Amazon CEO Andy Jassy delivered second-quarter 2026 results on July 30, 2026, reporting net sales up 20% year-over-year to $200.6 billion and earnings per share of $5.75. Alongside those numbers, Jassy raised Amazon’s 2026 capital expenditure guidance from $200 billion to $220 billion, an increase he attributed largely to surging memory chip prices — the same supply-chain pressure Apple’s Tim Cook flagged on his own final earnings call the same week. Despite committing to spend $220 billion on infrastructure this year, up sharply from the $131 billion Amazon spent in 2025, Jassy told analysts the company still won’t “have enough capacity to meet all the demand we have in 2026” — a direct acknowledgment that AWS’s own growth is currently being capped by physical infrastructure limits, not customer demand.
To help fund that spending, Amazon sold $25 billion worth of corporate bonds in early July 2026, a substantial debt raise specifically earmarked for continuing the AI infrastructure build-out. When pressed by analysts on how the company would ultimately finance the full scope of its AI ambitions, Jassy offered little additional detail beyond confirming the spending commitment itself, saying there was “nothing to share” on further financing specifics at this time.
A New, Fast-Growing Line: Amazon’s Own AI Chips
One of the more concrete growth signals Jassy pointed to was Amazon’s custom AI chip business, which he said now has an annual revenue run rate exceeding $25 billion, growing at triple-digit percentage rates year-over-year. That business — built around Amazon’s own Trainium and Inferentia chip families, designed as lower-cost alternatives to Nvidia’s GPUs for many AI workloads — is a direct bet that Amazon can capture more of the AI infrastructure value chain itself, rather than depending entirely on outside chip suppliers the way most of its cloud competitors do.
The Trillion-Dollar Bet
Perhaps the most striking single claim from Jassy this quarter was his statement that he now believes AWS will “very possibly” become a $1 trillion annual revenue business “in time” — a specific, if unspecific on timeline, prediction from the head of a cloud division that generated a fraction of that figure even in its most recent strong quarter. Jassy pointed to a growing roster of large, well-known enterprise customers moving substantial workloads to AWS — citing an example like the PGA Tour’s own use of AWS cloud infrastructure — as evidence of the kind of broad, cross-industry adoption he believes can eventually support that scale of revenue.
How Wall Street Reacted
Investor reaction to Amazon’s quarter was mixed once the scale of the new spending commitment set in. Amazon shares moved on the earnings news as analysts weighed strong underlying sales growth and Jassy’s bullish AWS commentary against the reality of a further $20 billion jump in planned capital expenditure funded partly through fresh debt — a now-familiar pattern across nearly every major cloud provider this earnings season, with Alphabet, Meta, and Microsoft all disclosing similarly large upward revisions to their own AI infrastructure budgets within the same few weeks. That synchronized spending pattern across all four major hyperscalers is itself a notable data point: whatever supply constraints and rising memory costs each company cited individually appear to reflect the same industry-wide conditions, not a company-specific issue unique to Amazon. Amazon’s own retail and advertising businesses, the company’s original core, continued growing at a healthy clip alongside AWS during the same quarter — a reminder that Amazon’s AI infrastructure bet is being funded by, and layered on top of, an already large and profitable underlying business, rather than replacing it.
What This Means for Philippine Founders
Jassy’s admission that AWS still can’t meet current demand despite a $220 billion spending commitment is a genuinely useful, real-world data point for any Philippine startup planning its own cloud infrastructure needs for the next 12-18 months: capacity constraints at this scale, at the world’s largest cloud provider, suggest Philippine businesses should build in real buffer time when provisioning new AWS compute for a planned launch or scale-up, rather than assuming instant availability. Amazon’s own AI chip push is also worth watching directly — a credible, lower-cost alternative to Nvidia GPUs reaching real commercial scale is a genuine opportunity for Philippine AI startups to lower their own infrastructure costs, once Trainium and Inferentia-based instances are broadly available and well-supported for the specific AI workloads a Philippine team is actually running. Jassy’s $1 trillion AWS prediction, if it materializes even partially, would also mean a correspondingly larger, more diversified set of enterprise reference customers and case studies for Philippine businesses evaluating AWS against its competitors — a genuinely useful body of evidence to draw on when making an infrastructure decision, beyond marketing claims alone. Whether that prediction plays out on Jassy’s own optimistic timeline or a slower one, the direction is the more durable signal for planning purposes: cloud infrastructure spending industry-wide is still scaling up, not leveling off.
Share this article