The Downside Risks Every Nvidia and AMD Investor Must Know

TL;DR Nvidia faces customer concentration risk (Microsoft, Meta, Amazon, Google), $50B in lost China market access, and customer-developed custom chips. AMD trades at 148x earnings where everything must go perfectly, with historical 32% drawdowns during market shocks. Both stocks are priced for flawless execution.

Why the Bear Case Matters More Than the Bull Case

50% of data centers worldwide have been stopped or delayed.

That single data point should make any AI semiconductor investor pause. The bull cases for Nvidia and AMD are well documented and widely circulated. But in my experience, what separates successful investors from the rest isn't knowing why things could go right — it's understanding exactly how they could go wrong. Here are the specific risks I'm watching for both companies.

Nvidia Risk #1: Dangerous Customer Concentration

Nvidia's revenue depends on a handful of massive customers: Microsoft, Meta, Amazon, and Google.

This level of concentration represents extraordinary dependence on a very small number of relationships. If any of these companies slows AI spending — even temporarily — Nvidia feels it immediately and everywhere. When your quarterly revenue moves in tens of billions, a single customer's purchasing decision can reshape your earnings report overnight.

Nvidia Risk #2: The AI Monetization Question

Here's the uncomfortable question the bears keep asking: has AI actually generated enough real revenue to justify the hundreds of billions being spent on chips?

Right now, spending is happening at a pace that assumes AI will become one of the most profitable technologies in human history. If that timeline stretches longer than expected, if companies decide to pause and digest what they've already purchased, demand for Nvidia chips could slow sharply. A stock priced for perfection doesn't handle slowdowns well.

The data center construction delays aren't theoretical. Half of global data center projects have reportedly been stopped or delayed. That's not a small number.

Nvidia Risk #3: China and Custom Chips

US export restrictions have already cut Nvidia off from the Chinese market. Jensen Huang himself estimated this represents roughly a $50 billion opportunity they can no longer fully serve. If restrictions tighten further, the impact grows.

Simultaneously, Google, Amazon, and others are building custom chips to reduce their dependence on Nvidia. The bear case for Nvidia isn't that the company fails — it's that growth slows just enough to make a very expensive stock look overpriced.

Consider Adobe as a precedent. Adobe hasn't contracted. Its growth merely decelerated. The stock dropped from an all-time high near $700 to around $250. Growth slowdowns in premium-valued stocks can be devastating.

AMD Risk #1: The Weight of 148x Earnings

AMD trades at 148 times earnings. The implication is stark.

Everything has to go right. Not mostly right. One hundred percent right, if not better. At these multiples, there's virtually no margin for error, no room for a single quarter of disappointing results, no tolerance for execution delays.

AMD Risk #2: Historical Volatility

AMD is one of the most volatile stocks in the market.

During downturns, it has historically dropped an average of 32% during market shocks — sometimes losing over 30% in less than two months. This volatility becomes particularly dangerous when combined with a 148x P/E ratio. Once a selloff begins, multiple compression can accelerate the decline in a self-reinforcing spiral.

AMD Risk #3: Everything Must Align Simultaneously

AMD's position is more precarious than most investors realize.

The MI 450 has to ship on time. Major customer deployments with Meta and OpenAI have to perform as expected. The overall AI spending environment has to remain robust. All three conditions must hold simultaneously. Betting on three major assumptions going perfectly for a stock at 148x earnings requires substantial conviction — or perhaps just insufficient awareness of the downside.

What This Means for Investors

My point isn't that either company should be sold. Both Nvidia and AMD are exceptional businesses with genuine competitive advantages.

But understanding the bear case is just as important as understanding the bull case. Knowing your risks and accepting them is fundamentally different from being unaware of them. The same stock, at the same price, becomes a completely different investment depending on whether you've accounted for what could go wrong.

Both stocks are currently priced for flawless execution. If you're considering a position, first ask whether your portfolio can absorb the downside if the plan doesn't unfold perfectly.

FAQ

Q: How real is the threat from customers building their own chips?

A: It's real in the medium term. Google's TPU and Amazon's Trainium are already deployed in production workloads. However, these chips are optimized for specific tasks and can't fully replace general-purpose GPUs. They'll erode Nvidia's market share gradually rather than displace it overnight.

Q: What would happen to AMD's stock if AI spending slows?

A: At 148x earnings with an average historical drawdown of 32% during market shocks, a meaningful slowdown in AI spending could trigger a severe correction. AMD's current valuation assumes aggressive, uninterrupted growth — any disruption to that trajectory would likely cause outsized price declines.