NVIDIA’s dividend hike and buyback reveal a new question: can the ultimate growth stock become a value stock without losing its edge?
KEY TAKEAWAYS
NVIDIA is still a growth company, but its massive dividend increase and buyback authorization make it look more like a value stock than ever before. The company is generating more cash than it can productively reinvest fast enough.
The dividend hike is not generosity. It is evidence that NVIDIA’s cash generation has become almost absurd.
Jensen Huang is not simply returning cash to shareholders. He is also using NVIDIA’s balance sheet to invest across the AI infrastructure ecosystem.
The Cisco comparison matters because real technology and real demand do not prevent overbuilding. Hypergrowth always eventually slows.
The strongest argument in NVIDIA’s favor is that Jensen appears to understand the transition before the market forces it on him. That discipline may separate NVIDIA from past hypergrowth casualties.
MY HOT TAKES
NVIDIA is no longer easy to label. It is acting like a growth stock, a value stock, and a strategic investment fund all at once.
The dividend increase is the tell. Companies do not raise dividends that dramatically unless they know the cash machine has entered a new phase.
The vendor-financing concern is fair, but it is not automatically fatal. NVIDIA is funding an ecosystem from a position of strength, not desperation.
Cisco is the warning label, not the base case. The technology can be real and still disappoint investors if expectations get too far ahead of normalized demand.
Jensen Huang’s real test is not building the fastest plane. It is landing it without turning shareholders into nervous passengers clutching the armrests.
You can quote me: “Jensen Huang is essentially running a sovereign wealth fund out of chip profits.”
Cash machine. Just two nights ago, I was giving a finance lecture to a group of young students and I was explaining the differences between growth and value stocks. I started with what I learned as a Freshman economics student back in the early 1980s (ouch–that was a minute ago). “Growth stocks,” I asserted “are stocks from companies that are expected to have continued, outscaled earnings growth.” I continued, “these companies do not typically pay dividends because they re-invest all those great earnings into their fabulous growth initiatives.” Then I asked the class, “who can give me an example of a popular growth stock?” I heard a few Apples, at least four or five Googles (er, Alphabets, if you want to get technical), but the overwhelming number of declarations went to NVIDIA. I did hear a “Bitcoin,” but ignored it, as I will for this morning’s post–but clearly worth a bigger discussion on another day.
All those companies are indeed superlative choices of what we would consider growth companies. But there was one problem. All of those three companies pay dividends! I had to think on my feet. Were there any better examples of companies that had displayed double digit earnings growth with continued expected future growth? I could come up with a few, but they were either too obscure for these neophytes, or…well, they paid dividends too.
NVIDIA, Apple, and Alphabet all achieve those great growth trajectories by being technology innovators. Technology innovation takes lots of capital. To remain in the earnings growth fast lane and to expand their protective moats, it would seem logical for those companies to invest every available dollar back into that growth engine. Wouldn’t it?
I tried to explain but kept finding myself describing these three classic growth companies using “value stock” attributes. I thought to myself that I had better sharpen up the language on this before the next lecture.
Not a day passed since that lecture, and NVIDIA dropped its Q1 earnings–last night. I watched carefully as the numbers streamed out on Bloomberg. Blah, blah, blah–exactly as I expected, but probably not good enough for the hungry masses–same as usual, also as I expected. And then it came–a dividend boost. Not just a boost, a 20x dividend boost. That was not as I expected. I was surprised–as was the street. What would I tell my students at the next lecture?
Here is what I would tell them. When a company raises its dividend twenty-five times in a single move–from a penny a share to twenty five cents–and simultaneously authorizes an additional $80 billion in share buybacks on top of the $38.5 billion it already had sitting in the authorization, that is not generosity. That is a confession. It is a CEO standing at the podium, staring at a cash waterfall so overwhelming that he has literally run out of productive places to put it fast enough. NVIDIA generated $50 billion in net cash from operations in a single quarter. It returned $20 billion to shareholders in Q1 alone. The company is not struggling to grow. It is struggling to deploy the fruits of that growth faster than they accumulate. That is a remarkable problem to have. It is also the precise moment a company crosses an invisible line–from hypergrowth engine to something else entirely.
So what do you call a company that generates more cash than it can reinvest, raises its dividend dramatically, and authorizes massive capital returns? On Wall Street, we call that a value stock. Which means my students were right about NVIDIA, they just described it about eighteen months too early. The question now is whether Jensen Huang knows exactly what he is doing, or whether he is wandering into territory that has humbled greater CEOs before him.
To his credit, Jensen is not simply handing money back and calling it a day. He has a second answer to the cash problem, and it is audacious. In 2026 alone, NVIDIA has committed over $40 billion in equity investments across the AI infrastructure stack. The centerpiece of that cash investment bonanza is a $30 billion stake in OpenAI, the largest single equity deal in NVIDIA’s history. Beyond that, you have investments in Corning for optical fiber manufacturing, in data center operator IREN, in silicon photonics specialists. Last year, NVIDIA generated $97 billion in free cash flow. Jensen is essentially running a sovereign wealth fund out of chip profits, and so far the returns are staggering. The company’s $5 billion bet on Intel has already grown to over $25 billion in value (for now 😰), one of the fastest returns of that magnitude in American corporate history.
