Every dividend investor eventually asks the same question about a non-dividend stock: if the company isn’t paying me, where is the money going? The blunt answer is that it’s being reinvested. The interesting answer — the one worth an entire guide — is into what, at what rate of return, and for how long can it keep going.

That’s the growth engine. A well-built engine takes a dollar of retained earnings and, five years later, hands you back several dollars of durable free cash flow. A broken engine takes the same dollar and, five years later, hands you a write-down. The single most important skill in non-dividend investing is being able to tell the two apart on the way in rather than on the way out.

The economic argument for reinvestment

Warren Buffett has framed this cleanly for six decades. Roughly paraphrased: a company should only retain a dollar of earnings if it can produce more than a dollar of long-term market value with it. If it can’t, that dollar belongs back in shareholders’ hands via a dividend or a buyback. The math is arithmetic, not philosophy.

Non-dividend investing lives entirely in the first two rows of that table. The engine is the machine that keeps the return on retained capital above the cost of capital, year after year.

The four forms of reinvestment

Every dollar a non-dividend company keeps ends up in one of four buckets. Recognizing which bucket dominates for a given ticker is half the analysis.

DiviDrip surfaces the ratio between these four buckets under the Capital Analytics tab. A company that spends 90% of its reinvestment budget on M&A gets a different — usually lower — Capital Reinvestment Score than one that spends the same total dollars on organic R&D and capex. The mix matters at least as much as the amount.

Flywheels: what actually creates compounding

Reinvestment on its own doesn’t compound. Ten dollars into R&D that produces a product nobody wants is just an expense. The magic ingredient is the flywheel — a self-reinforcing loop where the output of one cycle becomes the input to the next, and each turn requires less energy than the last.

Three real historical examples explain the concept better than theory does.

Amazon’s original flywheel

Jeff Bezos famously sketched the original Amazon flywheel on a napkin in 2001. Every arrow points into another arrow: lower prices attract more customers, more customers attract more third-party sellers, more sellers means better selection, more volume drives down cost structure, lower costs allow still- lower prices. That loop ran for two decades. The stock returned roughly 200,000% cumulatively from the 1997 IPO without a single dividend paid.

NVIDIA’s software-hardware flywheel

NVIDIA’s CUDA software layer, launched in 2007, was given away free. Every engineer who wrote CUDA code became locked into NVIDIA hardware. As the AI wave crested from 2022 onward, that fifteen-year investment paid off in a data-center revenue line that went from roughly $340M in fiscal 2016 to over $47B in fiscal 2024 — a roughly 140-fold increase. The flywheel: more developers → more optimized workloads → better hardware performance → more developers.

Netflix’s content-data flywheel

Streaming launched at Netflix in January 2007. The company used subscriber viewing data to inform which content to commission, which reduced the hit-rate randomness that plagues traditional Hollywood studios. Better original content attracted more subscribers, which produced more viewing data, which improved the next round of commissioning decisions. Streaming revenue went from under $1.5B in 2008 to more than $37B by 2024 — a roughly 25-fold increase — while the underlying flywheel remained intact.

Every one of these companies paid zero dividends across the entire compounding phase. Every dollar of profit went back into the flywheel. That’s the growth engine at work.

The four warning signs an engine is stalling

Engines fail slowly, then all at once. If you own a non- dividend stock, you want to catch the stall while the price chart still looks fine. Four signals, in rough order of severity:

The Operating Momentum score on the Capital Analytics tab rolls three of those four signals into a single z-score relative to the company’s own five-year history. That’s the fastest way to notice a stalling engine without having to build the spreadsheet yourself.

Applying it: your first non-dividend engine check

Here’s the workflow that consistently separates strong engines from weak ones. Every step is on the Capital Analytics tab in DiviDrip by TwylightCrow.

1. Start with the Capital Reinvestment Score. If the composite grade is below 60 / 100, the engine has real problems. Above 85 / 100 is fortress territory. Between those bands is where most of the interesting screening happens.
2. Read CFROI next. If cash flow return on invested capital is above the company’s weighted average cost of capital by at least 400 basis points for three straight years, the engine is genuinely producing value. Below that spread, you’re paying for growth that isn’t compounding.
3. Confirm the Rule of 40. Growth plus margin above 40 keeps you inside the healthy efficient-scaling zone. Below 30, the engine is burning fuel faster than it’s producing.
4. Check Buyback Effectiveness. Retained earnings that never make it back to a lower share count are just management’s money, not yours. See The Anatomy of a Stock Buyback for the full trap.
5. Clear the forensic gate. Beneish M-Score and Altman Z-Score are the last checks. If either flashes red, close the tab and go find a different engine. There are 3,000+ non-dividend stocks in the DiviDrip universe — there’s no reason to fight a forensic-red setup.

