ANALYSIS The same AI infrastructure buildout that is rewarding companies pouring capital into cloud and custom silicon is now punishing those caught on the wrong side of the component squeeze — and a single earnings night made the divergence unmistakable.
Why it matters
Amazon and Apple reported earnings on the same Thursday evening, and the market's verdict was historic in its asymmetry. Amazon surged 14% on Friday while Apple dropped 9%4. Apple is on track for its biggest single-day decline since March 2020, while Amazon is on track for its biggest jump since April 20123. ◆ The split is not a generic growth-versus-value story; both companies beat revenue expectations. The fault line runs through AI capital expenditure: who is spending into the cycle and who is being squeezed by it.
The big picture
Amazon reported revenue of $200 billion, up 20% year over year, beating the LSEG consensus estimate of $196.47 billion6,8. AWS sales expanded 37% year over year, its fastest growth since 2021, trouncing analysts' expectations for 31% growth. Operating income rose 43% year over year to $27.46 billion, beating the $23.57 billion consensus forecast. CEO Andy Jassy said the company expects capital expenditures to reach $220 billion this year. Amazon's in-house chip business — where it rents compute capacity based on its own custom silicon — has reached a revenue run rate of $25 billion.
Apple, by contrast, reported earnings, revenue, and iPhone sales all above market expectations, including a 22% surge in handset sales1. But the company issued guidance for September-quarter revenue growth of 9% to 11%, missing analysts' expectations for 12% growth, citing "supply constraints". "The major takeaway from Apple's report was how much the increase in memory prices is really starting to weigh on the company," one analyst said on Bloomberg Technology. "They talked a lot about component shortages".
ANALYSIS The irony is structural: the very AI spending boom that is filling Amazon's cloud order book is tightening the memory and chip supply that Apple needs for its devices.
Meanwhile, the infrastructure buildout continues to accelerate beyond the hyperscalers themselves. Morgan Stanley is leading advanced talks to lend $15 billion to Nexus Data Centers to build a large AI infrastructure campus in Hubbard, Texas, for Anthropic7. Google agreed to backstop Anthropic with its investment-grade credit rating as part of the deal.
Between the lines
The market's reaction rippled across the global semiconductor supply chain. SK Hynix closed nearly 30% higher, marking its best day on record, while Samsung closed nearly 27% higher, also clocking its highest single-day gains5. Taiwan's TSMC surged almost 10%. Japanese chip stocks rallied sharply: Advantest climbed more than 16%, Disco rose 12.74%, and SoftBank Group jumped 13.8%. The iShares Semiconductor ETF (SOXX) surged more than 8% overnight.
ANALYSIS The chip rally marks a sharp reversal from the same week's bruising sell-off over lofty AI valuations. Amazon's results — and specifically the AWS acceleration — appear to have provided the proof point that spending translates to revenue, resetting the narrative in a single session.
Mizuho's analyst called Amazon's AI story "a game changer," arguing that the 37% AWS top-line acceleration "convinced more people that AI is improving their overall business". On Amazon's custom silicon advantage, the analyst noted: "Not only does it enable them to get supply, because it's so difficult to get Nvidia graphics processors, it also enables them to offer that to their customers, probably at a lower cost. So you win on both sides, supply and price or cost".
ANALYSIS That framing highlights the strategic wedge opening between companies that control their own silicon supply and those that depend on the same constrained merchant market. Amazon's $25 billion custom-chip run rate is both a margin advantage and a supply hedge; Apple's guidance miss suggests it lacks an equivalent buffer for the memory and component surge driven by AI demand.
The Anthropic data center deal underscores how the capex cycle is now pulling in new categories of capital. A $15 billion project-finance loan backstopped by Google's credit rating represents infrastructure spending at a scale and structure more familiar from energy megaprojects than from software companies.
What's next
Apple guided for 9% to 11% revenue growth in the September quarter. ANALYSIS That quarter coincides with the company's next iPhone launch cycle, meaning the supply constraints Apple flagged will be tested against peak seasonal demand. Amazon's raised capex forecast of $220 billion for the year sets a new benchmark that other hyperscalers will be measured against. The Nexus-Anthropic financing, if finalized, would signal that AI lab infrastructure is becoming a standalone asset class — with investment-grade credit from strategic backers replacing traditional venture equity as the funding mechanism for frontier compute.