Nvidia, AMD and Microsoft just delivered the kind of AI numbers that end arguments. Nvidia posted $81.6 billion in quarterly revenue, AMD’s data center sales climbed 57% in a year, and Microsoft’s Azure cloud grew about 40%. Analysts rate all three a Buy, with almost no dissent across more than 130 published calls.
Buy the trio and it looks like a diversified slice of the AI build-out. Follow the cash, and the three names start to behave like a single position, priced on the same customers, the same handful of mega-deals, and the same year of spending.
Three Record Prints, One Demand Engine
Start with the scoreboard, because it is genuinely loud. Nvidia’s quarter ended April 26 brought record revenue, up 85% from a year earlier, with data center sales of $75.2 billion. The chipmaker also lifted its quarterly dividend to 25 cents from a single cent and authorized $80 billion in fresh buybacks, the kind of capital return a company makes when cash is overflowing.
AMD ran on the same fuel. First-quarter revenue reached $10.3 billion, up 38%, while the data center segment hit $5.8 billion, up from $3.7 billion a year before. Microsoft, reporting its fiscal third quarter to March 31, booked $82.9 billion in total revenue, and its cloud capacity is still selling out faster than the company can build it.
The ratings line up neatly behind the prints. Across the analyst calls tracked on the three names, sell recommendations are effectively absent, which is itself a signal worth holding onto.
| Company | Latest revenue | AI / data center signal | Analyst calls (Buy / Hold / Sell) |
|---|---|---|---|
| Nvidia (NVDA) | $81.6B, up 85% | Data center $75.2B, up 92% | 51 / 3 / 0 |
| AMD | $10.3B, up 38% | Data center $5.8B, up 57% | 28 / 6 / 0 |
| Microsoft (MSFT) | $82.9B, up 18% | Azure growth near 40% | 39 / 7 / 0 |
Read those figures from the Nvidia first-quarter results filing and the picture is of three companies winning different races. The trouble is that they are running on the same track.
The Money Moves in a Circle
Here is the part the three-stock pitch tends to leave out. The demand driving every one of those prints traces back to a small cluster of buyers, and the sellers are increasingly helping to finance the buyers.
- Nvidia said in September 2025 it intends to invest up to $100 billion in OpenAI as the model builder deploys 10 gigawatts (GW, a measure of data center power capacity) of Nvidia systems. The capital goes in as the chips go out.
- AMD struck its own pact with OpenAI in October 2025, agreeing to power 6 GW of capacity and issuing the customer a warrant for up to 160 million AMD shares that vests as orders scale. The buyer can become one of the seller’s largest shareholders.
- Microsoft owns roughly 27% of OpenAI and remains a primary cloud host through Azure. The Azure revenue that fund managers cheer is partly OpenAI compute spending, which in turn pays for Nvidia chips.
Nvidia’s reach does not stop at OpenAI. The chipmaker has also backed other model builders, including the xAI funding round that pulled in $20 billion with Nvidia among the backers. None of this is hidden and none of it is improper. It does mean the revenue each company books is partly underwritten by capital the same companies are putting back in.
Nvidia Still Owns the Rails
For now, the lead is not close. Nvidia still supplies roughly four-fifths of the world’s AI accelerators, and the company’s grip extends well past the silicon itself.
The less-watched number this quarter was networking. Data center networking revenue hit $14.8 billion, up 199% from a year earlier, as buyers bought the switches and interconnects that lash thousands of GPUs (graphics processing units, the chips that run AI training) into a single machine. That hardware, plus the CUDA (Compute Unified Device Architecture, Nvidia’s proprietary software layer), creates switching costs that keep customers from walking.
The scale is hard to overstate. Nvidia became the first company to cross a $5 trillion market value on this wave, and it is now pushing into adjacent markets, including its first Windows PC processors built on Arm.
The risk sits inside the same balance sheet. An earlier filing flagged that two large customers made up close to 40% of data center revenue. Dominance and dependence are riding in the same car.
AMD’s Opening, and Its Price
AMD is the genuine challenger, but it is buying its way in. The company holds an estimated 5% to 7% of the AI accelerator market, and it guided second-quarter revenue to about $11.2 billion as its Instinct line gains ground.
