AI spending faces a new test: cash returns

Investors are looking beyond data centre and chip spending to whether it is generating durable revenue and free cash flow, with Alphabet under pressure and Microsoft faring better.

Wall Street is beginning to judge artificial intelligence spending less by its scale than by the cash it produces. That shift became clearer after recent earnings from the largest US technology groups, when Alphabet’s shares fell sharply and Oracle rallied, while Microsoft held up better than many investors had expected. Reuters reported that the market is now scrutinising whether heavy outlays on data centres and chips are translating into durable revenue and healthy free cash flow, rather than simply signalling ambition.

The change reflects the sheer size of the capital commitments required to build AI infrastructure. Big tech companies are still racing to secure servers, semiconductors and power capacity, but the sums involved are now large enough in some cases to outstrip operating cash generation. That has made free cash flow, the cash remaining after capital expenditure, a more useful measure for investors trying to separate sustainable growth from debt-fuelled expansion. Reuters noted that this is especially relevant as some companies increasingly tap bond markets to help finance their build-outs.

Alphabet has become the clearest example of the new market scepticism. According to its filing with the US Securities and Exchange Commission, third-quarter revenue reached $127.6 billion, up 25% year on year, while Search and other revenue increased 12% and Google Cloud revenue more than doubled. Yet the market focused on the pressure behind that growth: capital expenditure outpaced operating cash flow in the quarter, leaving free cash flow negative and forcing investors to confront how much of the AI spending boom is being absorbed before it can be monetised. Reuters reported that the company’s shares fell after results that missed expectations on search growth, even as cloud demand remained strong.

Microsoft, by contrast, has so far convinced investors that it can spend heavily and still preserve financial discipline. Its official first-quarter fiscal 2027 results showed revenue of $92.3 billion, up 19%, and operating income of $45.9 billion, up 21%. The company also reported free cash flow of $18.6 billion, comfortably ahead of market expectations, even after $45.2 billion of investment in computing infrastructure. Azure revenue rose 48%, underscoring how AI-related demand is already feeding through into commercial cloud usage. Melissa Otto of S&P Global Visible Alpha Research told Reuters that the contrast between the two companies was stark, with Microsoft showing disciplined growth while Alphabet consumed cash and missed expectations in its core business.

The pressure is not confined to firms with weaker cash generation. Microsoft is also reported to be considering a $25 billion bond issue, and Oracle has sought to calm investors by setting a cap on additional external financing and saying it expects free cash flow to turn positive in the second half of fiscal 2028. The broader message from the latest results is that AI spending is no longer enough on its own to win favour in equity markets. Investors now want proof that the outlay can be converted into earnings, while still leaving room for future investment without overextending the balance sheet.

This reassessment matters well beyond Silicon Valley. Reuters reported that the Bank of Korea has warned that if major US technology companies slow AI investment next year, South Korea’s semiconductor exports could come in $50 billion to $100 billion below earlier forecasts, trimming economic growth by about 0.4 percentage point. That risk helps explain why investors are watching US big tech so closely: the pace of AI infrastructure spending now shapes not only company valuations, but also demand for memory chips and the prospects for supplier economies that depend heavily on the semiconductor cycle.