Alphabet eased many of the concerns about the sustainability of AI infrastructure investment by reaffirming plans to expand investment in artificial intelligence (AI) along with cloud results that beat market expectations.
Hwang Su-uk of Meritz Securities said in a report on the 23rd, "We believe Alphabet's second-quarter results delivered all the numbers the market wanted to see," adding, "It was an earnings release that met expectations for increased capital expenditures (CAPEX) to address AI proliferation and revenue monetization."
Alphabet raised this year's CAPEX guidance to $200 billion from the prior midpoint of $185 billion. It did not provide a specific figure for next year's CAPEX but projected it would increase significantly from this year.
The pace of AI monetization was also faster than the market expected. Google Cloud revenue jumped 81.8% from a year earlier, far exceeding the market consensus (in the 60% range) and even surpassing 80%, which had been seen as the top end of optimism. The cloud operating margin was 35.6%, up 2.7 percentage points from the previous quarter. The order backlog also grew to $517 billion, and it was seen as positive that revenue recognition began for TPU systems first supplied to customer data centers and that these volumes were reflected in the backlog.
Hwang said the results offered three takeaways.
First, AI computing resources are still in a shortage state. He said, "Alphabet chose a bridge strategy of actively using external computing capacity until internal data center expansions are completed," adding, "In the short term, expense burdens will pressure the operating margin, but the company said the return on investment for long-term contracts is high."
Next, he assessed that Alphabet made clear its intent to continue competing in frontier AI models based on the "scaling law." Hwang said, "Much of the Q&A on the earnings call focused on Gemini's competitiveness, and the company said it regards securing frontier models as very important and is putting substantial computing resources into them," interpreting this as "it has not given up on AI competition and will continue to compete to develop top-tier models."
He also pointed to a shift in thinking about funding sources as noteworthy. He said, "The market has tended to view AI CAPEX only through big tech's free cash flow (FCF)," adding, "But Alphabet noted that, in addition to operating cash flow, it is considering liability and, if needed, equity issuance as funding options."
In fact, over the past 12 months, Alphabet expanded the size of its liability from $16 billion to around $100 billion. He said, "Even though free cash flow temporarily turned negative, it could raise CAPEX guidance because it does not rely solely on cash flow for investment funding," adding, "This suggests that AI investment can continue through diverse funding methods."
He noted that AI investment risk is shifting to funding methods. Hwang said, "If 2023–2024 investment was cash-centered, going forward AI CAPEX may move into the realm of optimal capital structure, making it more susceptible to interest rates and the credit market," adding, "The AI investment risk the market will watch may not be whether big tech gives up competing, but rather become a credit issue."