Google’s parent company isn’t just betting on artificial intelligence — it’s building the infrastructure to host the future. On its latest earnings call, Alphabet revealed it will raise capital expenditures for 2025 to $85 billion, up $10 billion from previous estimates. That push comes as demand surges across its cloud business and AI services, reshaping what was once a search-first company into an infrastructure heavyweight.
At the core of this growth sits Google Cloud, which brought in $13.6 billion in Q2, a 31 percent year-over-year jump. More notable, though, is the profit story: the division reported $2.8 billion in operating income — more than double what it earned the previous year. After over a decade of losses, profitability has arrived, fueled largely by AI-related services and infrastructure demand.
CEO Sundar Pichai noted a sharp rise in high-value contracts, including more than twice as many deals worth over $250 million compared to last year. Customers, he said, are signing on fast — and they’re sticking around. However, the company isn’t celebrating just yet. Pichai also flagged persistent hardware supply constraints and looming infrastructure bottlenecks.
The company’s technical buildout reflects that reality. Alphabet spent over $14 billion on servers in Q2 alone, with more data centers expected to come online before year’s end. According to CFO Anat Ashkenazi, two-thirds of this quarter’s spending went to servers, with the remainder funding networking and facilities.
Beyond infrastructure, Alphabet’s search business is holding strong in the face of rising generative AI competition. AI Overviews are driving new engagement, particularly among younger users. Pichai insisted this is expanding the kinds of questions users ask — not replacing them.
Meanwhile, Waymo’s autonomous vehicle unit crossed 100 million miles on public roads, and Alphabet is steering more capital into next-gen bets, including smart glasses and AI-driven interfaces. While search remains central, it’s clear the future Alphabet envisions requires far more than queries — it needs compute, scale, and a global tech backbone.
