Valuing Companies With No Earnings (Without Hand-Waving)

In 2020 and 2021, more than 70% of US companies completing initial public offerings reported net losses in the year before listing. That figure isn't a product of the SPAC frenzy alone — it represents the continuation of a trend that began in the 1980s, when loss-making IPOs accounted for roughly 20% of the total. By the peak, the median company going public had never reported a profitable quarter. Which means that for a growing share of publicly traded equities, every valuation tool in the standard kit breaks down before you start.

This is the problem. Price-to-earnings requires earnings. EV/EBIT requires positive EBIT. Owner earnings multiples require something that at least resembles a stream of real cash. When the denominator is negative or absent, the formulas either produce a nonsense number or simply refuse to work. And yet the stock trades. Someone prices it. Someone decides it's worth owning at that price.

The temptation at that point is to wave your hands: cite the total addressable market, multiply by an assumed penetration percentage, scale to a revenue projection, apply a multiple — and call the result analysis. Not a thesis. There are legitimate frameworks for valuing no-earnings companies, but they require more explicit assumptions, more honesty about where the uncertainty lives, and more discipline about what a revenue multiple or a growth projection actually implies about the underlying business. This post lays out the ones that hold up.

When Earnings Aren't There Yet

Not all no-earnings companies present the same analytical problem. The framework you need depends on what kind of 'no earnings' you're dealing with.

The first category is pre-revenue companies — truly early-stage businesses where cash flows are speculative projections at best. These are nearly impossible to value with rigor in a public market context. You're estimating the size and timing of cash flows that don't exist, for a business that hasn't proven its unit economics, in a market whose competitive structure is still forming. Most public market investors don't seriously attempt it, not because it's wrong to try, but because the error bars are wide enough that the output barely constrains the decision. This is primarily venture territory.

The more common case for public market investors is the company generating significant — often fast-growing — revenue while reporting GAAP losses. Think early Airbnb, Spotify through most of its first years as a public company, the cohort of enterprise software businesses that went public in the 2010s. The GAAP losses in these businesses often reflect accounting constructions that don't map to cash reality: heavy stock-based compensation, amortization of acquired intangibles, or aggressive upfront investment in customer acquisition that GAAP expenses immediately but generates returns over multiple years. Free cash flow may already be positive — or on a visible trajectory toward it. The real question isn't 'does this company make money?' It's 'what do the cash economics look like when you strip out the accounting noise?'

The third category — and the hardest — is companies that are both GAAP-negative and cash-flow-negative, burning through investor capital to fund growth. Amazon in 1999 and 2000 was in this category. Peloton in 2021 was in this category. The difference between them wasn't the accounting entries or the revenue growth rates. It was whether the underlying unit economics could ever support attractive returns at scale. That is the question that matters — and it's the one that too many analyses skip entirely.

Revenue Multiples: A Starting Point, Not an Endpoint

When a company has no earnings, the valuation conversation defaults to enterprise value divided by revenue. It's the least bad option in a limited toolkit. Used carefully, it's a starting point. Used carelessly, it produces numbers that are just as speculative as the TAM approach — only with a denominator that makes them look more rigorous.

EV/Revenue = Enterprise Value ÷ Annual Revenue To stress-test any given multiple, translate it back into an implied earnings assumption: Implied Mature Market Cap ≈ Revenue × Target Net Margin × Target P/E Multiple Present Value of That Outcome = Implied Mature Market Cap ÷ (1 + Discount Rate)^n If this present value exceeds today's market cap, the multiple may be justified — provided the margin, multiple, and timeline assumptions are realistic. If not, you're paying for optimism that hasn't been stress-tested.

The discipline is in the translation, not the multiple. A company at 10× revenue, assuming 20% net margins at maturity and a 20× P/E multiple on those earnings, implies something roughly fair relative to eventual cash generation — if the margin assumption holds. A company at 50× revenue with the same margin assumptions implies the market is pricing in roughly a tripling of revenue before reaching that steady state. Both can be defensible for the right business. Both can be wildly optimistic for the wrong one.

Cerebras Systems offers a timely example. The AI chip company targeted a Nasdaq IPO in May 2026 with an anticipated valuation of $26–$27 billion. In fiscal year 2025, Cerebras reported revenue of approximately $510 million — up roughly 20× year-over-year — but recorded operating losses of $145.9 million before a $363 million gain from a financial liability that made GAAP income appear positive. Strip that accounting item out and the company is operationally loss-making. At $26–$27 billion, the valuation implies roughly 50× trailing revenue. That may be defensible if Cerebras can diversify its customer base beyond a single hyperscaler, capture share of the AI inference market, and expand margins toward software-like levels. But each of those 'ifs' is a named assumption that has to survive contact with Nvidia's installed base, hyperscaler vertical integration, and competitive entrants with deeper capital. The multiple is a shortcut. The assumptions are the actual work.

For context: in 2025, profitable SaaS businesses traded at a median enterprise value-to-revenue multiple of approximately 7.8×, versus around 6.7× for unprofitable peers — a gap that reflects the discount investors apply for cash-flow uncertainty. When a company sits at 50× revenue, the investment case requires margin expansion and competitive durability that exceed what most comparable businesses have achieved. That's a hypothesis to test, not a number to accept.

