The Peak Margin Canard

There’s a phrase that gets thrown around Wall Street whenever a great growth company dares to invest in its own future: “Peak margins. Story over. Nothing else to see here.” 

It’s a comfortable conclusion. Read what’s visible, react to what’s measurable, call it a thesis. The problem is that the real answers aren’t on the shoreline. They’re out in the ocean. And most of Wall Street never goes there. 

The Shallow Trade 

Let’s call this what it is: a surface-level, directional trade dressed up as financial analysis. 

Margins matter. They are one of the two levers that grow earnings, and when compression arrives, one cylinder is no longer pulling its weight. That’s a legitimate warning signal — and it deserves to be treated as one. The mistake isn’t noticing it. The mistake is treating a warning as a verdict. The right response to margin compression is investigation, not reaction. The shoreline reader sees the warning and sells. The ocean reader sees the warning and asks why. 

The math makes the case better than any argument can. 

Note: Above is a hypothetical example for illustration purposes. 

Revenue doubled. Gross profit grew 67%. You just told your clients to sell that $2 billion of incremental profit — because the margin percentage went down. 

That’s what happens when you mistake the shoreline for the sea. 

The Coiled Spring 

When a great company invests aggressively — infrastructure, new products, market expansion — margins compress on purpose. But compression can also signal pricing power erosion, unfavorable mix shift, or competitive pressure. The cause matters as much as the fact. Investment-driven compression in a confirmed growth cycle is a very different animal than structural deterioration — and conflating the two is where the real analytical failure lives. The market anchors on the narrative, gets confused, and the stock drifts sideways. And while it drifts, the cash flow engine keeps building. 

I call this the Coiled Spring. But understanding why it forms requires looking at all three dimensions at once. At the structural level, the long wave is intact — the secular tailwind hasn’t broken. At the cyclical level, capital is flowing, the industry is in expansion mode, the medium wave is confirming. And at the company level, the business is doing exactly what a great company should do when it sees a generational opportunity: loading up, investing, deliberately absorbing margin to own the next phase of growth. 

All three waves aligned. Fundamentals resonating. But the market can’t see it because it’s fixated on one line item that is temporarily, intentionally lower. Pent-up alpha accumulates beneath the surface — until consensus finally catches up to what the full picture has been saying all along. And when it does, the spring releases. Fast. 

Amazon: The Ocean Was Always There 

For over a decade, Jeff Bezos invested at the expense of margins without apology. Earnings were negligible. Margins were thin. The peak margin crowd called it a broken business model — repeatedly. 

Except there was an ocean hiding inside the compression. AWS — one of the most powerful enterprise businesses ever built — was scaling quietly beneath the surface while shoreline readers were busy writing Amazon off. When it finally emerged, margins expanded violently and the stock became a generational compounder. 

Margin compression caused by deliberate, high-conviction investment in a confirmed opportunity is not the same as margin deterioration. Conflating the two is the canard. 

NVDA: Nothing to See Here? 

The peak margin narrative on NVIDIA had real ammunition — Blackwell ramp costs, custom ASIC pressure, export restrictions. Gross margins did compress from their 2024 peaks. But a company generating $44 billion in a single quarter at “compressed” margins is still an unprecedented cash flow machine. The underlying wave never broke. 

In The Power of Inference,¹  we noted that over half of NVIDIA’s engineers work on software, not chips — a structural competitive signal hiding in plain sight, invisible to anyone reading only the margin line. That’s the ocean. The peak margin crowd never got there. They stopped at the gross margin line, called it a thesis, and missed one of the defining growth stories of the decade. 

Peak margin is not peak everything. It never was. 

When It’s Actually Right 

The argument isn’t always wrong. It has genuine teeth when revenue is decelerating alongside the margins — when the company is losing pricing power, not investing in future growth. When compression is structural: commoditization, competitive disruption, a model approaching obsolescence. When revenue growth quality is deteriorating. When the industry cycle itself has peaked. 

One cylinder down, and the engine itself losing power. That’s a different story. Sell it. 

But you have to do the work to know which situation you’re in. The answer doesn’t live on the shoreline. 

The Only Question That Matters 

When you see margin compression in a high-growth company, one question governs everything: Is revenue growth intact — and is it durable? 

If yes, you don’t have a deteriorating business. You have a cash flow machine in investment mode, temporarily obscured by a narrative that rewards surface-level thinking. The stock trades sideways for a while. That’s fine. That’s the spring loading. 

If no — revenue decelerating, growth quality eroding — the canard has teeth. 

The peak margin canard has been costing investors alpha for decades. It’s always seductive, always sounds smart, and is often dead wrong — because it asks a shoreline question and never looks out at the sea. 

Don’t fall for it. 

Footnote
¹  Mark Scalzo, “The Power Of Inference: How New Benchmarking Helps Further Reveal NVIDIA’S AI INFLECTION“, Validex.co, October 17, 2025.