Here's the story of how two companies added $8 trillion in value, each in different ways, but both powered by artificial intelligence.
Nvidia designs the processors that train and run artificial intelligence models. Alphabet builds the models, runs the cloud infrastructure and sells the AI-powered services that reach billions of users. One sells the pickaxes; the other mines the gold and sells the jewellery. Both are now worth more than virtually every economy on Earth.
Nvidia's market capitalisation has broken through $5 trillion for the second time, reclaiming a landmark first reached in October 2025 and lost during a volatile start to 2026. Alphabet, up roughly 25% year to date and 158% over the past twelve months, is closing the gap fast at $4.8 trillion, briefly overtaking Nvidia in after-hours trading amid reports that Anthropic had committed $200 billion in cloud spending over five years. The two companies are now separated by less than $400 billion, a margin that could vanish with a single earnings surprise.
Three years ago, Nvidia was worth $300 billion and Alphabet $1.2 trillion. The combined $8 trillion they have added since is larger than the GDP of every country bar the United States and China.
The hardware colossus
Nvidia's ascent has been the more dramatic of the two because it came from a lower base and was driven by a single, overwhelming insight: that the graphics processing units it had spent decades refining for video games were the only chips capable of training the neural networks behind generative AI.
When OpenAI's ChatGPT ignited public demand for AI in late 2022, every major technology company scrambled to build or expand GPU clusters, and Nvidia was the only supplier that could deliver at scale. Full-year revenue for the fiscal year ending January 2026 reached $215.9 billion, up 65%, with data centre sales accounting for the vast majority. Analysts project fiscal 2027 revenue of $371 billion, with net income potentially exceeding $200 billion.
The company's competitive moat extends beyond the chips themselves. Its CUDA software platform, built over nearly two decades, is the standard programming environment for AI development, creating switching costs that make it difficult for customers to move to rival hardware even when alternatives exist. Nvidia's next-generation Vera Rubin processors, due later this year, promise to cut inference costs by 90% compared with the current Blackwell generation, a claim designed to keep customers locked in as the industry shifts from training new models to running them at scale.
The intelligence layer
Alphabet's rally has been powered by a different but equally compelling narrative: the company is monetising AI across every layer of its business simultaneously.
Google Cloud crossed $20 billion in quarterly revenue for the first time in the first quarter of 2026, growing 63% year on year, with a contracted backlog that nearly doubled in a single quarter to $460 billion. Google Search revenue rose 19% to $60.4 billion, defying years of predictions that AI chatbots would cannibalise traditional search. Net income for the quarter surged 81% to $62.6 billion.
Where Nvidia sells infrastructure, Alphabet is building the entire stack: its own Tensor Processing Units reduce reliance on Nvidia's GPUs, its Gemini large language model competes directly with OpenAI's GPT and Anthropic's Claude, and its cloud platform hosts many of the AI startups that are driving demand for compute. The Gemini app has reached 750 million monthly active users, up from 450 million at the start of 2025. Google now shows advertisements in roughly a quarter of AI-generated search results, turning what was supposed to be an existential threat into a new revenue stream.
The spending arms race
The numbers fuelling these valuations are staggering even by the standards of big technology. Alphabet has guided capital expenditure of $185 billion to $190 billion for 2026. Microsoft has committed $190 billion. Amazon is expected to spend $200 billion. Meta has raised its range to between $125 billion and $145 billion.
Almost all of this money flows, directly or indirectly, to Nvidia. The company's GPUs and networking equipment sit at the centre of a capital cycle in which cloud providers spend tens of billions on chips, AI startups commit hundreds of billions to cloud contracts, and investors bid up the shares of everyone involved on the assumption that demand will continue to compound.
NVIDIA anticipates a combined $1 trillion in data centre revenue from its Blackwell and Vera Rubin processor families across 2026 and 2027. If that figure is even directionally correct, it would represent the fastest revenue ramp for any product category in corporate history.
Circular economics and hidden risks
The interdependence between the two companies is deepening. Alphabet is simultaneously a major Nvidia customer and a competitor building its own chips. Anthropic, in which Alphabet has invested up to $40 billion, has committed $200 billion to Google Cloud while also signing compute deals with Amazon, Microsoft and SpaceX. The money circulates in an ever-expanding loop: investment begets cloud contracts, cloud contracts beget chip orders, chip orders beget revenue that justifies further investment.
Related reading
- AI can write code and prove theorems. It still can't book me a restaurant
- Karpathy's 630-line script points the way to self-improving AI agents
- The era of zero-click: Google is spending $185bn to break the thing it depends on
The risk is that the loop breaks. If AI revenue growth disappoints, if a geopolitical shock disrupts chip supply chains, or if the market simply decides that valuations have run ahead of fundamentals, the correction could be severe. Nvidia trades at roughly 30 times forward earnings, a level that leaves little room for error. Alphabet's price-to-earnings ratio sits at a similar multiple, elevated by the assumption that cloud and AI revenue will continue to accelerate.
For now, investors are choosing to bet that the AI infrastructure buildout is still in its early stages and that the two companies best positioned to profit from it are the same two racing towards $5 trillion. The question is no longer whether they will get there, but which one arrives first at $10 trillion.
