AI demand rises as the cost of supplying it divides the market
Cloud and server demand is accelerating, but heavier capital spending and margin pressure are not producing a one-way AI rally across company shares.
Economics & Markets··Morning
Spending rose, valuations strained
After signaling heavier AI spending, Tesla fell 14.5% and Alphabet 7.1%, erasing roughly 200 billion dollars and 300 billion dollars in value. Both reported negative second-quarter free cash flow. Alphabet raised annual capex guidance to 195-205 billion dollars, while Tesla's spending rose 142% to 5.79 billion dollars. Yet Google Cloud revenue grew 82% and customers spent about 50% above commitments. Third-party capacity intended to serve that demand could depress margins.[1], [3]
Demand creates different winners in the supply chain
Intel's second-quarter revenue rose 25% to 16.1 billion dollars and data-center revenue 59% to 6.3 billion dollars; management said customers wanted more than it could produce. Google Cloud's search for outside capacity and Intel's production constraint show the same AI-infrastructure pressure at different points. Cloud capacity and server processors involve different products, contracts and timelines, so their figures cannot be combined. The shared conclusion is that demand is forcing capacity decisions in services and chips.[2], [3]
The theme grows without a single return path
J.P. Morgan's guide placed AI ETFs among the top five themes by assets despite second-quarter volatility, as overall mutual-fund inflows slowed and money shifted toward ETFs. Those flows do not guarantee future returns. Intel's growth and cloud customers' above-commitment spending show strong usage, while Tesla's and Alphabet's declines expose capital intensity and margin pressure. Markets are pricing the AI opportunity separately from the burden of financing it. The observable signal is whether added capacity becomes revenue without further margin erosion.[1], [2], [3], [4]
Related columns
For more information on this topic, you can read the related columns.