Amazon completed a $25 billion（約4兆円） bond sale on July 7 but faced notably weaker demand than recent hyperscaler offerings, with orders at just 2.5 times the amount available. The sale underscores investor pushback on the scale of AI infrastructure spending; AI-sector bonds this year have already reached nearly double the $136 billion（約22兆円） issued in all of 2025, while Amazon's own free cash flow has plummeted from $25.9 billion（約4.1兆円） a year ago to $1.2 billion（約1900億円）, signaling the financial strain of massive capital expenditures on cloud and AI buildout.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Amazon raised $25 billion（約4兆円） in bonds on July 7, bringing its 2026 debt issuance to $92 billion（約15兆円）—more than Alphabet, Meta, or Oracle have each issued this year. To attract buyers, Amazon offered 18 to 21 basis points of extra yield on its longest bonds, yet demand was the weakest for any hyperscaler since Meta's $30 billion（約4.8兆円） bond sale in October 2025, with orders at just 2.5 times the bonds on offer (down from 3.2 times in March).
Why it matters
The slowing demand signals investor hesitation about the sustainability of the AI buildout. AI-sector bonds issued this year are already almost double the $136 billion（約22兆円） issued for the full year 2025, and hyperscalers like Amazon are burning cash at unprecedented rates—Amazon's free cash flow over the trailing 12 months fell to $1.2 billion（約1900億円） from $25.9 billion（約4.1兆円） a year ago, driven by a $59.3 billion（約9.5兆円） year-over-year increase in property and equipment purchases. For business decision-makers, tighter bond-market appetite may gradually constrain how aggressively tech giants can fund AI infrastructure expansion.
What to watch
Amazon's hyperscaler bond spreads expanded 6 to 15 basis points on the day of the deal. The company will report earnings either later this month or in early August, which may reveal whether the capex pace (Amazon spent $43.2 billion（約6.9兆円） in Q1 2026 on AWS and generative AI) can be sustained profitably.
The $25 billion（約4兆円） Amazon bond sale exposes a tension at the heart of the AI investment boom. Hyperscalers have collectively issued $270 billion（約43兆円） in debt this year, with an estimated $194 billion（約31兆円） from tech giants building and operating massive data centers. Amazon alone spent $43.2 billion（約6.9兆円） in the first quarter of 2026 on AWS and generative AI infrastructure, forcing the company to tap the bond market aggressively. Yet the market is showing signs of fatigue. Bank of America characterized the demand as a "surprise" and noted that it was the weakest performance for any hyperscaler since October 2025, suggesting investors are becoming more selective about which AI-driven debt they will absorb.
The underlying concern is straightforward: spending is moving faster than the revenue path to justify it. Amazon's free cash flow has collapsed—from $25.9 billion（約4.1兆円） a year ago to just $1.2 billion（約1900億円）—even as operating cash flow climbed 30% to $148.5 billion（約24兆円）. The math reveals the squeeze: enormous near-term capital outlays (land, power, buildings, chips, servers, networking) must be paid in cash before AWS can monetize the capacity six months to two years later. Amazon's CFO acknowledged this explicitly on the April earnings call. For investors, the widening bond spreads (6 to 15 basis points) and weakening order books signal not a collapse in confidence, but a gradual repricing of risk. BofA described the pushback as not "too significant" at this point, and demand remains robust overall. Still, tighter credit conditions may force a recalibration of how quickly hyperscalers can expand, which in turn could reshape timelines for AI adoption across enterprise and cloud markets.
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
