Alphabet raised its full-year 2026 capital expenditure forecast to between $195 billion and $205 billion as demand for Google Cloud AI infrastructure continues to exceed available compute capacity. The company previously expected CapEx of $180 billion to $190 billion. Alphabet spent $44.9 billion during the second quarter, with the vast majority directed toward technical infrastructure supporting AI models, Google Cloud customers, Search, YouTube, and other services.
Approximately 60% of Alphabet’s second-quarter technical infrastructure investment went toward servers, while 40% funded data centers and networking equipment. CFO Anat Ashkenazi said Alphabet accelerated capacity deliveries to address rising demand but expects supply constraints to persist. The company plans to use additional third-party compute capacity during the third quarter as a temporary bridge while it expands internally owned infrastructure. Alphabet also expects CapEx to increase significantly again in 2027.
Google Cloud revenue increased 82% year-over-year to $24.8 billion, while Cloud operating income more than tripled to $8.8 billion. Cloud backlog reached $514 billion after increasing by more than $50 billion sequentially, with just over half expected to convert into revenue during the next 24 months. Alphabet said core Google Cloud Platform services, AI infrastructure, and enterprise AI solutions all contributed to growth. The company also began recognizing revenue from sales of complete TPU systems deployed in customer data centers.
• Q2 2026 CapEx reached $44.9 billion
• Full-year 2026 CapEx guidance increased to $195 billion-$205 billion
• Previous CapEx guidance stood at $180 billion-$190 billion
• Approximately 60% of technical infrastructure spending went toward servers
• Approximately 40% went toward data centers and networking equipment
• Google Cloud revenue increased 82% to $24.8 billion
• Google Cloud operating income reached $8.8 billion
• Google Cloud operating margin increased to 35.6%
• Cloud backlog reached $514 billion
• Alphabet expects to recognize just over 50% of the backlog during the next 24 months
• Alphabet will use third-party capacity as a temporary bridge while internal capacity expands
• Management expects CapEx to increase significantly again in 2027
• Higher infrastructure investment will increase depreciation, energy, and data center operating expenses
• Q2 free cash flow was negative $5.9 billion because of elevated CapEx
• Alphabet ended the quarter with $242.5 billion in cash and marketable securities
Alphabet’s infrastructure strategy spans Google-designed TPUs, NVIDIA accelerators, Google Axion CPUs, data centers, networking, and software designed to support workloads across GPUs and TPUs. Google Cloud now offers NVIDIA Vera Rubin systems and Google TPU 8t and 8i platforms. Its Virgo Network architecture is designed to connect as many as one million AI accelerators across multiple data center sites as a unified computing environment.
Alphabet said its software stack supports JAX, PyTorch, vLLM, and SGLang, allowing customers to move workloads between GPU and TPU environments. The company also introduced an agent-optimized Axion CPU that it says delivers 30% better performance per dollar than comparable offerings. Google continues to allocate compute first to frontier model development, followed by its consumer platforms and enterprise AI services.
The company said Gemini API traffic reached approximately 22 billion tokens per minute, up from 16 billion one quarter earlier. Nearly 500 Google Cloud customers each processed more than one trillion tokens during the past year, while more than 2,000 enterprises consumed over 100 billion tokens. Alphabet said the continued increase in model usage remains one of the main sources of pressure on available infrastructure.
“Our goal is to invest as long as we see an attractive return on that investment,” said Anat Ashkenazi, CFO of Alphabet. “While we have increased our capacity quite significantly over the past three years, the demand still outpaces that investment.”
🌐 Analysis: Alphabet’s revised CapEx outlook places the company among the largest infrastructure investors in the AI market and shows that compute availability, rather than customer demand, remains the immediate constraint on Google Cloud growth. The allocation of 40% of technical infrastructure investment to data centers and networking equipment also highlights the growing importance of scale-out and scale-across fabrics, optical connectivity, power delivery, and distributed data center capacity alongside accelerators.
Google’s decision to sell complete TPU systems into customer data centers expands its infrastructure model beyond conventional cloud consumption. It also positions Alphabet to compete more directly across the AI systems market while preserving internal capacity for Gemini training, inference, Search, and other first-party workloads.
