Google Is Spending Even More on AI Than Investors Expected

Key Facts

• Google raised its 2026 capital expenditure forecast to between $195 billion and $205 billion, an increase of up to $15 billion from the April guidance.
• CFO Anat Ashkenazi stated that capital expenditures will be 'significantly' higher next year as the company continues to acquire more compute capacity.
• Google signed a $920-million-per-month agreement with SpaceX for access to its data centers.
• Google Cloud reported a backlog of $514 billion in contracted work, up from $460 billion in the prior period.
• Q2 2026 total revenue reached $119.8 billion (up 24% YoY), with Google Cloud revenue at $24.7 billion (up 82% YoY).
• Google delayed its Gemini 3.5 Pro model due to internal benchmark failures, while Anthropic and OpenAI advanced their frontier models.
• Gemini consumer app reached 950 million monthly active users, comparable to ChatGPT, though active usage data suggests ChatGPT leads.
• Alphabet's stock fell 3.9% following the earnings call amid investor concerns over returns on infrastructure spending.

AI Infrastructure Spending Reaches New Heights

The revised capital expenditure guidance underscores a broader trend across the technology industry: demand for artificial intelligence computing resources continues to outstrip supply. Google, despite operating one of the largest cloud computing platforms globally, must still contract third-party capacity to meet the needs of its customers and internal projects. The $920-million-per-month deal with SpaceX exemplifies the lengths to which even hyperscalers must go to secure compute power.

The increased spending is directed primarily toward building new data centers and purchasing compute capacity from hyperscalers, neoclouds, and other suppliers. This expansion is not limited to 2026; the CFO explicitly stated that expenditures will be significantly higher next year, indicating a multi-year commitment to scaling AI infrastructure. For context, the revised forecast of up to $205 billion represents a substantial increase from the $180 billion to $190 billion range that analysts had broadly anticipated.

Google Cloud's Accelerating Momentum

While the spending spree raises eyebrows, the underlying demand trends provide some reassurance to investors. Google Cloud's backlog swelled to $514 billion, reflecting long-term commitments from enterprise customers who are increasingly turning to the platform for AI workloads. The division has emerged from being a solid third-place cloud provider in the U.S. to the one with the strongest momentum for AI workloads, thanks in large part to its Tensor Processing Units (TPUs). These custom chips offer compelling performance and lower operating costs for inference tasks, making them particularly attractive for businesses deploying large language models.

In the second quarter of 2026, Google Cloud generated $24.7 billion in revenue, an 82% year-over-year increase. This growth rate far outpaces that of its parent company's overall revenue, which grew 24% to $119.8 billion. The cloud division's acceleration is a testament to the surging demand for AI-powered services, from generative AI applications to machine learning pipelines.

Challenges in AI Model Development

Despite its infrastructure advantages, Google faces increasing competition in the frontier model race. Anthropic's Fable and Mythos models have dominated the AI conversation for over a month, while OpenAI's Codex is rapidly gaining enterprise traction. Google's planned Gemini 3.5 Pro, which was intended to be its major push into coding and enterprise workflows, has been delayed because it failed to meet internal benchmarks. Additionally, its cyber AI model remains in preview for select partners rather than being released broadly.

The consumer front offers a mixed picture. Google's Gemini app reached 950 million monthly active users, a figure that rivals ChatGPT's user base. However, much of that usage may stem from deep integration with Google Search, such as AI Overviews, rather than standalone app engagement. Third-party app intelligence platforms indicate that ChatGPT is used considerably more actively, suggesting that Gemini's reach does not translate into equivalent daily or weekly engagement.

Investor Concerns and Industry Implications

The revelation of increased spending this year and further rises next year has unsettled some investors, contributing to a 3.9% decline in Alphabet's stock price since the earnings call. This skepticism mirrors broader concerns across the AI sector about whether hyperscalers can generate sufficient returns on their unprecedented infrastructure investments. The capital expenditures required to build and operate AI data centers are staggering, and the path to profitability remains uncertain for many players.

Yet, for enterprise customers, the implications are clear: more AI capacity is becoming available, but significant investments and competition for compute are likely to persist well into next year. Companies that rely on cloud AI services may face higher costs or longer lead times for provisioning resources. Meanwhile, the competitive dynamics among cloud providers—Amazon Web Services, Microsoft Azure, and Google Cloud—continue to shift, with Google Cloud capturing a disproportionate share of AI-related workloads.

Google's rising capital expenditure also highlights a broader reality: demand for computing infrastructure continues to outpace supply across the industry. This supply-demand imbalance is driving innovation in chip design, data center construction, and energy procurement. It is also prompting hyperscalers to forge partnerships with unconventional partners, such as SpaceX, to access specialized data center capacity.

Outlook for the Remainder of 2026

Looking ahead, Google's capital expenditures will remain elevated, and the company shows no signs of slowing its AI infrastructure buildout. The commitment to spending significantly more in 2027 suggests that the current cycle of investment will extend well beyond the near term. For investors, the key question is whether the revenue growth from Google Cloud and AI products will eventually justify the tens of billions of dollars being poured into infrastructure. For now, the backlog of $514 billion provides a strong foundation, but the actual conversion of contracted work into recognized revenue will take time.

Meanwhile, Google's competitors are not standing still. Amazon and Microsoft are also ramping up their own AI spending, and startups like Anthropic and OpenAI continue to push the boundaries of model capabilities. The AI landscape is evolving rapidly, and Google's ability to lead in both infrastructure and model performance will determine its long-term position. The delayed Gemini 3.5 Pro model looms as a critical test: if Google can regain its footing in frontier AI, it could solidify its cloud and consumer offerings. If not, it risks ceding the high ground to rivals despite its vast infrastructure investments.

Source: TechRepublic News