When the Winner Stops Competing (And Starts Acquiring)

There's a moment in every market cycle that announces itself quietly, without fanfare. Nvidia's $20 billion licensing deal with Groq—announced just before Christmas, when most traders were mentally already gone—is that moment.

I don't mean it's shocking. Nvidia buying up potential rivals is as newsworthy as snow in December. But the structure of this deal, the language they chose, the way it was framed to avoid saying "acquisition"—this tells you something about where we are.

Groq was supposed to be different. Founded by the very people who created Google's TPU, it was built on an undeniable technological thesis: that future AI wasn't about raw GPU horsepower, but about specialized chips optimized for inference—faster, cheaper, less power-hungry. The kind of underdog story that makes capitalism feel real for a minute. The kind of thing that makes you believe competition still works.

Three months ago, Groq raised $750 million at a $6.9 billion valuation. Blackrock invested. Samsung invested. Donald Trump Jr.'s firm backed it. Real capital, real convictions. In September, when they closed that round, Groq had momentum.

Then Nvidia called.

The Polite Extinction

Here's what we tell ourselves: Nvidia licensed Groq's inference technology. Hired some brilliant engineers. Let Groq continue as an "independent company" under new management. Everyone wins. The market gets smarter chips faster. Competition thrives.

This is fiction, and we all know it.

Bernstein's analyst put it best—this is "essentially an acquisition without being labeled one," a way to avoid the antitrust glare that's been sharpening on Big Tech. You pay $20 billion. You get the IP. You take the founder and the best people. You integrate the technology into your system. The original company stays on life support as a shell, a regulatory fig leaf. The "non-exclusive" part? It's window dressing. Groq's going to keep operating GroqCloud, sure. But its reason for existing—to be the alternative to Nvidia—is now in Nvidia's pocket.

This isn't new. Cisco did it through the late '90s. Microsoft perfected it in the 2000s. But watching it happen in real time, to a company that genuinely had better technology for a specific problem, is different than reading it in a business school case study.

What we're watching isn't Nvidia flexing its muscles. It's Nvidia eliminating the only real threat in the inference space before that threat could grow teeth.

The Cash That Breaks Competition

Nvidia's sitting on $60.6 billion in cash and short-term investments. They're generating $22 billion a quarter in new cash. For a company at that scale, $20 billion is "82 cents per share," according to Bernstein. It's nothing. It's the cost of not having a competitor.

This is the structural problem we're not talking about. When one company is so far ahead that it can simply buy potential rivals before they become threatening, competition doesn't disappear—it gets absorbed. The would-be competitor doesn't die. It gets hired. Its IP gets licensed internally. Its founder gets a job offer he's not really supposed to refuse.

Markets work when alternatives exist. When alternatives start looking like acquisition targets at Series D, the market stops working. It just looks like it still does.

Amazon did this with cloud infrastructure. Meta did it with social platforms. Microsoft's doing it with AI startups. Each time, the founders walk away rich. Each time, the losing investors (the Series A people, mostly) get okay returns. Each time, the market structure ossifies a little more. And each time, we're told this is healthy competition.

The S&P 500 hit another record high this week. The 87.5% rally since October 2022 continues. Nobody's selling mega-cap growth. Why would they? Nvidia's not just winning—it's making sure nothing else can.

What Valuation Can't See

Meanwhile, bankruptcies hit a 15-year high. Over 700 companies filed for bankruptcy in 2025. Spirit Airlines. Rite Aid, which has now ceased operations entirely. Claire's. Smaller names that nobody talks about because they don't move indices. Companies that couldn't access the $22 billion annual cash flow, that couldn't absorb losses, that faced actual competition.

The S&P 500's CAPE ratio is 39.4. It's only been that high in 25 months since 1957. The chart that follows those moments shows something consistent: the market tends to perform poorly in the years afterward. Sometimes significantly poorly.

We're in a bifurcated economy now, and it's not subtle anymore. On one side: companies with such dominant positions they can acquire threats before they threaten. On the other: everyone else, which includes the actual economy where bankruptcy rates are highest since before the last financial crisis.

The premium stocks keep rising because they deserve it, in a sense. They've won. They've strategically positioned themselves such that they can't lose.

That used to be called a monopoly. Now it's just called "leadership."

Three Days Until 2026

Precious metals are up. Gold's up nearly 70% for the year. Silver's up 150%. Bitcoin's stuck below $87,000, failing to break a range for weeks. The Fed just signaled one rate cut for all of 2026—one—after cutting three times to end this year. Tariffs remain in place. The deficit is still there. The consumer is still there, but more fragile.

And here we are, on the eve of 2026, with a market priced for perfection, held up by a handful of companies that have solved the game: become so large that competition becomes acquisition, and call it innovation.

Groq's going to keep building faster chips. Nvidia's going to integrate them. Groq's going to be a division inside the empire, its founder working for the company that spent $20 billion to ensure he never competed with it again.

This is the market we have. The question isn't whether it's sustainable. The question is what breaks first when it isn't.