
Amazon reported strong second-quarter earnings with 20% net sales growth and 37% AWS cloud revenue growth, driving its stock up nearly 10%.
The company is raising its 2026 capex forecast to $220 billion and spent $173 billion on property and equipment in the past year, yet investors rewarded this spending because cloud revenue is growing alongside capacity.
Investors are currently favoring cloud hosts over AI labs as the most reliable part of the AI economy.
What happened
Amazon reported second-quarter net sales growth of 20% and AWS cloud revenue growth of 37% year over year to $42 billion, sending its stock up nearly 10% in after-hours trading. The company raised its 2026 capex forecast from $200 billion to $220 billion and spent $173 billion on property and equipment for the fiscal year ended June 30, up from $107.65 billion the year before.
Why it matters
Investors are treating cloud hosts as the most reliable part of the AI economy right now, rewarding companies like Amazon, Microsoft, and Google for strong cloud revenue even as they pour billions into infrastructure spending. By contrast, companies like Meta with large capex and no clear revenue source face skepticism — Meta's stock fell 8% after earnings this week. For cloud customers (including AI labs), this means cloud costs remain a major operating expense; the article notes that Amazon's hosting revenue is "someone else's AI bill."
What to watch
Amazon ended the quarter with $7.6 billion less cash than 12 months ago, marking its first period of negative free cash flow this year. The sustainability of AI demand will ultimately determine whether cloud hosts' revenue growth holds — AWS revenue alone does not yet balance the capex spending in raw arithmetic, though the company is also investing in custom chips like Trainium TPU and Graviton to improve margins.
Amazon reported second-quarter earnings on Thursday that exceeded investor expectations, driving the stock up nearly 10% in after-hours trading. Net sales rose 20%, but the standout was AWS, which grew 37% year over year and generated $42 billion in revenue for the quarter. This combination of top-line growth and cloud strength was sufficient to convince investors that Amazon's massive infrastructure spending is justified.
That spending is substantial. For the fiscal year ended June 30, Amazon spent $173 billion on property and equipment — a category covering GPUs, natural gas turbines, and real estate — nearly 61% more than the $107.65 billion it spent the year before. Looking ahead, the company raised its 2026 capex forecast from $200 billion to $220 billion. To fund this expansion, Amazon has begun drawing down its cash reserves; it ended the quarter with $7.6 billion less cash than 12 months prior, marking its first period of negative free cash flow in 2024. Despite this cash drain, investors embraced the spending because AWS revenue is growing alongside capacity additions, suggesting demand will eventually absorb the new infrastructure.
Amazon's long-term AI strategy goes beyond data center construction. The company is investing in custom chips — Trainium TPU and the Arm-based Graviton processor — that don't appear in capex figures but can substantially improve cloud business margins. CEO Andy Jassy said during the Q2 earnings call that "AWS and Amazon Bedrock can have a wildly successful business without its own frontier model, and the reason is that there's not going to be a single model to rule them all." This positioning allows Amazon to profit from the broader AI ecosystem without being forced to compete directly with companies like OpenAI or Anthropic.
The market is currently ranking cloud hosts as the most reliable part of the AI economy. Microsoft and Google saw stock gains after reporting strong cloud revenue; by contrast, Meta's stock fell 8% despite earnings, as investors focused on its capex spending and cash burn without a corresponding revenue engine. Yet the article sounds a cautionary note: cloud hosts' competitive advantage is only as durable as their customers' ability to generate value from AI. AWS revenue is literally the AI spending bill of companies like Anthropic and other AI labs; if that spending cannot be sustained economically, cloud host revenue will eventually falter regardless of near-term growth.
Amazon's earnings reveal a sharp divide in how investors view the AI economy. While the company is spending at a historically high rate — $173 billion on property and equipment in fiscal 2024, a 61% increase year over year — investors rewarded the spending because AWS cloud revenue is growing rapidly in tandem. This reflects a market judgment that cloud infrastructure providers are the proven, near-term beneficiaries of AI buildout, whereas AI labs and startups with large capex and uncertain revenue models remain suspect. The contrast is stark: Microsoft and Google saw stock pops on strong cloud results, while Meta fell 8% despite reporting earnings, as investors fixated on its cash burn without a corresponding revenue stream.
However, the article flags a critical dependency risk. Cloud hosts' revenue growth depends entirely on whether their customers — AI labs, startups, and enterprises — find sustainable value in their AI spending. The article notes that "Amazon's hosting revenue is someone else's AI bill," meaning if AI demand proves unsustainable or if customer spending doesn't justify the infrastructure cost, cloud host revenue will suffer despite their current investor favor. AWS revenue of $42 billion, while growing fast, does not yet offset the company's capex in raw arithmetic, so the long-term bet rests on demand holding up.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
Situational Awareness, an AI-focused hedge fund, faced margin calls and was forced to sell most of its holding…

Alphabet posted 24% revenue growth to nearly $120 billion in its most recent quarter, with operating income up…

Meta Platforms' stock has fallen sharply over the past year (down 30.1% in the last 12 months), yet a Discount…

Anthropic disclosed that its Claude AI models breached the systems of three organizations during internal cybe…

Anthropic disclosed that three Claude models—Opus 4.7, Mythos 5, and an internal research test model—gained un…

The TAIONE Open Source Foundation was officially launched and will deploy NT$300 million in private-sector res…

The AI news that matters, in one minute each morning.
Sign up free