If you've been adding GOOGL or planning to reduce NVDA recently, you must read this before taking any action.
If you've been buying GOOGL or planning to sell NVDA recently because of the Anthropic $200 billion contract news, I suggest you pause. The other side of this deal is: 40% of Google Cloud's revenue over the next 5 years is tied to Anthropic alone, and the cash Anthropic has to pay Google each year is almost equal to its entire annualized revenue. This isn't a partnership; it's two companies betting on each other's survival. Understanding this layer is key to knowing the risks behind GOOGL's continued rise and NVDA's continued fall.
What Happened
On Tuesday, The Information reported that Anthropic has committed to investing $200 billion in Google Cloud over the next five years. It also disclosed: Anthropic's annualized revenue jumped from $30 billion two months ago to $44 billion, with a 70% gross margin, and a new round of financing values it at $900 billion. Alphabet rose 2% after-hours, with a market cap of $4.7 trillion, just one step away from Nvidia's position as the world's most valuable company.
The most common market interpretation is: the landscape of the AI cloud 'tripolar rivalry' is set – Google becomes the leader with the dual lock-in of Anthropic + TPU, AWS gets OpenAI, and Microsoft falls behind. It sounds clear. But when I flipped the numbers, I found this interpretation missed three things.
First: 40% of Google Cloud's future revenue is tied to one company.
Reuters' report contained a key figure: this $200 billion accounts for over 40% of Google Cloud's entire future revenue backlog of $460 billion. That means 40% of Google Cloud's revenue over the next 5 years depends on Anthropic alone. This level of concentration would be flagged red in any institutional due diligence report.
More subtly: Alphabet itself is also investing in Anthropic, up to $40 billion. So this is a closed loop: Google invests money in Anthropic → Anthropic uses the money to buy Google's cloud → Google's cloud data looks better → The value of Alphabet's Anthropic equity rises → Alphabet's paper profits surge. The $37.7 billion in paper profits on Alphabet's books in Q1 mainly came from Anthropic's valuation jump.
Second: The money Anthropic has to pay Google each year equals its entire annualized revenue.
Simple math: $200 billion divided by 5 years equals $40 billion in annual cash expenditure. Anthropic's current annualized revenue is $44 billion. A company's annual cash expenditure commitment is almost equal to its current total annualized revenue. The only way this works is if Anthropic's revenue continues to grow rapidly every single year going forward, and every round of financing goes smoothly without its valuation being crushed.
This isn't saying Anthropic will definitely have problems. But when you buy GOOGL, you need to clearly understand that you're not betting on Google's execution ability; you're betting that Anthropic will run smoothly every year for the next five years.
Third: TPU isn't being sold to anyone other than Anthropic.
This piece of information is the most counterintuitive. CoreWeave, Lambda, Nebius – three cloud companies specializing in AI compute rental – have all explicitly stated they won't adopt TPU in the short term. The public reason: 99% of customer demand is still for Nvidia GPUs.
This information discounts the story that Google has become an AI-native cloud. TPU currently has only one truly significant customer: Anthropic itself. The real meaning of the 'AI-native cloud' label, for now, is just Anthropic-dedicated infrastructure.
How it relates to you
If you've been buying GOOGL recently, the core assumptions you're betting on are that Anthropic will cough up $40 billion every year for the next 5 years without incident AND that TPU can open up markets beyond Anthropic. You need to assess the probability of these two assumptions yourself. GOOGL's current valuation has already priced in the optimistic side of this news; the 2% after-hours gain is basically the ceiling.
If you've been selling NVDA recently because Nvidia is being replaced, you can stop. 99% of AI cloud customers still want GPUs – this is a piece of underlying data that hasn't changed much. If NVDA falls along with sentiment tonight or this week, it's actually an opportunity, not structural weakness.
The most awkward one is Microsoft. The one truly hurt in the AI cloud 'tripolar rivalry' isn't 'Da', it's Azure – having lost the exclusive with OpenAI, there's no new story of the same caliber in its backlog. MSFT's rise today is just the market beta; the delayed reaction is still to come. If you're holding MSFT for AI, this is worth reconsidering.
$Alphabet(GOOGL.US) $NVIDIA(NVDA.US) $Microsoft(MSFT.US)
The copyright of this article belongs to the original author/organization.
The views expressed herein are solely those of the author and do not reflect the stance of the platform. The content is intended for investment reference purposes only and shall not be considered as investment advice. Please contact us if you have any questions or suggestions regarding the content services provided by the platform.