FAQ

What does the phrase "growth engine" actually refer to?

It’s a shorthand for the internal mechanism that turns retained earnings into more retained earnings. Every non-dividend company keeps 100% of its net income by design, and the useful question isn’t whether it grows the top line — that’s table stakes — but whether the capital it keeps compounds at a higher rate than you could earn deploying it yourself. When the engine works, each incremental dollar the company holds back produces more than one dollar of future free cash flow. When the engine sputters, retained earnings decay into asset write-downs, dilutive acquisitions, or shares bought back at absurd prices.

How is this different from just looking at revenue growth?

Revenue growth alone tells you almost nothing about the engine. Uber grew revenue from roughly $8B in 2018 to $37B in 2023 while burning cash the entire time — high growth, negative engine. Meanwhile Costco grew revenue at a much slower clip through the same period but converted almost every incremental dollar to durable free cash flow. The growth engine is a joint read of top-line acceleration, incremental margin conversion, capital intensity, and the reinvestment runway ahead. DiviDrip’s Capital Analytics tab is built specifically to surface those four dimensions in one screen instead of forcing you to model them yourself.

What are the four forms of reinvestment a growth company can pursue?

(1) R&D — hiring more engineers, funding more product bets. Highest option value, longest payback. (2) Capital expenditure — building data centers, factories, distribution. Slow to compound, but creates the deepest moats when done right (NVIDIA’s CUDA ecosystem, Amazon’s fulfillment network). (3) Acquisitions — the fastest way to buy revenue, and statistically the most destructive form of reinvestment for shareholders. Roughly half of large M&A deals destroy value versus the counterfactual of just holding cash. (4) Buybacks and debt paydown — the quietest form, sometimes the smartest, and the easiest to fake through option-grant absorption. The Anatomy of a Stock Buyback guide covers that trap in detail.

What is a "flywheel" and why do investors obsess over the word?

A flywheel is a self-reinforcing loop where each spin makes the next one easier. Amazon’s original flywheel — famously sketched on a napkin by Jeff Bezos in 2001 — was: lower prices, more customers, more sellers, better selection, lower cost structure, still lower prices. Each pass around the loop cost less energy than the last. Non-dividend investors chase flywheels because they are the closest thing to compound interest that exists inside an operating business. Not every growth stock has one. Snap and Peloton, for example, grew revenue for years without a genuine flywheel, and the growth was subsidized by capital markets until it wasn’t.

When does a growth engine actually stall?

Four warning signs, in rough order of severity. First: incremental revenue starts requiring more capital than the last cohort of revenue — capital intensity is rising. Second: gross margin plateaus and then softens, meaning the unit economics have hit their ceiling. Third: R&D as a percentage of revenue climbs but growth doesn’t follow — spend is decoupling from output. Fourth: the CEO begins talking about "adjacencies" instead of the core business — which is often the tell that management believes the primary flywheel has slowed. DiviDrip’s Operating Momentum score turns three of those four signals into a single z-score so you can screen for engine stalls before they show up in the price chart.

How do I actually evaluate a non-dividend company’s engine on DiviDrip?

Open the ticker and click into the Capital Analytics tab. Start at the top: the Capital Reinvestment Score gives you the composite grade (0–100). Below that, look at CFROI — cash flow return on invested capital — which is the raw efficiency read. Then Rule of 40 to confirm growth + profitability haven’t decoupled. Then Buyback Effectiveness to check whether retained earnings are actually landing in your pocket via a shrinking share count. Then Beneish + Altman as the forensic safety gate. If any one panel is deep red, keep hunting. The strongest engines are the boring cases where every panel reads healthy green — those are the compounders you can own through drawdowns without second-guessing the thesis.

Try it

Pick a name you already own and one you don’t. Open the Capital Analytics tab on both. Line the five panels up side by side — Capital Reinvestment Score, CFROI, Rule of 40, Buyback Effectiveness, Forensic Safety — and ask yourself which engine looks more durable. The value of the exercise isn’t which stock wins; it’s that you can now see what durability looks like in numbers instead of guessing from the price chart.

For the underlying math and history, see the Capital Reinvestment Score and CFROI glossary entries. For the mental-model layer that sits above the engine analysis, read The Capital Appreciation Playbook.

This guide is educational. Historical flywheels don’t guarantee future compounding — every engine eventually reaches a stall point, and the job of the non-dividend investor is to notice the stall while the numbers still look fine.