- Amazon raised 2026 cash capex guidance to $220B from $200B, with the $20B increase driven by higher memory chip costs [1] - AWS revenue grew 37% YoY to $42.2B in Q2, its fastest growth in 18 quarters, with operating margins at 39.4% [3] - AWS contracted backlog hit $496B, up $132B in a single quarter, with demand already reserved through 2028 [2] - Amazon's AI and custom silicon businesses each surpassed $25B annualized run rates, growing at triple-digit percentages [3] - CEO Andy Jassy said Amazon will not have enough capacity to meet demand in 2026, and expects the same dynamic in 2027
[4] Amazon raised its full-year 2026 capital expenditure forecast to approximately $220 billion, up $20 billion from prior guidance of $200 billion, citing elevated memory chip costs as the primary driver. The revision came during the company's Q2 2026 earnings call on July 30, alongside results that showed AWS revenue surging 37% year-over-year to $42.2 billion — its fastest growth rate in 18 quarters [1] [3].
Even at the new spending level, CEO Andy Jassy told analysts the company will still fall short of meeting customer demand. "We will still not have enough capacity to meet all the demand we have in 2026. And I believe this dynamic will also be true in 2027, too," Jassy said on the call [4]. AWS contracted backlog — work that has been signed but not yet delivered — jumped $132 billion in a single quarter to reach $496 billion, with commitments already stretching into 2028
.
[2]Amazon's Q2 results broadly beat expectations, with total revenue hitting $200.6 billion (up 20% YoY) and operating income rising 43% to $27.5 billion. The company guided Q3 revenue of $197 billion to $202 billion and operating income of $22.5 billion to $26.5 billion. Shares surged approximately 9% in after-hours trading [3] [5].
The Numbers: $220B and Rising Memory Costs #
The $20 billion capex increase from $200 billion to $220 billion was attributed specifically to higher memory costs rather than an expansion of Amazon's buildout footprint. Q2 capital expenditures alone totaled $54.2 billion, a 68% year-over-year increase, reflecting the pace at which Amazon is deploying infrastructure [3].
AWS property and equipment reached $223 billion in Q1, up from $190 billion the prior quarter. Jassy noted that AWS remains on pace to double its power capacity by the end of 2027 compared with 2025, a metric that underscores the physical scale of the buildout [4].
The capex increase positions Amazon as the largest single spender among hyperscalers in 2026. Alphabet has guided $175 billion to $185 billion, Meta $125 billion to $145 billion, and Microsoft $110 billion to $120 billion — bringing the Big Four total to roughly $725 billion, up 77% from 2025 [6].
AWS Performance and AI Revenue #
AWS delivered $42.2 billion in Q2 revenue with a 39.4% operating margin, generating $16.6 billion in segment operating income. The margin included approximately $600 million in energy derivative gains. At its current pace, AWS is running at a $169 billion annualized revenue rate [3].
Amazon's AI business and its custom silicon business — categories that overlap — each independently crossed a $25 billion annualized run rate, growing at triple-digit percentages year-over-year. The custom silicon figure is significant given Anthropic's 10-year, $100 billion Trainium capacity commitment signed in April 2026, which alone would account for substantial future demand [3] [2].
Backlog and Demand Visibility #
The $496 billion AWS backlog represents one of the most concrete demand signals in the infrastructure sector. The $132 billion quarterly increase reflects both new enterprise AI commitments and renewals of existing cloud contracts at higher volumes [2].
Cumulative multi-year commitments from major AI labs now exceed $225 billion, anchored by the Anthropic deal and similar arrangements. The demand visibility stretching to 2028 gives Amazon unusual forward clarity on infrastructure requirements but also constrains near-term capacity allocation [3].
For the broader data center ecosystem — including power providers, cooling vendors, and electrical equipment suppliers — Amazon's disclosure confirms that supply-side constraints on AI infrastructure remain the binding factor through at least 2027, not demand uncertainty [1].
Hyperscaler Capex Context #
Amazon's $220 billion commitment arrives in a quarter where every major hyperscaler raised or maintained aggressive capex targets. The combined $725 billion in 2026 planned spending across Amazon, Alphabet, Meta, and Microsoft represents the largest single-year infrastructure investment cycle in technology history [6].
Investor reactions have been mixed across the group. While Amazon shares rose sharply on the strength of AWS growth and margin expansion, Alphabet's earnings call earlier in the week triggered a sell-off despite similar capex escalation. The market appears to be differentiating based on revenue conversion rates rather than penalizing capex spending uniformly [5] [6].
The sustained capacity constraints reported by Amazon and its peers have direct implications for power procurement, data center construction timelines, and GPU supply chains. Amazon's statement that it will double power capacity by end of 2027 implies gigawatts of new generation and transmission coming online across its global footprint over the next 18 months [4].
What's Next #
Amazon did not break out specific data center locations or power procurement details on the earnings call. However, the company's ongoing buildouts span Northern Virginia, Oregon, Ohio, Mississippi, Saudi Arabia, and multiple other jurisdictions where permit filings and utility interconnection requests have been documented throughout 2025 and 2026.
The key question for infrastructure watchers is whether the $220 billion figure represents a ceiling or a floor. Memory costs could continue to rise, and Amazon's own language — demand extending to 2028, capacity insufficient through 2027 — suggests the spending trajectory has further room to increase in subsequent guidance updates.
Companies mentioned #
Further sources #
[1] Data Center Knowledge — Amazon Lifts AI Infrastructure Spending to $220B as Dem… ↗
[2] TechTimes — AWS Backlog Hits $496B as Amazon Raises AI Spend to $220B and Capac… ↗
[[3] Digital Applied — Amazon's $200B Quarter: Ads, AWS, and the Anthropic Gain ↗](https://www.digitalapplied.com/blog/amazon-q2-2026-earnings-ads-aws-anthropic-gain)
[[4] Fortune — Andy Jassy said Amazon will spend $220 billion this year — and still … ↗](https://fortune.com/2026/07/30/andy-jassy-amazon-capex-demand-aws-pga-tour/)
[[5] CNBC — Amazon hikes 2026 capex to $220 billion due to higher memory costs ↗](https://www.cnbc.com/2026/07/30/amazon-amzn-q2-earnings-report-2026.html)
[[6] Tom's Hardware — Google, Microsoft, Meta, and Amazon capex spending to hit $725… ↗](https://www.tomshardware.com/tech-industry/big-tech/big-techs-ai-spending-plans-reach-725-billion)
The stories that matter, in one email. Free — unsubscribe anytime.