The cynical read, and Wall Street is already whispering it, is that this looks a lot like vendor financing. The argument goes that NVIDIA is investing in companies that turn around and use that capital to buy NVIDIA’s chips, creating a loop that flatters revenue without necessarily reflecting organic end demand. It is a fair concern to raise. But it misses something important. NVIDIA is not a cash-strapped supplier stretching its balance sheet to keep customers afloat. It is the most profitable technology company on the planet, deploying surplus capital into an ecosystem it helped create, and collecting extraordinary returns along the way. There is a meaningful difference between those two things, even if the structure looks similar from a distance.
And yet the history of Wall Street is littered with companies that looked different from a distance, right up until…well, they didn’t. Which brings me to the question I have been turning over since last night’s call. Between 1995 and 2000, Cisco’s revenue rocketed from $2 billion to $19 billion. At its peak in March of 2000, it briefly became the most valuable company in the world. The technology was real. The routers were real. The internet was real. The demand was real. And then it wasn’t–at least not at the pace Cisco had built its entire financial architecture around. The company wrote off $2.25 billion in inventory in a single quarter and posted a loss of nearly $3 billion. Its stock still has not returned to that peak, more than twenty five years later. John Chambers, a brilliant operator by any measure, later admitted he had never built models to anticipate what a sharp demand normalization would look like. He treated hypergrowth as a permanent condition rather than a phase.
To be really, REALLY clear, I am not calling NVIDIA a bubble. The AI infrastructure buildout is real. Agentic AI–the phase Jensen described last night when he said demand has gone parabolic–represents a genuine structural shift in how much compute the world needs on a continuous basis. Agents do not run in training bursts. They run constantly, which means GPU utilization becomes baseline infrastructure cost, not a periodic capital event. That is a fundamentally different demand profile than anything Cisco ever enjoyed. But the Cisco question is not really about whether the technology is real. It is about whether the CEO running the most cash-generative company in history has built something that survives the day the growth rate steps down from parabolic to merely extraordinary. That day always comes. The only variable is whether you saw it coming.
Here is why I think Jensen has seen it coming. The dividend hike and the buyback authorization are not the moves of a CEO who believes the music plays forever. They are the moves of someone who understands, perhaps better than anyone on the street, that the hypergrowth chapter has a last page, and who may be engineering the transition before the market forces his hand. Chambers never accepted that Cisco was becoming a value company. Jensen, by cutting that twenty five times dividend increase and stacking $80 billion in buyback authorization, may be the first hypergrowth CEO in a generation to embrace that transition on his own terms, on his own timeline, while the numbers are still extraordinary.
There is more evidence in the details. NVIDIA guided Q2 revenue at $91 billion, and that guidance assumes zero data center revenue from China, a market Jensen himself has called a $50 billion annual opportunity. The company shipped no Hopper chips to China in Q1, compared to $4.5 billion in the same quarter a year ago. That is not optimistic guidance dressed up with a bow. That is a conservative floor with a potential upside catalyst sitting just offshore, waiting on geopolitics. Meanwhile, the Vera CPU, which is NVIDIA’s first chip purpose-built for agentic AI, opens what Jensen described as a brand new $200 billion addressable market. The platform is not standing still while Jensen figures out what to do with the cash. It is actually accelerating.
So here is what I will tell my students at the next lecture. NVIDIA is still a growth company. It is also, for the first time, behaving like a value company. And in the rarest of cases–Microsoft under Satya Nadella comes to mind–a company can be both simultaneously, if the person at the top is disciplined enough to manage the handoff. Jensen Huang closed last night’s earnings call with five words: “demand has gone parabolic.” He is right. The numbers prove it. But the most important question for long-term investors is not whether demand is parabolic today. It is whether the man running the most cash-generative enterprise in the history of American business has the judgment to land this plane when the tailwinds ease. Based on everything we saw last night–the conservative guidance, the ecosystem investments, the product cadence, the capital return discipline–I think the odds are in his favor. NVIDIA is, indeed, an incredible company, but I am watching Jensen Huang very, very carefully. The soft landing is the hardest maneuver in aviation. And on Wall Street, nobody gives you extra credit just because the flight was spectacular. The landing is everything.
YESTERDAY’S MARKETS
Equities rallied yesterday with the S&P 500 up 1.0%, the Dow up 1.2%, and the Nasdaq up 1.5%, as Iran deal optimism pulled Brent crude down more than 5% to around $105 per barrel following reports that three supertankers transited the Strait of Hormuz. The 10-year Treasury yield eased slightly to 4.65% after touching a 16-month high of 4.67% the prior session, as bond markets continued to price in the inflationary overhang from elevated energy prices.
NEXT UP
Initial Jobless Claims (May 16th) is expected to come in at 210k, close to last week’s 211k claims.
Housing starts (April) may have slipped by -5.3% after jumping by 10.8% in March.
Building Permits (April) probably rose by 2.5% after sinking by -11.4% in the prior period.
S&P Global Flash Manufacturing PMI (May) may have slipped to 53.8 from 54.5
S&P Global Flash Services PMI (May) is expected to have inched up to 51.2 from 51.0.
Important earnings today: Williams-Sonoma, Advanced Auto Parts, Deere, Walmart, Ralph Lauren, Workday, Zoom, Deckers Outdoors, Take-Two Interactive, and Ross Stores.