
Amazon and Microsoft are each spending roughly $200 billion(約32兆円) this year on AI data center buildout in an unprecedented arms race to capture cloud market share and prove the investment will yield returns. Investor confidence is strained after Google's recent stock drop on heavy capex and negative free cash flow, raising questions about whether either company can monetize massive infrastructure spending fast enough. This week's earnings reports from both firms will be closely watched for signs of slowing demand or margin pressure that could suggest the race has become unsustainable.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Amazon and Microsoft are each deploying roughly $200 billion(約32兆円) in capital expenditures in 2026 to build out AI data centers, an unprecedented level of investment driven by demand for cloud and AI services. Microsoft reports earnings Wednesday and Amazon on Thursday, when investors will scrutinize revenue growth, profit margins, and customer backlogs at their respective cloud businesses.
Why it matters
Investor patience is thin as both companies race to prove massive spending will generate returns within the next few years. Last week, Google's stock fell 7% after raising capital-expenditure projections and reporting negative free cash flow, raising questions about whether hyperscalers can convert infrastructure investment into profits fast enough. Amazon and Microsoft together represent 8% to 9% of the S&P 500, so retirement investors have direct exposure to the outcome.
What to watch
AWS has a backlog of $364 billion(約58兆円) in signed customer contracts (plus a recent $100 billion(約16兆円) Anthropic deal excluded from that figure) and in-house chip revenue commitments exceeding $225 billion(約36兆円), while Microsoft's total remaining performance obligations stand at $627 billion(約100兆円) (99% higher year-over-year). Both companies are selling everything they can build, and near-term results will test whether available AI applications—currently limited to chat assistants and coding tools—can justify the scale of spending.
Amazon and Microsoft have dominated cloud computing for nearly two decades, but the artificial intelligence boom has pushed their rivalry to an intensity neither company has faced before. Each is set to spend roughly $200 billion(約32兆円) this year building data centers—Amazon formally guiding to that figure for 2026, and Microsoft spending $104 billion(約17兆円) in the first nine months of fiscal 2026 with expectations to reach roughly $190 billion(約30兆円) for the calendar year, per CFO Amy Hood's April guidance. This week, Microsoft reports earnings Wednesday and Amazon Thursday, with investors closely watching revenue growth, profit margins, and customer backlogs at Azure and AWS, respectively, alongside the capital costs underpinning the buildout.
The stakes are unusually high because investor patience has become unpredictable. Last Thursday, Google's parent Alphabet saw its stock crater 7% after raising capital-expenditure projections and reporting negative free cash flow, signaling that the market is questioning the massive sums being deployed and the timeline for return on investment. For Amazon and Microsoft, the two largest cloud players controlling half the market between them, the competitive dynamic has shifted from a steady rivalry to what equity investors describe as a "bloodsport" over capital allocation. Amazon's free cash flow dropped from $25.9 billion(約4.1兆円) a year ago to $1.2 billion(約1900億円) in the past 12 months, and it more than doubled bond debt to over $120 billion(約19兆円) to fund the spending. Microsoft has avoided new bond issuance, instead funding roughly $35 billion(約5.6兆円) per quarter from operating cash flow, though its free cash flow for the 12 months ending in March was $73 billion(約12兆円), up slightly year-over-year—a relative strength that some analysts credit to the different structures of the two companies' finances.
Customer commitments offer some reassurance to investors about the spending pace. AWS has a backlog of $364 billion(約58兆円) in remaining performance obligations (signed contracts for future revenue), which excludes a recent $100 billion(約16兆円) deal with Anthropic; Bank of America analysts noted that AWS's in-house chip revenue commitments exceed $225 billion(約36兆円). Microsoft disclosed nearly $627 billion(約100兆円) in remaining performance obligations, representing 99% year-over-year growth, though that figure encompasses the entire commercial business, not Azure alone. Amazon CEO Andy Jassy wrote in his annual shareholder letter that "we're not investing approximately $200 billion(約32兆円) in capex in 2026 on a hunch," noting that AWS capex is "almost all spoken for" and that customer commitments cover a substantial portion, with much of the spending expected to be monetized in 2027–2028. AWS Chief AI and Technology Officer Matt Wood told Fortune that "the more capacity we open up, we sell it immediately."
Yet the applications driving those returns remain nascent. The handful of breakout AI products currently available are limited to chat assistants and coding tools—a state Wood compared to the early internet, when only a few websites existed. "We're going to have millions of them, just like we have millions of websites today," he said, and "there are going to be entirely new categories of products that just didn't exist a year ago that are going to be absolute breakthrough smash hits." This week's earnings calls will test whether investors believe that vision, or whether the massive spending will eventually force both companies to slow deployment and justify near-term returns.
The cloud market has become a two-player race where Amazon Web Services and Microsoft Azure operate as what researcher Melissa Otto calls "frenemies"—competitors with distinct strengths and often different customer bases, yet capable of competing for the same enterprises. AWS excels with flexible, customizable infrastructure ideal for machine learning workloads and startups, while Azure extends Microsoft's enterprise software ecosystem, making adoption easier for companies already running Microsoft products. When customers do compete, they often buy from both providers, fragmenting rather than consolidating the market. This structural reality explains why the spending has reached an unsustainable intensity: with no clear winner and early-stage AI still dominated by basic applications like chatbots and coding tools, both companies are locked in a capital arms race to secure infrastructure capacity and lock in customer commitments before the field consolidates. The financial pressure is real—Amazon's free cash flow plummeted from $25.9 billion(約4.1兆円) a year ago to $1.2 billion(約1900億円) in the past 12 months as capex accelerated, while Microsoft has maintained higher free cash flow ($73 billion(約12兆円)) but is channeling roughly $35 billion(約5.6兆円) per quarter into buildout. Last week's market reaction to Google's raised capex guidance and negative free cash flow signaled that investor patience for infrastructure spending without near-term return is eroding, making this week's earnings calls a critical moment for both companies to convince the market that their customer backlogs and commitments justify the unprecedented capital deployment.
AI-summarized, only the topics you pick — one digest a day via Email, Slack, or Discord.
Free · takes 30 seconds · unsubscribe anytime
No comments yet. Be the first to share your thoughts!
Log in to join the discussion





Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime