Amazon AWS profits rise amid AI capex surge

The pressure is clearest in the valuation debate around Anthropic, whose rich private-market price tag has become part of a broader argument over whether AI leaders are being valued ahead of profits. The seed headline captures the central tension: to support a $2 trillion-style valuation, a company needs something resembling the profit profile of a mature platform business, not a venture-stage balance sheet. That is why comparisons to Amazon matter. Amazon’s cloud and advertising businesses have shown that scale can eventually convert into earnings, but Anthropic has “barely turned a profit,” leaving investors to bridge a wide credibility gap between adoption and monetization.

The market backdrop adds to the stakes. U.S. Treasury yields remain elevated, with the 10-year around 4.7% and the 2-year near 4.2%, keeping a higher discount rate on long-duration growth assets. That makes profitable AI franchises more valuable than speculative ones, and it raises the bar for unprofitable names that depend on future cash flows. Even the broader S&P 500 tone remains only neutral in Adalytica’s S&P 500 trade signals, suggesting the market is not in a euphoric phase where any AI narrative can be rewarded indiscriminately.

Amazon’s stock has reflected that tension. Its shares have been volatile, but the latest price action shows investors are still willing to pay for AI-linked growth when it is backed by operating income, not just narrative. Microsoft and Alphabet, meanwhile, remain the other heavyweight comparables in the AI race: both have healthier balance sheets and deeper commercial distribution than most private AI players, which makes them more plausible long-term beneficiaries if enterprise AI spending keeps rising. Their own shares have also been bid up, but their earnings bases are far more established.

For investors, the distinction is crucial. The bull case is that Anthropic and its backers are building a foundational AI layer that will eventually monetize through enterprise subscriptions, model access fees and cloud usage, much as AWS and Microsoft Cloud monetized earlier waves of digital transformation. The bear case is that the economics of frontier AI remain too capital-intensive, with inference, training and infrastructure costs absorbing much of the value created, leaving limited room for the kind of profits implied by trillion-dollar valuations.

That is why Amazon’s latest numbers matter beyond one stock. They show that the AI boom is still being financed by incumbents with real cash flow, even as private-market valuations for AI companies stretch far ahead of proven earnings. The next catalyst is whether AI demand can keep lifting cloud revenue and operating leverage fast enough to validate those bets — or whether the market starts demanding the kind of profit discipline that Amazon itself only earned after years of scale.