Google is explicitly prioritizing the development of Artificial General Intelligence (AGI) over the commercial deployment of its own AI processors. CEO Sundar Pichai (photo) made this clear during the presentation of Alphabet’s quarterly financial results. Those figures included another striking detail: for the first time since 2004, the company reported negative free cash flow.
Although Google now also offers its Tensor Processing Units (TPUs) to cloud customers, the CEO stated that developing its own AI models remains the top priority. During the earnings call, Pichai was asked about how Google allocates its available TPU capacity. Analysts wanted to know how the company balances growing customer demand against its own need for computing power.
According to Pichai, that trade-off is clear. The capacity needed to remain at the forefront of Artificial General Intelligence (AGI) development is allocated first. Only then does Google distribute the remaining computing power across products such as Search, YouTube, and Google Cloud.
Within Google Cloud, the CEO says the available capacity goes primarily to Vertex AI, Gemini Enterprise, and other strategic services, including data analytics and cybersecurity solutions.
AI drives strong growth for Google Cloud
The financial results show why demand for AI capacity is so high. Google Cloud generated revenue of $24.75 billion in the past quarter, an 82% increase from a year earlier. Operating profit came in at $8.8 billion, more than three times last year’s figure.
According to Alphabet, this growth is primarily driven by demand for AI infrastructure and AI solutions. The company now also has a backlog of $514 billion in cloud contracts yet to be fulfilled.
The other business segments also performed strongly. Search’s advertising revenue grew by 17 percent, and YouTube saw its advertising revenue rise by 13 percent.
Notably, according to Google, generative AI is actually driving more traffic to the search engine. While it was previously expected that AI chatbots would replace traditional search queries, Pichai says that AI Mode within Search is actually increasing the total number of search queries.
Hiring temporary additional capacity
To keep up with the demand for AI computing power, Alphabet is once again increasing its investments. CFO Anat Ashkenazi expects the company to invest between $195 billion and $205 billion in infrastructure this fiscal year, higher than the previous forecast of $180 billion to $190 billion.
Nevertheless, Google is also grappling with the limited availability of AI hardware. Because not all necessary systems can be delivered on time, the company will temporarily lease additional capacity from third-party providers in the third quarter. This strategy is intended to bridge the gap until its new in-house infrastructure is ready.
According to Pichai, this is a deliberate choice. By accepting temporarily higher costs now, Google can continue to serve major cloud customers and secure multi-year contracts that will yield attractive returns in the long term.
Investments weigh on cash flow
Alphabet closed the second quarter with revenue of $119.8 billion, up 24 percent from a year earlier. Operating income rose 34 percent to $40.8 billion.
However, the substantial investments in AI infrastructure are placing an increasingly heavy strain on the company’s finances. For the first time since 2004, the company reported negative free cash flow of $5.9 billion. Investors reacted cautiously to this development: in after-hours trading, Alphabet’s stock lost about 4 percent.