Alphabet's 160% Rally in a Year: Why Owning the Whole AI Stack Changes Everything

For years, the dominant narrative about Google and AI was one of squandered advantage. The company had invented the transformer architecture. It had the data, the compute, the talent, and the distribution. And yet it had been persistently cast as the company that was losing the AI race — slow to ship, culturally cautious, and perpetually one announcement cycle behind OpenAI.

That narrative is now in full retreat. Alphabet’s stock has gained approximately 160% over the past twelve months, hitting an all-time high on May 8, 2026, and briefly — in after-hours trading — surpassing Nvidia to claim the title of world’s most valuable company. The market is pricing in something specific: that owning most of the AI stack is worth more than owning any single part of it.

The Stack Thesis

Gene Munster, managing partner at Deepwater Asset Management and one of the more closely watched technology analysts on Wall Street, articulated the bull case in terms that have become close to consensus: “Google is one of the two best-positioned AI companies because they own most of the stack. Chips, models, infrastructure and distribution. On top of that, they’re nicely profitable.”

The “stack” framing matters because it identifies what separates Alphabet from companies that are AI-enabled versus AI-native at the infrastructure level.

Chips: Google’s Tensor Processing Units (TPUs) are now on their seventh generation, purpose-built for the matrix multiplication operations that dominate AI training and inference. Google doesn’t sell TPU access at scale the way Nvidia sells GPU capacity — instead, TPUs give Google structural cost advantages on its own workloads that external cloud competitors cannot easily replicate.

Models: Gemini 4 is expected at I/O next week. Gemini 1.5 Pro already holds benchmark leadership positions in long-context reasoning. The model family spans ultra-efficient on-device versions to the flagship research-grade Ultra tier, covering the full deployment spectrum from mobile to data center.

Infrastructure: Google has invested more aggressively in physical AI infrastructure than at any point in its history. The company is on track to spend approximately $75 billion in capital expenditure in 2026, the majority earmarked for data center expansion, networking, and custom silicon.

Distribution: This is the most underappreciated element of the thesis. Google reaches more than two billion active Android devices. Chrome commands approximately 65% of global browser market share. Search processes an estimated 8.5 billion queries per day. YouTube is the world’s largest video platform. Each of these is a distribution channel through which Gemini capabilities can be delivered to users who never explicitly chose an AI product — they just kept using the Google products they already used.

Cloud: The Number That Changed the Conversation

The Q1 2026 earnings report that landed in late April was the catalyst that broke Alphabet’s stock to all-time highs. Google Cloud revenue reached $20 billion in the quarter — crossing that threshold for the first time — representing 63% year-over-year growth. The cloud backlog nearly doubled to $462 billion.

JPMorgan called the quarter “standout” and named GOOGL its “top overall pick” in technology, citing accelerating growth and a backlog that suggested demand visibility stretching years into the future.

The single most significant number beneath that backlog is Anthropic. The AI company behind the Claude model family — OpenAI’s primary frontier competitor — has reportedly committed to spending $200 billion on Google Cloud over five years. That commitment is, in itself, larger than Google Cloud’s entire annual revenue from two years ago. It reflects an irony that has not been lost on industry observers: Anthropic, which competes directly with Google’s Gemini models in the AI market, has structurally tied its infrastructure future to Google’s cloud.

A Briefly World-Leading Valuation

The moment Alphabet’s market cap ticked above Nvidia’s in after-hours trading on May 8 lasted only briefly — Nvidia closed the following session higher — but the symbolic significance was not lost. For most of the past 18 months, Nvidia has occupied the position of the AI infrastructure company that all others depend on. Alphabet’s temporary surpassing of it was a market statement that the value of AI infrastructure extends beyond GPU supply.

The question implicit in that valuation is whether the distribution advantage is real or theoretical. Google Search has not been displaced by ChatGPT or any other AI product; on the contrary, AI Overviews has increased query satisfaction metrics while maintaining advertising revenue. Android 17’s Gemini integration, announced at the Android Show on Monday, embeds AI capabilities at the OS level across billions of devices — a distribution reach that no AI startup can approach.

Bubble Warnings Are Getting Louder

Not everyone reads the Alphabet rally as justified. As the stock has climbed, so have the voices warning that the market is pricing in AI adoption curves that have not yet materialized in revenue at the pace implied by valuations.

The bear case centers on search durability: if AI-native interfaces eventually displace the ten-blue-links search paradigm at scale, Google’s core advertising business faces structural pressure that no amount of cloud growth can fully offset. Regulatory risk is a secondary concern, with antitrust cases in the US and EU still unresolved.

The more immediate concern among bears is concentration risk in the cloud backlog. A backlog of $462 billion sounds impregnable — but if Anthropic’s $200 billion commitment is included, and if Anthropic eventually diversifies its cloud spend or is acquired by a company with different infrastructure preferences, a significant portion of that backlog becomes contingent.

The I/O Test

The week of May 19 is Google’s next opportunity to either justify or strain the market’s confidence. The I/O keynote is expected to bring Gemini 4, hardware announcements including Project Aura XR glasses, and the full developer platform story for Android 17 and Aluminium OS.

A weak I/O — vague model announcements, no shipping hardware, developer features that feel derivative — would not immediately collapse the stock, but it would give bears ammunition and raise questions about whether the valuation has gotten ahead of the product timeline.

A strong I/O — Gemini 4 with demonstrable capability leaps, shipping hardware at accessible price points, developer tools that attract the builder ecosystem — would cement the thesis that Google’s AI transition is real and accelerating.

The stack is in place. The question is execution.