The Price of the AI Race: Google Burns $490 Million a Day
Google's quarterly reports confirm a dominant industry trend: massive AI infrastructure spending has led to the company's first negative cash flow in over two decades.
The quarterly reports published by Google last week are further proof of the trend currently dominating the technology industry: extremely extensive spending on AI, which is impacting financial results. Google reported a negative cash flow in the second quarter for the first time since going public more than two decades ago, against the backdrop of massive AI infrastructure expenses.
According to the reports, Alphabet — Google's parent company — spent 44.9 billion dollars between April and June on AI infrastructure, including chips, servers, and server farms. This translates to an expenditure of 490 million dollars per day. In contrast, the company earned 39.1 billion dollars during the same period.
Free cash flow is a metric that particularly interests investors, as it measures the cash remaining for companies to pay off debts or return to shareholders after covering operating costs and capital expenditures. The company's CFO, Anat Ashkenazi, emphasized to analysts that this trend is expected to continue, with capital expenditures projected to reach 195–205 billion dollars this year.
"The estimate is that free cash flow will continue to be under pressure as a result of our investments in technological infrastructure, which allow us to leverage the opportunities that AI creates and continue to generate attractive returns," she said. Ashkenazi noted that the company relies on operating cash flows, debt, and shares to finance these expenses, emphasizing that Google has no intention of selling additional shares beyond what has already been announced.
However, in view of increasing AI expenses, Alphabet's debt has climbed to nearly 100 billion dollars. In June, the company raised 85 billion dollars in its first share sale in over 20 years. These expenses are intended to meet a backlog of cloud contracts, which rose to 514 billion dollars at the end of June, up from 460 billion dollars in the previous quarter.
Google is not alone. Meta, Microsoft, and Amazon are also increasing spending on data centers, network equipment, accelerators, and the massive power infrastructure required to support these activities. Together, the four companies are expected to spend 730 billion dollars on AI this year, with spending projected to increase further next year.
Data revealed by Google aligns with a report from the Japanese newspaper "Nikkei Asia," which estimates that the extensive investment by five major US tech giants (including Oracle) in AI infrastructure has led to a "hidden debt" of 1.65 trillion dollars. This refers to long-term contracts that have been signed but have not yet entered into force, as the data centers are not yet operational. Once these sites go live, the companies will be obligated to pay for the generated computing power regardless of market demand. Consequently, Alphabet, Amazon, and Microsoft face the risk of high future payments for capacity that may lack sufficient customers.
Beyond infrastructure, tech companies are spending record amounts on lobbying in Washington to shape regulations regarding data centers and AI oversight. According to the "Financial Times," Google spent 4.7 million dollars on lobbying in the second quarter, while Microsoft spent 3.5 million dollars. Google's focus is on shaping policy around data center construction and electricity consumption. The company stated it is working to promote "federal legislation that promotes American dominance in AI, ensures responsible development, and benefits everyone."
