Will Google overtake Nvidia as the world’s most valuable public company on any trading day before year-end, powered by the AI compute boom?

The Almost-King of AI

Alphabet is one good afternoon away from declaring moral victory over Nvidia. The stock just ripped 10% in a day, is up roughly 140% over 12 months, and sits at more than $4.6 trillion in market value. Nvidia, after a brief flirtation with $5 trillion, has slid back under $4.9 trillion.

Mathematically, Google needs only about a 4% pop—roughly $401 a share—to match Nvidia’s current market cap. Options prices put the odds of touching that level by mid-May at around 53%. Traders are literally pricing this as a coin flip.

Here’s my bet: that coin keeps landing on Nvidia’s side. Between now and the end of 2026, I don’t think there’s a single trading day where Alphabet’s closing market cap is higher than Nvidia’s.

Trillion-Dollar Tug-of-War

This isn’t just a scoreboard gag for people who memorize ticker symbols. It’s a fight over where the AI boom actually pays out: to the company that sells the picks and shovels, or the company that owns the gold mine and charges rent.

On one side, Nvidia: the GPU cartel with a software religion (CUDA), a near-fanatical developer base, and data-center revenue that’s exploded on the back of AI training and inference. Hyperscalers—Amazon, Microsoft, Google, Meta—are pointing roughly $710 billion at AI capex this year, and Nvidia is still the default destination for a terrifying share of it.

On the other side, Alphabet: the quiet compute landlord. Google Cloud just crossed $20 billion in quarterly revenue. The company runs one of the densest AI infrastructure stacks on Earth: custom TPUs, massive data centers, long-term energy contracts, global fiber, and software tying it all together. Sam Altman’s line that “compute is destiny” could have been written on a Google slide deck ten years ago.

The surface question is short-term—who grabs the market-cap crown first? The deeper one is structural—does value accrue to the compute seller (Nvidia) or the compute owner-operator (Alphabet)? My read: in this cycle, the seller still wins.

The Capex Mirage Helps Nvidia More Than Google

The headline number is obscene: around $710 billion of 2024 capex guided from the Big Four hyperscalers. That beat Wall Street’s expectations by about 8%, setting off another melt-up in chip stocks. The Nasdaq chip index ran roughly 35% in a month. Analysts used words like “historic” with a straight face.

But peel the label off and you find something less heroic: a big chunk of that capex “surprise” is just price inflation on scarce components, not a bigger buildout. RBC and BNP Paribas estimate that in some cases nearly half the apparent increase is higher memory costs, not more hardware.

In other words, a lot of this arms race is the same number of tanks at a higher sticker price.

That nuance matters. If hyperscalers were dramatically accelerating the quantity of AI infrastructure, you could argue the long-term win shifts toward whoever monetizes that capacity across search, ads, video, and cloud workloads—i.e., Google. But if the main story is scarcity-driven price hikes on the most critical pieces, then the power sits with whoever controls the choke points.

Right now, Nvidia is the choke point. Scarcity plus must-have hardware is not Google’s business model; it’s Nvidia’s. The more the capex story turns out to be about pricing power on GPUs and memory, the more justification there is for Nvidia’s premium.

Google’s Compute Empire Is Real. The Repricing Isn’t—Yet.

Alphabet’s recent breakout isn’t a hallucination. The market finally noticed that Google doesn’t just slap AI features onto search results; it owns one of the most sophisticated compute factories on the planet.

Google Cloud is scaling past $20 billion in quarterly revenue. Independent rankings of AI compute power put Google at or near the top, thanks to TPUs and a global infrastructure spine that was built for exactly this moment. If the top AI models converge in quality, as insiders increasingly expect, then the game tilts toward whoever can deliver inference cheaply, at global scale, with five-nines reliability.

That’s an argument for Alphabet eventually being valued more like an AI infrastructure utility and less like a glorified ad company with side projects.

But “eventually” is doing a lot of work. Right now, investors still treat Alphabet as a diversified conglomerate: search and YouTube throw off cash, Cloud is ascendant but not dominant, AI is a growth accelerant more than a standalone profit center. Nvidia, in contrast, is priced as the purest bet on the AI hardware supercycle. Every incremental dollar of AI capex feels like it walks straight into Nvidia’s P&L.

That purity is a feature in a mania. It’s also why, when options markets handicap Google overtaking Nvidia, they’re basically betting on short-term flows—post-earnings wobbles, gamma games, rotation within the Magnificent 7—not on a profound, durable re-rating of how AI economics work.

Why I’m Fading the Alphabet Upset

Over the next few weeks, a brief Alphabet overtake is entirely plausible. Nvidia has a habit of selling off after earnings—down after four of its last five reports—even when the numbers are strong. Meanwhile, Google just gave investors a clean, simple story: we are an AI compute company now.

Stretch that horizon to the end of 2026, though, and the hurdle changes. We’re no longer asking “Can Google tag Nvidia in a volatile week?” but “Will there be any day in the next couple of years where the closing tape says GOOGL > NVDA?”

For that to happen and stick, you probably need some combination of:

• A real digestion phase in AI capex that hits semis hardest, compressing Nvidia’s multiple while leaving Alphabet relatively unscathed.
• Faster-than-expected proof that hyperscalers can push meaningful AI workloads onto in-house chips (TPUs, Trainium, etc.), denting Nvidia’s growth edge.
• A step-change in Google’s AI monetization—Search and YouTube revenue accelerating and margins holding up—so investors rewrite the whole model around “compute utility” economics.

All of those may happen eventually. I’m just skeptical they line up fast enough, and brutally enough, to knock Nvidia below Alphabet even once on a closing basis in the next 120 weeks.

More likely: we get a few scares, some furious back-and-forth in the leaderboard, maybe an intraday moment where Alphabet ticks above Nvidia and finance Twitter declares the regime change. By the closing bell, Nvidia is still the more expensive addiction.

The Satirical Verdict

So here’s the call you can score later: through the final U.S. trading day of 2026, Alphabet will not end a single session with a higher market cap than Nvidia.

Google may own more of the AI empire—chips, cables, data centers, and the apps on top—but in this particular gold rush, the market is still happier paying the guy who sells the shovels than the landlord who reads philosophy about resource allocation.

Alphabet will get plenty of credit for building the world’s most powerful AI factory. Nvidia, for now, gets the money. When history writes up this era, it may note that Google was rich enough to buy all the compute in the world—and kind enough to let Nvidia invoice it.