How much weight should investors give a company's current and projected gross margins when underwriting?

Harry Stebbings · Sep 25, 2024

Gross margin at $3M ARR is irrelevant to an early-stage investment decision.

You have no idea what the margin will be in ten years, and even a poor margin can fetch a great multiple in good times while a great margin can be trapped in a closed IPO market — there are too many variables to know.

44:00 20VC: Benchmark's Eric Vishria on Where is the Value in AI: Chips, Models or Apps | Why Nvidia Will Not Be The Only Game in Town | The Commoditisation of Foundation Models | Which AI Apps Have Sustaining Value vs Hype and Short Term Revenue

Cem Sertoglu · Nov 20, 2024

Investors should be more open-minded about what a young company's ultimate margin structure could become rather than judging it on present-day unit economics.

They passed on Taxify/Bolt at seed because they doubted the warranted take rate given Uber's market pricing and couldn't see a global trajectory from a small Baltics player — their biggest loss.

Scope: framed as a hindsight lesson from one miss

14:41 20VC: Turning $16.5M into $2.1BN; Lessons from the Greatest Venture Investment in European History: UiPath | Why VC is Not Being Commoditised | Why Price Does Not Matter | Lessons on Loss Ratio, Selling and Signalling with Cem Sertoglu

Mitchell Green · Mar 28, 2025

Passing on Snowflake at a $500M valuation because of its horrible gross margins was a 100% wrong call

They judged it on gross margins at the time and were completely wrong

38:18 20VC: Why Traditional VC is Broken: How VCs Learned Nothing from 2021 | Why LPs are More Important than Founders & Advice to Emerging Managers | Bull Case for Bytedance & Why TikTok's Ban Doesn't Matter with Mitchell Green, Lead Edge Capital

Martin Mignot · Aug 11, 2025

Bad gross margin in a company's early days should generally be ignored as a reason to pass on an investment

Many businesses, including all the LLM providers, had really bad gross margins early on, so it is a poor early signal

Scope: early days / early stage only; if gross margin is the only thing holding you back; in most cases

0:00 20VC: Figma, Scale, Wiz: Inside Index's Decacorn Factory | Decision-Making, Investment Process, Biggest Lessons, Biggest Misses | Why Gross Margin is a Fallacy at Seed | Never Turn Down a Deal on Price with Martin Mignot, Partner @ Index Ventures

Martin Mignot · Aug 11, 2025

Investors should beware of using gross margin as a filter in a company's early days; if low gross margin is the only thing holding you back, it should be ignored

Early on, founders should be focused entirely on growth and product, not margin optimization; Revolut, Snowflake, Deliveroo and the LLM providers all had very poor early gross margins

Scope: 'in most cases'; if gross margin is the sole objection

25:37 20VC: Figma, Scale, Wiz: Inside Index's Decacorn Factory | Decision-Making, Investment Process, Biggest Lessons, Biggest Misses | Why Gross Margin is a Fallacy at Seed | Never Turn Down a Deal on Price with Martin Mignot, Partner @ Index Ventures

Byron Deeter · Aug 25, 2025

Investors should not walk from deals purely because early unit economics don't pencil — a generational entrepreneur can power a company through to later model arcs that do work

He passed on Tesla because he couldn't see the Roadster's unit economics working — and they didn't — but Elon put the company on his back and pulled the later arcs through

Scope: framed as a personal regret and challenge to his own short-term discipline

34:38 20VC: Do Margins Matter in AI? | Is Defensibility Gone For Good? | Is Vertical SaaS Dead in a World of AI | What SaaS Rules Are BS and No Longer Apply in a World of AI | The Future of Venture: Why Chanel vs Walmart is BS with Byron Deeter

Harry Stebbings · Mar 16, 2026

Investors are wrong to reflexively dismiss companies for weak early gross margins, because many of the best companies had poor margins for years before expanding them.

DoorDash, Deliveroo and Spotify all ran with bad margins early and Spotify's margin expansion since has been remarkable.

8:29 20VC: The 8 Moats of Enduring Software Companies: How to Analyse for Durability and Defensibility in a World of AI | Why Dropouts are "AI Maxing" the World & Remote Early-Stage Companies are Dying with Gokul Rajaram

Gili Raanan · Mar 28, 2026

Gross margins should never be discussed with early-stage companies; they are a problem to tackle years later

Part of an investor's job is sequencing which challenges a founder tackles this year versus in 2027, 2028 — first build the foundations of a healthy business

Scope: for young portfolio companies; margins become the conversation around 2029

24:49 20VC: The Venture Model is Broken | You Need to be Greedy and Selfish to Win Early Stage Investing | Why Margins Do Not Matter for Early-Stage Startups | The Growth Rate that is Required in a World of AI with Gili Raanan, Founder @ Cyberstarts

Everett Randle · Nov 10, 2025

AI app companies should be evaluated on absolute gross profit dollars per customer and gross profit multiples rather than gross margins and revenue multiples

AI inference sits in COGS so margins look worse, but if gross profit per customer is 4-5x a SaaS company's, the absolute dollars and addressable market are far larger

Scope: applies to AI app companies with heavy usage, not SaaS

22:51 20VC: Benchmark's Newest General Partner Ev Randle on Why Margins Matter Less in AI | Why Mega Funds Will Not Produce Good Returns | OpenAI vs Anthropic: What Happens and Who Wins Coding | Investing Lessons from Peter Thiel and Mamoon Hamid

