The AI cash-burning spree has been so wild that Tesla and Google posted a rare negative cash flow in the second quarter.

Image Source: The Interface Gallery

Massive investments in AI are eating into the cash reserves of tech giants.

On the evening of local time July 22, Tesla and Google's parent company Alphabet simultaneously released their financial reports for the second quarter of fiscal 2026. Both companies saw substantial growth in revenue and core AI businesses, yet they shared a rare identical financial status: their free cash flow turned negative this quarter after being positive for years, with massive investments in AI and robotics acting as the core drag.

According to the financial report, Tesla posted revenue of $28.24 billion in the quarter, a year-on-year increase of 26%, hitting a record for the same period. The number of paid FSD subscription users continued to rise, and energy storage equipment deliveries set a new quarterly peak. However, its operating profit was only $398 million, plummeting 57% year-on-year, and the operating profit margin fell to 1.4%.

Tesla's R&D investment and capital expenditure increased significantly. Projects such as AI infrastructure, Robotaxi, and the Optimus humanoid robot continued to advance, putting certain pressure on the profitability of the quarter.

Against this backdrop, Tesla's capital expenditure in the second fiscal quarter reached $5.789 billion, surging 142% year-on-year. Its operating cash flow was $4.69 billion, and free cash flow turned negative, with an outflow of about $1.09 billion. This marks the first time in nearly two years that its cash flow has turned negative.

On the other side, Alphabet's total revenue for the quarter was $119.796 billion, up 24% year-on-year. Among this, Google Cloud's revenue was $24.768 billion, skyrocketing 82% year-on-year. The enterprise version of Gemini has covered 90% of the Fortune 100 companies. Its advertising and search businesses maintained double-digit growth, and its net profit attributable to shareholders reached as high as $112.107 billion, a year-on-year increase of nearly 300%.

However, Alphabet's capital expenditure in the second quarter was $44.924 billion, and its operating cash flow for the period was $39.069 billion. It finally recorded a negative free cash flow of $5.855 billion, which is also the first time the company has seen negative single-quarter cash flow since its listing.

In terms of stock prices, neither of the two performance reports boosted market sentiment.

After Tesla released its financial report, its stock price closed at $374.01, with a total market value of $1.4 trillion, and it once fell by more than 5% in after-hours trading. Alphabet's stock price once dropped by more than 4% after hours, closing at $341.91 with a total market value of $4.16 trillion. Investors are generally concerned that sustained high capital expenditure will squeeze profits and cash reserves for a long time.

Looking specifically at where the two companies' investments are going, Tesla's capital expenditure focuses on the physical AI track. The first area is AI computing power factories: the Cortex 1 and 2 computing power bases in Texas are being continuously expanded for model training of FSD and the Optimus humanoid robot. At the same time, the construction of Austin's self-developed semiconductor factory is advancing to ensure the supply of in-vehicle and robot chips. The second is the transformation of the Optimus humanoid robot production line: the Fremont factory has shut down the old Model S and X production lines, fully replaced the robot assembly equipment, and built a new dedicated manufacturing plant in Texas. The procurement of the entire production line equipment and infrastructure investment cost a huge amount. The third is the production line supporting the mass production of Cybercab self-driving taxis and the expansion of global Robotaxi operation and maintenance stations, coupled with the expansion of Megapack energy storage and 4680 battery production capacity. Multiple manufacturing infrastructure projects are spending money simultaneously, directly consuming the current operating cash.

Google's large expenditures are concentrated on the construction of AI computing power infrastructure. The company is significantly expanding its global data centers, purchasing a large number of TPUs and NVIDIA accelerated chips to build the Gemini training cluster, and simultaneously expanding the Virgo supercomputing network to support the model invocation demand of 22 billion Tokens per minute. The company is also increasing investment in the R&D of self-developed Axion CPUs and secure AI platforms, and building and acquiring supporting AI R&D plant and equipment to pave the way for a substantial increase in capital expenditure throughout the year.

However, neither company has plans to reduce or postpone their investments.

Vaibhav Taneja, Tesla's Chief Financial Officer, stated that the company's capital expenditure in 2026 is expected to exceed $25 billion and will continue to grow in the next two to three years. Elon Musk, Tesla's CEO, emphasized that Tesla should advance its investments "as quickly as possible" while avoiding waste to support long-term development directions such as autonomous driving, AI, and robotics.

Sundar Pichai, Google's CEO, said in the financial report: "Our investments in artificial intelligence are redefining the development potential of the company's entire business lines." In his view, the above performance proves that Google's differentiated, full-stack artificial intelligence layout is creating tangible and quantifiable value for global individual users, enterprise clients, and partners.

According to Alphabet's disclosure, the company has raised its full-year 2026 capital expenditure range to $195 billion - $205 billion, and clearly emphasized that the investment scale in 2027 will be further expanded.

In fact, in addition to Google, OpenAI, Anthropic, Microsoft, and others also released investment plans on the same day. OpenAI raised its planned computing power expenditure by 2030 from $600 billion to $750 billion; Anthropic signed an investment agreement of up to $50 billion with AMD; Microsoft will provide billions of dollars to Mistral AI, a French AI startup, to build GPU data centers in Europe.

In June, Amazon added $13 billion for investment in AI and cloud infrastructure in India. From 2026 to 2030, its total investment in local AI cloud will exceed $21 billion; in July, Meta announced that it would add $40 billion to its large data center campus in Louisiana, USA, and plans to increase its computing power capacity to at least 5 GW.

Shay Boloor, Chief Market Strategist at Futurum Equities, believes that in the past, tech platforms were asset-light businesses, with revenue growth far exceeding investment needs. After AI reconstructs the business model, companies must continuously spend money to build physical computing power and manufacturing production lines. The scale of capital consumption is unprecedented, and pressure on cash flow will become the norm.

Currently, Google's Gemini, Tesla's FSD, and other products have shown commercial increments, but computing power and robot manufacturing are three-to-five-year long-cycle investments, with no cash flow return visible in the short term. Once the subsequent AI demand falls short of expectations, or intensified industry competition drives down product gross margins, the cash flow gap caused by previous sustained high capital expenditures may directly lead to a decline in performance and valuation.

This article is from the WeChat Official Account "The Interface News", written by Song Jiannan, and published by 36Kr with authorization.