NVIDIA Just Printed $81.6 Billion in 90 Days. Here Is Why I Am Still Bullish.

NVIDIA just dropped a Q1 fiscal 2027 print that broke every model on Wall Street. Revenue of $81.6 billion. Up 85% year over year. Up 20% in just one quarter. Data center alone did $75.2 billion, up 92% year over year. Networking inside that segment grew 199%. Let that sink in for a second. A trillion dollar company is growing networking revenue at almost triple. That does not happen in nature. That happens when you are the only person in the room selling the thing everyone needs.

I have been long NVIDIA since well before most of finance Twitter could spell Blackwell. I get the bear arguments. I have heard every one of them. Concentration risk. Hyperscaler capex peak. China export controls. Custom silicon from Google and Amazon. I have read it all, and I still think this stock is mispriced. Here is the honest math on why.

NVIDIA is not an AI play. It is the toll booth on the most aggressive infrastructure build out in the history of capitalism, and the toll is going up, not down.

Revenue of $81.6 billion. Data center revenue of $75.2 billion, of which $60.4 billion was compute and $14.8 billion was networking. Networking grew 199% year over year. That number is not a typo. Hyperscaler revenue inside data center sat at roughly 50%. The other 50% came from sovereign AI buildouts, neoclouds like CoreWeave, enterprise, and industrial. That diversification is the part the bears never mention.

For years the bear case was "what happens when Meta and Microsoft slow down." The answer in Q1 is "every government on earth is now buying GPUs and the neoclouds are buying as much as they can finance." Saudi Arabia announced a sovereign AI deal worth tens of billions. France is doing it. UAE is doing it. India is starting. This is no longer a four customer story. This is becoming a fifty customer story, and the new customers are price insensitive because they are buying national security, not Q4 earnings leverage.

Blackwell was supposed to ramp in late 2024. It slipped to early 2025. Bears called the trade dead. Then it shipped, then it ramped faster than Hopper did at the same point in its cycle, and by Q1 2026 it is already pulling the entire data center segment up 92% year over year. Blackwell 300 is shipping now. Rubin is the next platform after that, and Jensen has already told everyone it lands in 2026 with another step function in performance per watt.

This is the part Wall Street keeps getting wrong. They think of NVIDIA as a cyclical hardware company on a 4 year cycle. The reality is that NVIDIA is now on an annual cadence. Every year they ship a new architecture. Every year the performance per dollar goes up. Every year the cost to train a frontier model collapses. That collapse does not slow demand. It accelerates it, because it unlocks new use cases that were previously uneconomic.

Yes, Google has TPU. Yes, Amazon has Trainium. Yes, Microsoft is working on Maia. None of them are going to dent NVIDIA in this cycle, and here is why nobody on CNBC will say it plainly. Custom silicon works great if you are willing to lock yourself into one cloud, accept slower software, and write your model around the hardware. NVIDIA works for everything. CUDA is sixteen years of compounded developer mindshare. Every PhD coming out of every machine learning program in the world learned on NVIDIA. Every research paper benchmarks on NVIDIA. The software moat is real, it is deep, and it is widening, not narrowing.

The hyperscalers will keep building custom silicon for the same reason Walmart sells store brand cereal. It is a negotiating tool. It is not a replacement.

The bear case I hear most is "hyperscaler capex cannot keep growing at this rate forever." Correct. It cannot. But it does not have to. Look at the actual numbers. Microsoft, Meta, Google, and Amazon combined are spending around $400 billion on AI capex this year. That number gets debated and dunked on by people who think it is too high. Now go look at what those four companies generate in operating cash flow. It is a rounding error for them. They could double capex without breaking a sweat.

And here is the part nobody talks about. The marginal return on AI capex is rising, not falling. Inference workloads are exploding. Every new model release means more queries, more tokens, more compute. The unit economics get better as scale rises. None of that screams "this is the peak."

Country level GPU buildouts are not a one quarter story. They are a five to ten year story. Every developed nation has now figured out that compute is the new oil, and they want a domestic supply. That means recurring, multi billion dollar NVIDIA orders from buyers who do not care about Q4 earnings calls. Saudi Arabia signed for tens of billions. UAE same. France, UK, India, Japan, all coming. This is structural demand that did not exist in the 2024 model. Wall Street has not priced it.

I am still long because the multiple is not insane. NVIDIA trades around 30 times forward earnings while growing the top line at 85% and the bottom line faster. Compare that to Costco at 50 times for 6% growth, or Walmart at 38 times for 5%. The bears compare NVIDIA to the dot com top. The math says NVIDIA is cheaper relative to growth than half the boring grocery stocks in the index.

The other reason I am long is optionality. Robotics. Autonomous vehicles. Drug discovery. The next leg of compute demand is not just chatbots. It is physical world AI. And every single one of those use cases runs on NVIDIA today.

I am not a permabull. The risks I take seriously are these. A real recession that forces enterprise to cut AI budgets. A major model release that does more with less compute, the DeepSeek shock 2.0. A regulatory action in China that takes a quarter or two off the comp. An accounting scandal at one of the neoclouds that spooks the supply chain.

None of those break the long term thesis. All of them could cause a 20 to 30% drawdown. Position size accordingly. Do not bet the farm on a single name no matter how good the story is. Read Position Sizing in a High-Valuation World before you size up.

NVIDIA at $81.6 billion in a quarter is not a top. It is a checkpoint. The infrastructure cycle has multiple years left, the moat is widening, sovereign demand is real, and the multiple is reasonable relative to growth. I am still long. I am not adding at every rip. I do scale in on weakness. The conviction is built on the numbers, not the noise.

Read next: Cognition Just Hit $26 Billion | Zuckerberg Just Bet the House on AI

*⚠️ Disclaimer: This post is for educational and entertainment purposes only. MentorSurge is not a financial advisor. Nothing on this site is investment advice. NVIDIA is a high volatility stock and past performance is not indicative of future results. Always do your own research. Always consult licensed professionals before making decisions with real money.*