Customer Concentration Risk: When One Buyer Owns You

Nvidia (NVDA) booked staggering AI revenue last year — and a handful of customers wrote most of the checks. When a few buyers own your top line, one budget cut can rewrite the entire investment thesis.

What Is Customer Concentration Risk?

Customer concentration risk is the danger that comes from depending on a small number of customers for a large share of revenue. If your biggest buyer leaves, slashes orders, or simply demands a lower price, your earnings can crater overnight.

It is one of the most under-appreciated risks in fundamental analysis, because it rarely shows up in the headline numbers. A company posting roughly 40% revenue growth looks unstoppable — until you learn that one customer drove most of it.

Concentration does not change what a company earns; it changes how reliably it earns it — and reliability is exactly what the market pays a premium multiple for. Two firms with identical revenue can deserve very different valuations if one is diversified and the other is not.

How Do You Spot It in a 10-K?

Go straight to the risk factors and the revenue notes. U.S. companies are required to disclose any single customer that accounts for more than roughly 10% of total revenue, so the filing usually names the dependence outright.

Look for language like "one customer represented approximately 16% of revenue" or "our ten largest customers accounted for the majority of sales." The management discussion section often quantifies it further by segment.

For names deep in the AI supply chain — Broadcom (AVGO), Arista Networks (ANET), and Advanced Micro Devices (AMD) — these disclosures are where the real risk lives. If a company buries a customer-concentration risk factor while celebrating blistering growth in its press release, the filing is telling you the truth the marketing is not.

Where Concentration Bites Hardest

The AI buildout has created some of the most concentrated revenue structures in modern tech. A few cash-rich hyperscalers and AI labs are funding enormous order books — wonderful while it lasts, dangerous if any single buyer pulls back.

Oracle (ORCL) is the cleanest current example: reports suggest roughly 47% of its cloud backlog traces to a single AI customer. Qualcomm (QCOM) shows the flip side — for years its single largest customer worked to design Qualcomm's modems out of its phones, a slow-motion concentration unwind that pressured the stock for years.

Is Concentration Always Bad?

No — and this is where most investors get it wrong. A large, contractually committed anchor customer can be a genuine competitive advantage: it underwrites investment, signals product quality, and can be brutally hard for rivals to dislodge.

The question is the nature of the dependence. A multi-year contract with switching costs and a financially strong counterparty is very different from an informal relationship with a customer that could leave at the next renewal.

The right question is never "how concentrated is it?" but "how durable and how mutual is the dependence?" — a sticky two-way relationship is a moat, a fragile one-way one is a time bomb. This distinction is the heart of sound investment strategy: risk is not the enemy, mispriced risk is.

The Mistakes Investors Make With Concentration

The first mistake is celebrating concentrated growth without discounting it. Explosive revenue from one customer should be valued at a lower multiple than the same revenue spread across thousands — yet the market often pays up for the growth and ignores the fragility.

The second is assuming today's anchor is permanent. Customers in-source, switch vendors, or hit their own troubles. Apple's gradual move to design out QCOM modems is the canonical lesson that even a marquee customer can become an exit.

The third is forgetting the financial health of the buyer. A backlog is only as good as the counterparty's ability to pay — which is why an order book anchored by a still-unprofitable customer deserves more scrutiny, not less.

Pro Tips for Pricing Concentration Risk

Quantify it before you value it. Pull the percentage of revenue from the top customer and top ten, and ask what the income statement looks like if the largest one disappears. If that scenario breaks the thesis, size the position accordingly.

Next, weigh contract structure over relationship warmth. Length, exclusivity, switching costs, and take-or-pay terms tell you far more than a glowing earnings-call anecdote about a partnership.

Finally, watch for diversification as a catalyst. A concentrated company that steadily broadens its customer base should see its risk discount shrink — and that re-rating, not the next revenue beat, is often where the real return comes from. The same dynamic is playing out across the AI infrastructure names we track in our market commentary.

When Should You Walk Away?

When the dependence is one-directional, fragile, and unpriced. If a single customer can leave at will, drives the bulk of profit, and the stock trades as if that revenue were diversified and permanent, the risk-reward is stacked against you.

Be especially cautious when the anchor customer is itself financially shaky, because then you are layering counterparty risk on top of concentration risk. No growth rate is worth owning if a single phone call from one customer can erase a third of it — at that point you are not investing, you are underwriting someone else's budget. Knowing when a great story hides an unacceptable dependence is what separates disciplined investors from momentum chasers.

Ready to analyze these stocks yourself? Search any ticker on MainRatios to see valuations from 6 legendary investors - free.