Picture a solo founder whose freelance fees and software subscriptions climb in lockstep with revenue. Double the number of contracts, and the designer's freelance budget doubles too, the collaboration tools get bumped to the next pricing tier, and the workload doubles right along with them — yet the balance sitting in the bank account looks about the same as it did last year. When you're working hard and still can't figure out why nothing sticks, a financial statement can turn into an unexpected textbook. Today we'll open up Nvidia's 10-K and learn how to read an R&D spending ratio. It's a semiconductor company's story, but the lens you pick up here maps directly onto the numbers of a one-person business.

A falling ratio doesn't mean shrinking investment

A spending ratio is simply the share of revenue that a given expense eats up. For R&D, that means research and development spending as a percentage of revenue. There's a trap here that's easy to fall into: a falling ratio doesn't necessarily mean a company is investing less. If revenue grows faster than the expense does, the ratio drops even while the dollar amount keeps climbing. Read the ratio alone, and you'll draw exactly the wrong conclusion. What you actually need to do is read two numbers together — where the percentage is heading, and where the dollar figure is heading. The moment these two diverge is the signal that a business is graduating from a one-person operation into a platform.

The moment Nvidia's numbers diverge

Nvidia's three-year numbers show this divergence in sharp relief. Its R&D ratio fell from 27.0% in FY2023 to 8.6% in FY2026. Looking at that line alone, you'd think the company had scaled back research. But over the same period, R&D spending in dollar terms grew 2.5x, from $7.3 billion to $18.5 billion. Investment didn't shrink — revenue simply grew much faster. Revenue climbed from $27 billion to $215.9 billion, putting the three-year compound annual growth rate north of 88%.

Why did this happen? Because the CUDA development and architecture Nvidia poured resources into early on became the industry standard. Once you're the standard, pushing revenue up another notch doesn't require nearly as much additional development. The structure shifts: new revenue simply stacks on top of a foundation that's already built.

Once you become the standard, revenue outruns investment

Here's that story laid out as a single flow.

flowchart TD
    A["Upfront investment in CUDA"] --> B["Becomes the industry standard"]
    B --> C["Revenue grows without added R&D"]
    C --> D["R&D ratio: 27% down to 8.6%"]
    C --> E["R&D dollars: up 2.5x"]

The same principle shows up in profitability. Operating margin climbed from a low of 15.6% in FY2023 to 54.2% in FY2024, 62.5% in FY2025, and 60.4% in FY2026. But the more telling detail is that free cash flow has held steady at 44–47% of revenue — meaning that's the cash actually landing in Nvidia's hands, not just profit on paper. Nor did the company lever up to get there: its debt-to-equity ratio fell from 0.23x in FY2024 to 0.05x in FY2026. Rather than borrowing to bulk up while times were good, Nvidia funded its growth with the cash it was already generating.

Draw the same two lines for your own business

Now let's bring this back to your own numbers. Start by listing your revenue for the past three years alongside the core expense you've repeatedly paid to generate it. For a one-person business, that's typically freelance fees, tool subscriptions, or ad spend. Then plot two lines: one for that expense as a percentage of revenue, and one for the expense in absolute dollars.

Watch which direction those two lines move. If the expense ratio keeps climbing right alongside revenue, you're still in the one-person-operation stage — more work always demands proportionally more of your own hands, which is exactly why nothing accumulates in the bank. But if the absolute spending keeps rising while the ratio falls, something you built once has started carrying repeat revenue on its own. That something is your CUDA. It might be a standardized proposal template, an automation script you wrote once that keeps running on its own, or an onboarding process that, once set up, just runs itself.

The real question is where you put your next investment. Keep pouring money into variable costs that scale directly with revenue — per-project freelance work, say — and the ratio will never come down. Put it instead into an asset you build once and sell repeatedly, and even though the dollar amount feels heavier for now, a day will come when revenue stacks on top of it and the ratio finally bends.

One thing to try today

Next to your revenue for the past three years, write down the ratio and the absolute amount for your single largest recurring expense. If both lines are climbing together, redirect one purchase next month away from a per-project cost and toward an asset you'll keep using for years. The point where that ratio first bends is exactly where your business crosses over from a one-person operation into a platform.