Alphabet raises investment in AI infrastructure, sparking market concerns
On July 22, local time, Alphabet, the parent company of Google, announced its second-quarter financial report, significantly raising its full-year capital expenditure target to a range of US$195 billion to US$205 billion, causing investors to pay attention to the scale of investment in artificial intelligence (AI) infrastructure and future cash flow pressure. Although core indicators such as revenue and cloud business exceeded market expectations, the company's stock price fell by more than 4% in after-hours trading. Analysts said the market is evaluating whether continued increased investment in AI can translate into future revenue growth.
On July 22, local time, Alphabet, the parent company of Google, announced its second-quarter financial report, significantly raising its full-year capital expenditure target to a range of US$195 billion to US$205 billion, causing investors to pay attention to the scale of investment in artificial intelligence (AI) infrastructure and future cash flow pressure. Although core indicators such as revenue and cloud business exceeded market expectations, the company's stock price fell by more than 4% in after-hours trading. Analysts said the market is evaluating whether continued increased investment in AI can translate into future revenue growth. Investment of over US$200 billion attracts attention Alphabet's second-quarter financial report showed that the company's quarterly revenue increased 24% year-on-year to $119.8 billion, exceeding analyst expectations. Among them, Google Cloud's performance was particularly outstanding, with quarterly revenue increasing 82% year-on-year to US$24.8 billion; the company also recorded US$98 billion in other income, mainly from the increase in the value of the stocks of other companies it holds. The overall strong performance failed to attract the attention of investors, and the focus of the outside world is still focused on the capital investment pressure brought by AI infrastructure construction. In its financial report, Alphabet raised its full-year capital expenditure forecast to US$195 billion to US$205 billion from the previous range of US$180 billion to US$190 billion. This adjustment caused the company's stock price to fall in after-hours trading, and the market began to re-evaluate the scale of Alphabet's investment in data centers, computing resources and artificial intelligence model development in the next few years. Brian Mulberry, portfolio manager at Zacks Investment Management, pointed to $200 billion as a critical node. Alphabet currently holds a large amount of cash to invest in building AI infrastructure, but the company has not yet fully explained how these investments will translate into revenue growth. Yusuf Squali, chief internet analyst at Truist Securities, said he had expected Alphabet to maintain its full-year capital expenditure guidance range of $180 billion to $190 billion, but the company ultimately raised its target. If calculated based on US$205 billion, Alphabet's capital investment scale in the next few years may further expand. Squally previously predicted that the upper limit of the company's capital expenditures in 2027 may be around the range of 250 billion to 255 billion U.S. dollars; but now it is expected that this number may rise to 270 billion to 280 billion U.S. dollars, or even higher. "The market will take some time to digest this news." Alphabet has continued to expand its AI infrastructure in recent years, with the company’s capital expenditures reaching US$44.9 billion in the second quarter. To support data center construction, Alphabet also previously announced an $85 billion equity financing plan. At the same time, the company's cloud business continues to grow rapidly. At the end of the second quarter, Google Cloud's order backlog reached US$514 billion, up from US$460 billion in the previous quarter. Keep betting on the Gemini model In the field of AI, Alphabet is accelerating model research and development to cope with competition from companies such as OpenAI and Anthropic. Alphabet CEO Sundar Pichai said on the earnings call that the company is still a leader in many areas of AI, but also admitted that some areas still need improvement. Recently, the development of Alphabet's Gemini model has become the focus of market attention. The company launched the Gemini 3 model in November last year, which was once seen as a product that could compete with top models from OpenAI and Anthropic, but the competitive landscape has since changed. Bloomberg previously reported that Alphabet postponed the release of its latest flagship model, the Gemini 3.5 Pro, because the model failed to meet the company's internal goals in testing. Affected by this news, Google's stock price fell 2% last Friday (17th). Technology stocks as a whole were also affected by market fluctuations that day. In addition, Google's stock price was also affected in June when two key artificial intelligence researchers resigned and joined Anthropic, which attracted market attention.
During the conference call, Pichai said that Gemini 3.5 Pro is still in the testing phase and the company has begun pre-training on the next-generation model Gemini 4. He also said that currently, nearly 90% of Fortune 100 companies are using Gemini enterprise services and Google Cloud security infrastructure. On Tuesday (21st), Google released three new Gemini model versions, including Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber for the network security field. However, this launch does not introduce a new Pro version model, and the Gemini Pro series has not been updated since February. Meanwhile, other AI companies continue to roll out new models. Anthropic released the Mythos 5 and Fable 5 models in June, and OpenAI announced GPT-5.6 this month. In terms of search business, Alphabet's search-related revenue increased 17% year-on-year in the second quarter, in line with analyst expectations. The company said this growth was driven in part by artificial intelligence capabilities. Currently, Google Gemini application has 950 million monthly active users. Last year, the company launched "AI Mode," a feature that uses chatbot-like interactions to answer search questions and reduce the proportion of links displayed in traditional search results. In addition, Google will launch the “AI Overviews” function in 2024 to generate customized summaries for search results. In May this year, the company said that the number of monthly users of this feature exceeded 2.5 billion.
During the conference call, Pichai said that Gemini 3.5 Pro is still in the testing phase and the company has begun pre-training on the next-generation model Gemini 4. He also said that currently, nearly 90% of Fortune 100 companies are using Gemini enterprise services and Google Cloud security infrastructure. On Tuesday (21st), Google released three new Gemini model versions, including Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber for the network security field. However, this launch does not introduce a new Pro version model, and the Gemini Pro series has not been updated since February. Meanwhile, other AI companies continue to roll out new models. Anthropic released the Mythos 5 and Fable 5 models in June, and OpenAI announced GPT-5.6 this month. In terms of search business, Alphabet's search-related revenue increased 17% year-on-year in the second quarter, in line with analyst expectations. The company said this growth was driven in part by artificial intelligence capabilities. Currently, Google Gemini application has 950 million monthly active users. Last year, the company launched "AI Mode," a feature that uses chatbot-like interactions to answer search questions and reduce the proportion of links displayed in traditional search results. In addition, Google will launch the “AI Overviews” function in 2024 to generate customized summaries for search results. In May this year, the company said that the number of monthly users of this feature exceeded 2.5 billion.