The $205 Bn AI Bet Google Can’t Afford to Lose

A year ago, spending nearly $200 billion on AI infrastructure would have sounded reckless. Today, Google is doing exactly that. The company has raised its 2026 capital expenditure guidance to $195 billion-$205 billion while also reporting one of its strongest quarters ever.

For the second quarter of 2026, Alphabet’s revenue rose 24% year-over-year to $119.8 billion, driven by accelerating AI adoption across Google Search, Cloud and Gemini. Google Cloud emerged as the biggest growth engine, with revenue jumping 82% to $24.8 billion, while Google Search and other revenues grew 17% and YouTube advertising revenue increased 13%.

“Our AI investments are redefining what’s possible across every part of our business,” said Google CEO Sundar Pichai in a statement.

Google’s AI ambitions are now large enough that operating cash flow alone is no longer sufficient. In June, Alphabet raised $84.75 billion through an equity offering to fund AI infrastructure and compute expansion, while its debt has climbed to nearly $100 billion.

“We continue to generate very healthy, strong cash flow from operations. That’s our first source of funding. Then we look at debt, and most recently we did the equity raise. We have expanded our debt portfolio quite significantly over the past 12 months,” said CFO Anat Ashkenazi during the earnings call.

Ashkenazi added that Google had accelerated the delivery of AI infrastructure, resulting in higher capital expenditure. The company also reported negative free cash flow for the first time in at least a decade as AI spending surged.

AI Drives Search and Gemini Adoption

Alphabet attributed much of the quarter’s momentum to its AI products. Pichai said Google’s AI investments are translating into higher Search usage, with users increasingly adopting a unified experience combining AI Overviews and AI Mode.

“Our momentum starts with Search, where people are adopting one seamless Search experience across AI Overviews and AI Mode,” Pichai said. 

Google also recently expanded its Gemini portfolio with Gemini 3.6 Flash, Gemini 3.5 Flash-Lite and Gemini 3.5 Flash Cyber, underscoring its push beyond general-purpose AI. 

The cybersecurity-focused Gemini 3.5 Flash Cyber is positioned as a rival to Anthropic’s Mythos. “With our new Google AI Threat Defense, we are really excited to bring our new cyber model and CodeMender to help our customers defend against AI threats,” Pichai said.

Moreover, Google acquired Wiz in March 2026 to strengthen its AI-driven cybersecurity and multicloud security capabilities. “Nearly 90% of Wiz customers are using AI-powered security features,” Pichai said, adding that the customer base includes Morgan Stanley, telecommunications providers such as Telus, healthcare organisations such as Texas Children’s Hospital, software companies such as Atlassian, and several government agencies

Pichai further revealed that the company has also started pre-training Gemini 4 while Gemini 3.5 Pro remains in testing.

Gemini’s adoption also continued to accelerate during the quarter. Google said its APIs now process 22 billion tokens every minute, up from 16 billion in the previous quarter. 

Google also reported growing adoption across its AI products. More than nine million developers built with Gemini models each month, the Gemini app reached 950 million monthly active users, and AI Mode surpassed the one billion monthly active user mark.

The search giant also said its Gemma open models had been downloaded more than 900 million times, while the latest Gemma 4 models had crossed 300 million downloads since their launch in April.

Google Cloud Rocks

Google Cloud continued to outperform, with revenue reaching $24.8 billion, compared with $13.6 billion a year ago. Pichai said Cloud’s growth was driven by an integrated AI stack spanning chips, models, security, data platforms and enterprise agents.

He added that nearly 90% of the Fortune 100 now use Gemini Enterprise. Google is also seeing customers expand existing commitments by more than 50%, while transactions through Google Cloud Marketplace grew over seven times year-over-year.

Pichai revealed nearly 500 Cloud customers have each processed more than one trillion AI tokens over the past year, while more than 2,000 enterprises consumed over 100 billion tokens. 

Strong demand for enterprise AI also pushed Google Cloud’s backlog to $514 billion in the second quarter, up by more than $50 billion from the previous quarter.

While Google’s core businesses delivered a strong quarter, a large part of Alphabet’s record profit came from its investment portfolio.

The company booked $99 billion in realised and unrealised gains on equity securities, adding $77.1 billion to net income after taxes. Much of the gain was driven by the rising valuations of SpaceX and Anthropic, making investment income a significant contributor to the quarter’s earnings.

The TPU Underdog Story

This quarter was different for Google, as it recognised revenue from TPU sales for the first time. “We started delivering TPU systems to customer data centres in the second quarter. We continue to expect to recognise a relatively small portion of the revenues from our existing TPU system sales agreements this year, ramping as we exit 2026,” said Ashkenazi, adding that the vast majority of the revenues from these agreements will be realised in 2027.

Google is monetising its custom AI chips beyond its own cloud business. In May, the company partnered with Blackstone to create a standalone TPU cloud, allowing enterprises to rent Google’s Tensor Processing Units as a compute-as-a-service offering.  

Ashkenazi said TPU system sales are included in Google Cloud’s backlog alongside GCP agreements.

Meanwhile, Pichai said the company’s priority is to allocate enough TPUs to compete at the frontier of AGI development. “We are allocating as much of the available compute for frontier model development, as well as serving both our consumer business and our enterprise business in terms of first-party models,” he added.