Why We Are Shorting NVIDIA (NVDA)

By Mickelberry Capital

Let’s be clear from the start:

Shorting a stock is not about being contrarian for attention.
It’s about recognizing when price has detached from reality — and managing risk accordingly.

At Mickelberry Capital, we don’t short often. In fact, we avoid it unless conditions become extreme. NVIDIA (NVDA) is one of those rare cases where valuation, narrative, and market psychology have become dangerously misaligned.

This post explains why we are short NVDA, how we’re thinking about risk, and what investors should understand before blindly following momentum.

First: This Is Not a Call to “Bet Against AI”

Artificial intelligence is real.
Its impact will be long-lasting.
And NVIDIA is a legitimately strong company.

But great companies can still be terrible investments at the wrong price.

Our short thesis is not anti-AI.
It’s anti-paying any price for growth.

History is full of innovations that changed the world while destroying capital for investors who arrived too late.

The Core Issue: Extreme Valuation Disconnect

At current levels, NVIDIA is priced as if:

Growth will remain exponential indefinitely
Competition will be irrelevant
Margins will stay historically elevated
Capital intensity will not matter
Cycles no longer exist

That combination has never held true long term — in any industry.

From a valuation standpoint, NVDA is trading hundreds of percent above what we would consider fair value using conservative asset-based and cash-flow-based frameworks.

Even when adjusting for future growth, the margin of safety is gone.

When there is no margin of safety, risk dominates reward.

Narrative Saturation Is a Warning Signal

When a stock becomes synonymous with a single narrative — in this case, “AI equals NVDA” — discipline disappears.

Common warning signs we’re seeing:

Retail participation surging late in the cycle
Media framing NVDA as “untouchable”
Valuation concerns dismissed outright
Every dip treated as a guaranteed buying opportunity

Markets don’t break when fear is high.
They break when confidence becomes unanimous.

Capital Expenditure Reality Is Being Ignored

One point raised publicly by Michael Burry — and one we also flagged internally — is the way AI infrastructure is being depreciated and accounted for.

Even if longer depreciation lives are justified by durability, the capital intensity of AI hardware is enormous.

Key questions investors aren’t asking enough:

What happens when hyperscaler demand slows?
How many customers can sustain this level of capex?
What does normalized demand look like after the build-out phase?

These questions don’t need pessimistic answers — only realistic ones — to challenge current pricing.

Competition Will Matter — Eventually

NVIDIA is dominant today.

That does not mean:

AMD disappears
Custom silicon doesn’t scale
Margins remain untouched
Pricing power goes unchallenged

In capital-heavy industries, excess returns attract competition. That’s not an opinion — it’s economic gravity.

Markets are pricing NVDA as if this gravity has been suspended.

We disagree.

How We’re Approaching the Short (Risk Matters)

Shorting is inherently asymmetric — risk management matters more than being “right.”

Our approach:

Position size limited to ~5–10% of short-term capital
Structured exposure (not blind naked shorts)
Willingness to be early — but not reckless
Constant reassessment as conditions change

This is not a long-term “bet against America.”
It’s a valuation-based, cycle-aware trade.

What This Says About Our Broader Philosophy

At Mickelberry Capital, we:

Favor long-term ownership when valuation allows
Prefer boring, cash-generating businesses
Use tools like cash-secured puts for asymmetric safety
Avoid chasing narratives at peak enthusiasm

Shorting NVDA fits within that framework — not outside it.

We become cautious when:

Everyone agrees
Risk is dismissed
Price replaces analysis

That’s where discipline matters most.

Final Thought

NVIDIA may continue higher in the short term.
Markets can stay irrational longer than expected.

But valuation always reasserts itself.

We don’t need NVDA to fail.
We only need expectations to normalize.

And when expectations are this extreme, normalization alone can be enough.

As always:
Do your own research.
Manage your risk.
And remember — the goal is not to be exciting.

The goal is to survive, compound, and stay disciplined.

Subscribe to Mickelberry Capital to follow how we think, allocate, and adapt in real time.