Unit Economics and the Contribution Margin Test

The single most important analytical check on a no-earnings company isn't how fast revenue is growing. It's whether each incremental unit of business generates more value than it costs to produce — and whether that gap is wide enough to eventually absorb the overhead required to run the company at scale.

Contribution Margin = Revenue − Variable Costs For subscription businesses: Lifetime Value (LTV) ÷ Customer Acquisition Cost (CAC) > 3× is the common threshold for sustainable growth economics. Below 1×, you lose money acquiring customers at any scale. For marketplaces: Gross Profit per Transaction must cover the variable overhead of facilitating that transaction before any fixed costs enter the calculation.

This test separates 'unprofitable because we're investing in growth' from 'unprofitable because the business model doesn't work.' A company with positive contribution margins is funding a fixed-cost base that should become less burdensome as revenue scales — more revenue, same server infrastructure, same leadership team, same office leases. A company with negative contribution margins is in a categorically different situation. Scale makes the losses worse. No revenue growth fixes a broken contribution margin.

Peloton — ticker PTON — is the cautionary example that repays careful study. In fiscal year 2021, Peloton reported revenue of $4.0 billion, up approximately 120% year-over-year. Contribution margins were positive. LTV/CAC calculations, on their face, looked defensible. The stock reached a market capitalization above $50 billion. But the unit economics were predicated on a temporary structural condition: pandemic-era gym closures had inflated demand for at-home fitness equipment in a way that masked the true cost structure of the business at normalized demand. When gyms reopened in fiscal year 2022, subscriber growth stalled. The fixed cost base — content production, logistics infrastructure, retail stores — that looked manageable at $4 billion in revenue became unsustainable at the volumes that followed. By January 2023, the stock had fallen more than 90% from its pandemic-era peak.

I'm genuinely less certain about the durability of unit economics in early-stage businesses than I'd like to be. The evidence for structural unit economics — unit economics that hold up without the tailwind — often only becomes visible after the tailwind has faded. The honest question isn't 'do the current numbers look okay?' It's 'do the numbers look okay in a world where the conditions that helped create them are no longer present?' That question is harder to answer and usually more important.

Why TAM × Market Share Is Almost Always Wrong

The most seductive argument in no-earnings valuation is the least reliable. Total addressable market, multiplied by an assumed penetration percentage, scaled to a revenue estimate at some future date, multiplied by a target margin and a terminal P/E multiple. Systematically wrong. Not because the arithmetic is flawed — it isn't — but because the inputs are, at bottom, invented.

TAM estimates routinely overstate addressable markets by counting every dollar spent in adjacent categories as potentially capturable. A company selling digital payroll software doesn't compete for 100% of 'HR software spending' — it competes for a slice defined by company size, geography, regulatory complexity, and integration requirements. Market share assumptions discount the dynamics that favor incumbents: network effects that create lock-in, switching costs that reduce churn, and regulatory advantages that slow entry. The timeline to maturity — almost always 'five years' — is chosen because it's long enough to avoid near-term falsification, not because there's evidence the market develops on that schedule.

Amazon's early public history is the counterexample everyone reaches for, and it's worth understanding correctly. The investment case for Amazon in fiscal year 2001 — when it first turned a quarterly profit — wasn't that the company would capture 15% of global retail. It was that Amazon could earn structurally higher returns on capital than physical retail by eliminating the inventory-carrying costs that compress brick-and-mortar margins. The DCF analysis didn't require projecting share of a $10 trillion market. It required estimating what the unit economics would look like at a defensible, real-world scale. That is a testable hypothesis — and as Amazon's gross margins improved and capital efficiency compounded through fiscal years 2002 and 2003, the hypothesis accumulated evidence.

I ask the same question whenever a pitch rests primarily on TAM × share: is there a unit economics argument here? One built from actual cost and margin data? If not, that's the gap. TAM × share is what you reach for when you can't yet make the bottoms-up case — and reaching for it earlier than the evidence warrants is how investors overpay for businesses that never reach the margin structures implied by the multiple.

Key Takeaways

Valuing companies with no earnings is real analytical work. But the rigor lives in the assumptions, not the formulas.

Three checks that hold up across cases: First, classify what kind of 'no earnings' you're dealing with — GAAP-negative but cash-flow-positive is a different problem from cash-burn-negative with unproven unit economics. Second, if you're using a revenue multiple, translate it back into a specific implied steady-state earnings hypothesis: what net margin, what revenue scale, what timeline? Test those assumptions against industry comparables, not just against the bull-case narrative. Third, build the unit economics argument from the bottom up before accepting any TAM-based framing. If you can't get to a convincing contribution margin story, the TAM doesn't rescue it.

The margin of safety principle applies here in a modified form. For no-earnings companies, it means insisting on conservative assumptions at each step — the lower-end margin trajectory, the longer-than-expected path to positive cash flow, the cautious market share estimate. That conservatism isn't pessimism. It's how you arrive at a fair value estimate that functions as an actual compass — one that tells you when the price has moved far enough from plausible fundamentals to represent a real opportunity, and when it hasn't.