Everett Randle · Nov 10, 2025

Investors should not place much emphasis on current margins for AI companies; the useful work is reasoning from first principles about terminal gross margin structure and absolute gross profit dollars five to seven years out

comparing to SaaS is misleading because AI apps have a very different pricing and business model, and today's margins say little about the endpoint

Scope: 'today' — about current-stage AI companies; admits terminal margins are hard to know

24:54 20VC: Benchmark's Newest General Partner Ev Randle on Why Margins Matter Less in AI | Why Mega Funds Will Not Produce Good Returns | OpenAI vs Anthropic: What Happens and Who Wins Coding | Investing Lessons from Peter Thiel and Mamoon Hamid

Gokul Rajaram · Mar 16, 2026

Margins in years one and two — even years four and five — are not worth analysing; what matters is whether the company will have pricing leverage over its customers

Inference costs are already dropping so cost-side margin improvement happens by nature; the differentiated thing is defensibility that lets you raise prices, as with PayPal's repeated price increases and Uber's steadily expanding economics

39:43 20VC: The 8 Moats of Enduring Software Companies: How to Analyse for Durability and Defensibility in a World of AI | Why Dropouts are "AI Maxing" the World & Remote Early-Stage Companies are Dying with Gokul Rajaram

Gokul Rajaram · Mar 16, 2026

You should not forecast margins at all — you should assess whether the product is compelling enough to have the potential to raise prices later; durability and defensibility are the real worries because defensible companies will be able to expand margins

Price increases take many years to materialise (Disney+ and Amazon Prime raised prices repeatedly over time), so the analysis is about potential rather than projection

41:14 20VC: The 8 Moats of Enduring Software Companies: How to Analyse for Durability and Defensibility in a World of AI | Why Dropouts are "AI Maxing" the World & Remote Early-Stage Companies are Dying with Gokul Rajaram

Eric Vishria · Sep 25, 2024

Early-stage unit economics never extrapolate to scale, so precision about metrics like net dollar retention at $1–1.5M in revenue is false precision.

Across the five companies he has worked on that went from zero to over $200M in revenue, not one had economics at the early stage — or even at $30–50M — that extrapolated to $200M; too much changes inside these companies.

44:33 20VC: Benchmark's Eric Vishria on Where is the Value in AI: Chips, Models or Apps | Why Nvidia Will Not Be The Only Game in Town | The Commoditisation of Foundation Models | Which AI Apps Have Sustaining Value vs Hype and Short Term Revenue

Harry Stebbings · Mar 16, 2026

In today's environment it is impossible to have any view on a company's margin structure in years three, four and five

The world is changing too fast for those projections to mean anything

41:00 20VC: The 8 Moats of Enduring Software Companies: How to Analyse for Durability and Defensibility in a World of AI | Why Dropouts are "AI Maxing" the World & Remote Early-Stage Companies are Dying with Gokul Rajaram

Harry Stebbings · Oct 2, 2023

Margins matter and the opportunity cost of capital is real — even where you can make money, as in selling software to restaurants, there are easier places with higher upside to deploy the same capital

Some businesses are so hard that the difficulty isn't worth it relative to alternatives available for the same dollars

45:05 Opportunity cost of capital means margin and market quality matter even when a business can make money

Cem Sertoglu · Nov 20, 2024

Take-rate and margin questions only matter when you can't see a company becoming a global contender; if you can see the global outcome, unit economics at seed are irrelevant.

At the time Bolt looked like a small Baltics ride-hailing player, so the market and margin questions dominated the decision.

15:45 Margin scrutiny only matters absent a visible path to global scale

Jeff Jordan · Jan 16, 2023

Deeply negative early unit economics are investable when the founder can sketch a concrete roadmap of revenue and operational improvements and has the founder-product fit to execute it

Instacart was losing $20-25 per order on $10-12 of billings, but Apoorva, an ex-Amazon fulfillment engineer wired for continuous improvement, laid out ads, grocer economics deals, and process changes like dedicated lines and in-aisle scanning that took minutes out of each order — and the company is now significantly profitable

28:42 Concrete improvement roadmap and founder fit justify investing despite deeply negative unit economics

Parker Conrad · Oct 3, 2022

Most early-stage software companies' stated margins should be distrusted — they typically look very good early and get worse at scale

Costs handled ad hoc by founders or stray engineers never get allocated to cost of delivering the service, and rapid growth further distorts the assessment; margin profiles take a long time to stabilize

21:52 Early stage margins look good but deteriorate at scale due to unallocated costs

Peter Singlehurst · Mar 19, 2025

The growth-versus-profitability trade-off is a misnomer; the right question is incremental return on invested capital and long-run return on equity

32:06 Incremental return on invested capital not the growth vs profitability tradeoff is the right lens

David Cahn · Oct 27, 2025

Gross margins matter only as a directional indicator of how much product a company has built on top of the foundation models, not as an absolute measure of business quality

Gross margins rise over time — Snowflake was indicted early for low gross margins and became a very good business; in AI, compute costs fall every year so the trend line is clear

39:08 Margin level signals directional product depth not absolute business quality

Eric Vishria · Sep 25, 2024

In non-software businesses, early gross margins and unit economics genuinely matter, and a founder who isn't engaged on that topic is a disqualifying signal.

Sarah pointed out that in the company they were looking at the unit economics were bad with no path to improvement and the entrepreneur wasn't engaged; the mismatch between a compelling founder and the nature of that business was the deciding issue.

42:49 In non software businesses margins and founder engagement on economics matter