The cost of momentum shows up in the structure of its biggest win. The 6-gigawatt OpenAI partnership terms trade scale for ownership, and the math cuts both ways:
- The prize: a potential $90 billion in cumulative hardware revenue across multiple chip generations.
- The chip: the first 1 GW of MI450 GPUs is slated to deploy in the second half of 2026.
- The bill: a warrant covering up to 160 million shares, vesting as OpenAI buys more and as AMD’s stock clears price targets that escalate toward $600 a share.
That is the second-order trade laid bare. AMD’s data center figures, confirmed in its first-quarter earnings disclosure, look better partly because it handed a key customer a path to a large equity stake.
Microsoft Is the Cash That Keeps It Spinning
Microsoft is the steadiest leg of the stool, and the one most people misjudge. It does not need AI to fund itself, which is exactly why it can keep funding everyone else.
Capital spending tells the story. Microsoft poured $31.9 billion into property and equipment in the quarter and signaled the figure climbs above $40 billion next quarter as more capacity comes online. Its fiscal third-quarter financial statement shows operating cash flow large enough to absorb that without strain.
The company also reset its OpenAI arrangement, ending revenue-share payments and allowing any cloud provider to serve OpenAI models. That loosens one link in the chain, yet Microsoft’s roughly 27% stake means it still rises and falls with the same customer that anchors Nvidia’s and AMD’s growth.
Why the Basket Behaves Like One Position
Add it up and the diversification is mostly cosmetic. A position in all three is a leveraged bet that hyperscaler capital spending keeps climbing and that OpenAI and a few peers keep absorbing the output.
- ~80% of AI accelerators shipped still come from Nvidia, the trio’s shared single point of failure.
- ~$630 billion in projected hyperscaler infrastructure spending for 2026 is the demand the whole bull case rests on.
- Zero sell ratings across the roughly 134 analyst calls on the three stocks, a level of agreement that leaves no cushion if sentiment turns.
The wobble already has a name attached. Colette Kress, Nvidia’s finance chief, told investors in December 2025 that the headline OpenAI framework was still a letter of intent rather than a signed, definitive agreement. If the deals that justify the spending slip, the prints that justify the ratings slip with them, and they slip together.
Frequently Asked Questions
Are Nvidia, AMD and Microsoft Good AI Stocks to Buy Now?
Analysts overwhelmingly say buy: across the three, there are effectively no sell ratings and Nvidia’s price targets cluster between $245 and $300. The caveat is correlation. Because the three feed off the same customers and capex cycle, holding all three concentrates rather than spreads AI risk.
What Does Circular Financing Mean in AI?
Circular financing is when a vendor helps fund the customer that buys its products. Nvidia intends to invest up to $100 billion in OpenAI, AMD granted OpenAI warrants tied to chip orders, and Microsoft holds an OpenAI stake while hosting its compute. Notably, Nvidia’s own finance chief said in late 2025 the $100 billion plan was not yet a definitive agreement.
How Much of the AI Chip Market Does AMD Hold Versus Nvidia?
Nvidia ships roughly 80% of AI accelerators, while AMD holds an estimated 5% to 7%. AMD is gaining share, with data center revenue up 57% to $5.8 billion last quarter, but the gap remains wide and Nvidia’s networking and software lock-in widen its moat further.
When Does AMD Start Deploying Chips for OpenAI?
The first 1 gigawatt of AMD Instinct MI450 GPUs is scheduled to deploy in the second half of 2026, the opening tranche of a 6-gigawatt agreement. The associated warrant vests in stages, with the final tranche tied to AMD’s share price reaching $600.
What Is the Biggest Risk to the AI Stock Trade?
The biggest risk is reflexivity. Hyperscalers are projected to spend around $630 billion on infrastructure in 2026, and Microsoft alone guides to more than $40 billion of capex in a single coming quarter. If end demand or model-builder revenue disappoints, the financing loops between chipmakers, clouds and OpenAI can unwind quickly, pulling all three stocks down at once.
Disclaimer: This article is for informational purposes only and is not investment advice. Equities, including AI and semiconductor stocks, carry significant risk, and past performance does not guarantee future results. Readers should consult a qualified financial professional before making investment decisions. All figures are accurate as of publication on May 31, 2026.
