Half your revenue comes from a single customer, but with orders piling up lately, does that dependence actually feel reassuring? As long as your biggest customer keeps placing bigger orders, reliance doesn't look like risk — it's easy to feel like that one account is what's making this year comfortable. Last week, Nvidia showed the opposite instinct. At the exact moment demand was overflowing, it began loosening its concentration across three axes at once — customers, geography, and markets. In other words, it started diversifying away from customer concentration risk at its busiest moment yet.
Orders Are Backed Up, Yet Nvidia Is Leaning Less
Start with the signal that demand is overflowing. Dell's AI server order backlog stands at $60.9 billion — 1.3 times its $47 billion in quarterly revenue. Oracle said on its earnings call that it's running more than 300,000 Nvidia GPUs at 97.9% utilization. OpenAI raised $122.2 billion, formalizing years' worth of future infrastructure spending. It's fair to call this a moment when everything Nvidia makes sells the instant it's made.
Yet the moves that came out the same week looked almost defensive. With an estimated 50% of revenue coming from OpenAI alone and 44% from its top four customers combined, Nvidia struck a deal with Anthropic to spread that exposure, even if only slightly. With 40% of its supply chain concentrated in Taiwan, it partnered with Australia's Megaport, NEXTDC, and AirTrunk to add a data center foothold in a third country. And with a business still leaning heavily on data center GPU supply, it opened a new market in physical AI through a $3.5 billion robotics AI deal with Figure. Laid side by side, here's what those three axes — where Nvidia has been leaning, and which direction it's now pushing — look like.
Concentration hasn't disappeared on any of the three axes, but the direction is clear: Nvidia is adding a new pillar of support next to each one it already leans on.
Why Diversify When Business Is This Good
Nvidia's guidance for 70% growth rests on three assumptions: that its concentration with OpenAI holds steady, that Taiwan's supply chain stays stable, and that Chinese regulation stays at its current level. If all three wobble at once, the forecast falls apart badly. And earlier this month, Nvidia effectively announced it was pulling back from the Chinese market. One of the three assumptions is already shaking.
This is where the business logic shows itself. Diversifying away from customer concentration risk is only possible when you have spare capital and negotiating leverage. Extending financing to a new customer, partnering with infrastructure companies in unfamiliar countries, and tying up capital in a robotics platform aimed at the second half of 2027 all eat into near-term profit. Nvidia's capital allocation — May's dividend increase, a $25 billion bond issuance, and a $99 billion portfolio of investments that includes $30 billion in Intel — is a choice to spend that slack now, while it has it. Go looking for a new customer after the orders dry up, and the other side sets the terms. Go looking while orders are backed up, and you set the terms.
What This Means If You're a One-Person Business
The scale is different, but the structure is the same. If half your revenue comes from one client, a single staffing change, budget cut, or policy shift at that client can shake your entire year. Just as Nvidia laid out its three assumptions, start by writing down what assumptions your own revenue rests on: that the client will keep the same budget next year, that the fees on the platform you rely on won't change, that demand for your current flagship product will hold.
Next, work each axis one at a time. On the customer axis, secure a second client — even a small one — that's different in character from your big account; even a 10% revenue share changes your voice at the negotiating table. On the region axis, open up a second sales channel instead of staying locked into one platform. On the market axis, prepare a prototype that adapts the technology or know-how behind your current bestseller for a different customer segment — the same principle behind Nvidia redirecting its GPU supply capability into a robotics platform. Diversifying away from customer concentration risk comes down to where you spend the extra time and cash while things are going well.
One Thing to Try This Week
Write your revenue breakdown by client on a single sheet of paper. If your top client accounts for more than 40%, reach out to one prospective client with a different profile before the week is out. Right now, while you're busy with a backlog of orders, is the best time to make that call.
