Google records negative cash flow for first time as AI capex surges to $44.9B

What happened?

Alphabet, Google's parent company, reported negative free cash flow of $5.9 billion in the second quarter on July 24, 2026, the first negative result since its IPO in 2004, according to its earnings report. Capital expenditure (capex) reached $44.9 billion, doubling year-over-year and exceeding the $39.1 billion generated from operations. The company raised its full-year capex guidance to between $195 billion and $205 billion, up from the previous $180-$190 billion.

CFO Anat Ashkenazi revealed that Google delivered TPU systems to customer data centers for the first time, a shift from the previous model where they were only leased through Google Cloud. Approximately 60% of the quarter's technical infrastructure investment went to servers, with the remaining 40% allocated to data centers and network equipment. Depreciation of property and equipment rose to $7.1 billion from $5.0 billion a year earlier.

Why is it important?

This milestone marks a turning point in Google's AI investment strategy. Although the negative cash flow is small relative to the company's size (free cash flow over the trailing twelve months remains positive at $53.3 billion), it indicates that Alphabet is willing to take on debt and issue equity to finance its AI infrastructure. The company is in a supply-constrained environment, according to Ashkenazi, and is leasing third-party capacity as a bridge while its own data centers come online.

The sale of TPU systems to customers like Anthropic — which signed an agreement in October 2025 to access up to one million TPUs — transforms part of the capex into inventory that will generate revenue in 2027. However, spending on data centers is not directly recovered; it is amortized over years.

Historically, Google had maintained positive free cash flow since its IPO in 2004, even during the 2008 financial crisis and the COVID-19 pandemic. The fact that it now records a negative quarter underscores the magnitude of its AI bet, comparable to the cloud infrastructure investments made by Amazon and Microsoft in the 2010s. For example, Amazon Web Services (AWS) aggressively invested in data centers between 2010 and 2015, with capex growing from $1.3 billion in 2010 to $8.5 billion in 2015, but maintained positive cash flow thanks to its retail business revenue. Google, in contrast, does not have such a lucrative business to offset it, so it turns to debt and equity issuance.

Consequences and outlook

The rise in capex will pressure the income statement through higher depreciation and energy costs. The Cloud segment will face margin pressure in the third quarter due to leasing external capacity. The company has a $40 billion construction program in Texas and a $1.5 billion expansion in Alabama.

For investors, the signal is mixed: on one hand, Google is betting big on AI with proprietary infrastructure; on the other, negative cash flow and rising debt could generate volatility. Analysts will watch whether the investment translates into future revenue, especially with TPU deals. According to Tom's Hardware, Google is in a supply-constrained environment, and Ashkenazi stated that leasing third-party capacity will create modest margin pressure in Cloud in the third quarter.

Comparatively, the situation echoes Meta (Facebook) in 2022, when the company invested heavily in the metaverse, leading to a 64% drop in its stock over a year. However, Google has a track record of generating long-term returns from infrastructure investments, as demonstrated with its own fiber optic network and data centers for YouTube. The difference is that AI requires even larger and faster investment.

The AI chip market is dominated by Nvidia, whose GPUs are most in demand. Google, with its TPUs, aims to compete by offering a customized alternative. The sale of TPU systems to customers like Anthropic — which signed an agreement in October 2025 to access up to one million TPUs — is a step toward monetizing its technology. However, revenue from these sales will not be recognized until 2027, adding uncertainty about the return on investment.

What readers should know

• Negative cash flow is not a crisis: the company has $53.3 billion in trailing twelve-month free cash flow.
• Investment is concentrated in servers (60%) rather than construction, indicating a priority on computing capacity.
• TPU sales to external customers are a new revenue source, but revenue will be recognized mainly in 2027.
• The CFO warned that capex will continue to rise in 2027, suggesting the intensive investment phase will persist.
• The company is leasing third-party capacity, which will pressure Cloud margins in the short term.
• Spending on data centers is amortized over the long term, while servers (TPUs) can generate revenue faster if sold to customers.

In summary, Google is executing an aggressive AI infrastructure buildout strategy, sacrificing short-term cash flow to position itself in the chip and data center market. Success will depend on whether AI demand justifies the massive spending. If generative AI adoption accelerates, as analysts like Gartner predict — estimating the AI market will reach $300 billion by 2027 — Google could be well-positioned. However, if demand slows, the company could face overcapacity and significant financial pressure. Investors should closely monitor free cash flow trends and Cloud revenue in the coming quarters.