

The software stock selloff that has dominated markets this week has intensified scrutiny of Big Tech's AI investments. Bulls argue the spending is necessary to capture a generational opportunity; bears worry about returns that remain unproven at scale.

Management's commentary on 2026 capital plans will be parsed for any nuance. Are spending levels tracking as expected? Is AI demand justifying the investment? Are there signs of overcapacity building in the industry? These questions will shape the post-earnings narrative.

The Bedrock Bet

Amazon's AI strategy differs from competitors in one crucial respect: the company has positioned itself as a platform for multiple AI models rather than betting exclusively on proprietary technology. Through Bedrock, AWS customers can access models from Anthropic, Meta, Stability AI, and others alongside Amazon's own Titan offerings.

This "Switzerland" approach has resonated with enterprise customers wary of locking into a single AI provider. But it also means Amazon's success is partially dependent on third parties. Any commentary on Bedrock adoption rates and the competitive positioning versus OpenAI-powered Azure will be closely watched.

Retail: The Other Business

While AWS dominates investor attention, Amazon's retail operation still generates the majority of revenue. E-commerce sales are expected to show mid-single-digit growth, with advertising revenue—now a $60 billion annual business—providing meaningful margin expansion.

The retail segment faces several headwinds:

  • Consumer Spending Concerns: January's consumer confidence collapse raises questions about discretionary spending ahead
  • Prime Pricing: J.P. Morgan analysts predict a price increase to $159 in 2026, which could affect subscriber growth
  • Tariff Uncertainty: A significant portion of Amazon's third-party marketplace products originate from China

Any guidance suggesting retail weakness could offset AWS strength in the overall stock reaction.

The Trainium Factor

Amazon has made a major push into custom AI chips through its Trainium and Inferentia processor families. These chips offer cost advantages over Nvidia GPUs for certain AI workloads, potentially reducing Amazon's own infrastructure costs while creating a differentiated offering for AWS customers.

The Trainium 3 generation is expected to begin sampling later this year, with production ramp in 2027. Any updates on the chip roadmap or customer adoption of existing Trainium offerings will provide insight into whether Amazon can reduce its dependence on Nvidia's dominant GPUs.

What the Street Expects

Consensus estimates for Q4 2025:

  • Revenue: $211.2 billion (up 12.5% year-over-year)
  • Earnings Per Share: $1.98 (up 6.5% year-over-year)
  • AWS Revenue: $35.0 billion (up 21.6% year-over-year)
  • Operating Margin: 10.2%

Amazon has a strong track record of beating estimates, but the magnitude of the beat matters. Given elevated expectations around AI, merely meeting consensus could be received as a disappointment.

The Bigger Picture

Tomorrow's Amazon earnings arrive at a pivotal moment for AI-focused tech investing. The sector has experienced meaningful volatility as investors debate whether AI's transformative potential justifies current valuations and investment levels.

Amazon's results—and more importantly, management's confidence about AI demand sustainability—will help determine whether the rotation out of Big Tech continues or reverses. For the broader market, the implications extend well beyond one company's quarterly numbers.