Amazon heads into earnings with the cloud trade on edge. Shares fell roughly 4% into Thursday’s print as investors fixate on Amazon Web Services after Microsoft’s historic market-cap wipeout last week put a price on slowing cloud momentum and heavy AI spend. AWS grew 19% last quarter to $28.79 billion, narrowly missing expectations and reigniting debate over whether AI-fueled demand can offset capacity bottlenecks and margin pressure.
Microsoft’s $500 billion rout forced a collective rethink on the cost of AI. Azure’s growth slowdown and a 66% year-over-year surge in capital spending to $37.5 billion underscored a basic tension: runaway infrastructure budgets now, payoff later. Microsoft shares have stabilized, but the message landed. If the top end of hyperscale is seeing growth constrained by chips, power and construction timelines, the multiple on cloud cash flows gets marked down. That’s the setup confronting Amazon into its own results.
AWS posted 19% revenue growth to $28.79 billion last quarter, just shy of the $28.87 billion analysts modeled. A miss measured in basis points matters right now because the market is searching for a clean reacceleration narrative after two years of optimization headwinds. The Street needs proof that AI workloads are translating into broad-based, durable growth, not just pockets of pilot projects or one-off training bursts that make quarterly comps “lumpy.” Amazon doesn’t need blowout numbers; it needs a glide path that shows steady demand, improving utilization, and leverage despite the capex grind.
AI is forcing hyperscalers to behave like utilities. Data center builds are colliding with real-world limits: GPU availability, transformer lead times, and electricity. The result is a capital cycle running ahead of revenue recognition. Depreciation hits now; monetization ramps later. Investors just punished Microsoft for the timing mismatch. Amazon isn’t immune. It is pouring tens of billions into capacity, custom silicon, and power deals to support generative AI and classic cloud growth. The debate is not whether to spend, but how quickly those dollars become high-margin services rather than low-return infrastructure. If AWS margins hold or expand while capex rises, the stock gets air cover. If margins compress, the market will assume AI is dilutive for longer.
Clarity beats charisma. First, AWS growth cadence: is revenue growth improving sequentially and is the pipeline broadening beyond a handful of frontier-model customers? Second, supply and power: how much capacity is live, what’s the timeline to bring on meaningful new megawatts, and are chip constraints easing? Third, monetization: concrete examples of Bedrock, Q, and custom silicon wins that show customers shifting production workloads, not experiments. Fourth, backlog and commitments: growth in remaining performance obligations will help bridge the capex-to-revenue gap. Fifth, margin trajectory: color on utilization, pricing discipline, and unit economics for inference at scale. The market will also parse any hints on 2026 capex levels and whether spending intensity has peaked.
Microsoft’s message was not demand destruction; it was capacity and cost. That is an important distinction for Amazon. If Azure’s growth moderation stemmed from supply and power bottlenecks, AWS likely faces similar constraints. That makes guidance nuance critical. A credible path to easing constraints in the second half of the year can keep growth expectations intact without inflating near-term costs. On competition, Microsoft has leaned into AI co-pilots tied to its software estate. Amazon’s angle is platform flexibility: model choice via Bedrock, price-performance via Trainium and Inferentia, and a services-led approach through its massive partner ecosystem. The Street will reward whichever hyperscaler shows it can convert AI curiosity into sticky, scaled workloads without buying revenue with discounts.
AWS drives the majority of Amazon’s operating profit and, by extension, a disproportionate share of the valuation argument. After last week’s cloud reset, investors are wiring a higher risk premium into AI infrastructure. That pushes the focus squarely onto margins and free cash flow. For Amazon, the bull case is straightforward: prove that AI demand lifts utilization fast enough to outrun depreciation, keep pricing rational, and let operating leverage reassert. The bear case: AI workloads remain sporadic, inference margins trail classic cloud services, and capex stays elevated longer than planned because power and supply constraints take time to solve. Tonight’s commentary will tilt the balance.
AWS growth and margins will dominate, but the call’s “tells” matter. Any shift in tone on customer optimization versus expansion signals where we are in the cycle. Mentions of large language model inference at scale, enterprise data governance wins, and industry-specific AI services imply durable revenue streams. Concrete timelines for new regions, substations, or power purchase agreements would help quantify capacity relief. And pay attention to how management frames custom silicon adoption; successful migrations lower unit costs and can bolster margins even if list pricing holds steady.
A win is not a moonshot quarter; it’s a clean setup. Modest upside to AWS revenue, steady-to-better operating margin, a growing backlog, and credible capacity timelines would argue that Microsoft’s rout was more about timing than terminal value. A wobble would be soft AWS revenue with a cautious guide, margin pressure attributed to AI mix, and vague capex visibility. In that case, the market will assume the spending curve steepens before payback. With shares already down into the print, expectations are lower, but not low enough to forgive cloudy signals on AI monetization.
Microsoft’s stumble yanked the market’s focus from AI narrative to AI economics. Amazon now has to show its version of the same story can add up faster. If AWS can demonstrate that AI is a profit engine, not just a capex sink, the sector’s multiple can stabilize. If not, the cloud trade remains in reset mode, with power, chips, and depreciation dictating the tempo more than demos and keynotes. Tonight’s read on AWS will set the tone for how investors price AI infrastructure across the market in the months ahead.
