On November 9, Nvidia lost approximately $190 billion in market capitalisation following allegations involving its founder and CEO, Jensen Huang.
The company’s board announced an independent investigation into reports of a foreign intelligence effort targeting Huang. The allegations remain unproven, and the board says it has no reason to believe Nvidia secrets or U.S. national security were compromised.
Investors sold. The stock fell 4.3%.
But here’s the more interesting question.
What happens when a company’s reputation deteriorates faster than its customers’ ability to leave?
Because Nvidia isn’t a fashion label or a consumer brand.
You don’t stop buying Nvidia because you dislike its CEO.
You stop buying Nvidia when you can reproduce its economics somewhere else.
And those are two entirely different decisions.
1. The Market Can Leave Overnight. Customers Cannot.
Imagine a hyperscaler with $10 billion committed to AI infrastructure.
Its engineers are trained on CUDA. Its software is optimised for Nvidia hardware. Its data centres, procurement contracts and deployment schedules are built around Nvidia’s architecture.
Then a scandal hits.
Does the company cancel everything and switch to AMD?
Probably not.
Not because reputation doesn’t matter, but because switching could cost more than staying.
The stock market can change its opinion in milliseconds. An enterprise technology architecture can take years to replace.
Financial capital is liquid. Installed infrastructure is not.
That’s the first source of Nvidia’s market power.
Its moat isn’t simply GPU performance. It’s the ecosystem of developers, software, infrastructure and accumulated investment built around its products.
A competitor may offer a chip that’s 30% cheaper.
But if migrating delays a billion-dollar AI deployment by three months, that cheaper chip could become the most expensive decision the customer makes.
Consumer brands compete for preference.
Infrastructure brands compete to become the default.
And defaults are much harder to replace.
2. The Second-Order Economics
The headlines focus on Nvidia’s valuation.
But the more important consequences may emerge elsewhere in the AI economy.
First, competitors receive an opportunity they didn’t earn.
AMD doesn’t need Nvidia’s GPUs to suddenly become worse.
It only needs customers to become uncomfortable with depending entirely on one supplier.
A hyperscaler previously planning to spend 90% of its compute budget on Nvidia might begin evaluating alternative architectures.
Even if it doesn’t switch immediately, it could invest millions in making future migration possible.
The scandal effectively subsidises Nvidia’s competitors by strengthening the business case for diversification.
Nvidia’s biggest risk is encouraging them to finance tomorrow’s alternatives.
Second, reputational risk can become a financing cost.
Consider a hypothetical $10 billion AI data centre project.
At a 7% annual cost of capital, that’s $700 million.
If perceived risk pushes the financing cost to 8%, annual capital costs increase by $100 million on that assumed capital base.
That’s an illustration, not an observed consequence of this controversy.
But the economic mechanism matters.
AI infrastructure depends on enormous amounts of external capital. When uncertainty increases around a critical supplier, lenders and investors may demand stronger guarantees or higher returns.
Nobody needs to cancel an Nvidia order for that risk to travel through the ecosystem.
A reputational shock at the centre can become a financing shock at the edges.
Third, governments can weaken a moat that competitors cannot.
CUDA is difficult to replace through ordinary competition.
But governments can change the economics through export controls, procurement restrictions and national security requirements.
If Chinese customers are pushed toward domestic accelerators, their investment in local software ecosystems may accelerate.
More domestic customers attract more developers. More developers improve the alternatives.
Ironically, restrictions designed to contain technological competition can help create the conditions for a rival ecosystem.
The strongest long-term threat to Nvidia may not be a better GPU.
It may be customers and governments deciding they can no longer afford to depend on one.
3. A Brand Can Lose Affection Without Losing Pricing Power
Nvidia recently reported a 75% gross margin on $96.2 billion in quarterly revenue.
Those numbers don’t prove a monopoly. But they demonstrate an extraordinary ability to capture value from the AI economy.
And here’s the truth.
A company doesn’t necessarily need to be loved to maintain pricing power.
Look at Atlassian.
Customers complain about subscription increases, licensing costs and migration complexity.
Yet organisations continue using Jira and Confluence because those products are deeply embedded in their workflows.
Replacing them isn’t simply buying another subscription.
It’s rebuilding processes, integrations, reporting systems and institutional knowledge.
Microsoft benefits from similar economics in enterprise software. Adobe has historically benefited from them in creative workflows.
Nvidia operates at an even more fundamental infrastructure layer.
But there’s an important distinction.
There are two ways to create customer dependence.
You can make your product so valuable that leaving destroys genuine economic value.
Or you can make leaving so difficult that customers remain even when they’re dissatisfied.
Both produce strong retention numbers.
Only one necessarily produces genuine loyalty.
A customer who stays because you’re indispensable isn’t necessarily a customer who wants to stay.
This is why retention alone is a poor measurement of brand equity.
The better question is whether customers would choose you again if switching were free.
4. The $190 Billion Question
A $190 billion decline in market capitalisation doesn’t mean Nvidia destroyed $190 billion in brand equity.
Market value and market power are different things.
Market value reflects investors’ expectations about future earnings and risk.
Market power reflects a company’s ability to influence the commercial terms under which customers and competitors operate.
One can deteriorate without the other immediately disappearing.
Nvidia could face a serious governance crisis while continuing to sell enormous quantities of GPUs at premium margins.
Both realities can coexist.
But that doesn’t make reputation irrelevant.
The real danger appears when reputational damage changes how customers behave.
When boards revise procurement policies.
When governments accelerate domestic technology programmes.
When competitors receive funding to develop alternatives.
When customers spend money reducing switching costs.
These changes can unfold over years, even while Nvidia’s revenue continues growing.
The moat can be weakening while the financial statements still look spectacular.
5. Brand Equity Is Economic Power
I’ve been thinking about brand equity differently.
Less about recognition, sentiment and popularity.
More about economic power.
Why can one company charge 20% more than another?
Why do customers tolerate annual price increases?
Why do businesses continue buying from a company despite controversy?
The answer isn’t always trust or affection.
Sometimes it’s compatibility.
Sometimes it’s accumulated investment.
Sometimes it’s the cost of starting again.
The strongest brands combine emotional preference with economic dependence.
But those are not the same thing.
Nvidia has built one of the most powerful technological ecosystems in history. A single day’s market decline doesn’t erase that.
Yet its reputation crisis could encourage customers to begin dismantling the very dependencies that make the company so profitable.
And that’s the second-order effect worth watching.
Not whether people still admire Jensen Huang next week.
But whether the companies spending hundreds of billions on AI infrastructure still want to be just as dependent on Nvidia five years from now.
The real threat to Nvidia isn’t that customers suddenly stop buying its GPUs.
It’s that the customers who cannot leave today begin spending money to make sure they can leave tomorrow.
That’s when reputational damage becomes economic damage.
And that’s when a moat actually starts to disappear.
Sources: Nvidia Board of Directors statement (November 8, 2026); Reuters and Associated Press reporting (November 9–10, 2026); Nvidia Q2 FY2027 financial results. Allegations remain unproven. Financial examples are illustrative.