Lab
The inventor that creates breakthroughs before the market even asks.
The Pattern
What the Lab pattern is
Capital burns at a Lab for years before any customer asks for what it's building. A small team is chasing a moonshot that may or may not exist. The white coats, the GPU racks running at 2 AM and the whiteboards full of equations nobody requested are where that money goes while the science develops.
Think: Y Combinator, early OpenAI, Google X, Boston Dynamics, pre-pivot Magic Leap, pre-Alphabet DeepMind, pre-revenue biotech.
The Engine
How it makes money
A Lab runs on its lead researcher, or its research team, whose vision and credibility justify the capital going in. The money matters, but it's the fuel. The person is the keystone. Investors write checks to a person they believe will produce breakthrough IP (intellectual property), and they aren't writing them to a treasury. If that lead researcher leaves, the investors on your cap table (the list of who owns the company) usually walk within 18 months.
A Lab exists to invent things the market doesn't know how to ask for yet. Its strength is holding one clear technical thesis through years of uncertainty, until it produces defensible IP or results that competitors can't copy. A thesis here means the core bet about what will work. If that bet lands, the Lab gets upside big enough to define a whole category.
The Discipline
Operating principle
Every Lab is betting that demand will show up for something that doesn't exist yet. That puts the founder's conviction up against a market that hasn't said anything yet.
Staying alive comes down to runway management, meaning how many months of cash you have left. Every decision comes back to how long you can keep the lights on while the science develops. It works as a loop. A scientific milestone is what brings in capital, and that capital buys the time it takes to produce the next milestone. Some Labs ride that loop all the way to commercialization (actually selling a product), and some never get there.
Healthy
Healthy Labs pick one thing to be the best in the world at and stay there, instead of spreading into a broader portfolio.
Unhealthy
The R&D (research and development) bill is high and it runs for years, because typical horizons are 5-10 years before any commercial revenue comes in.
When the bet is wrong, a Lab ends up spending a decade producing papers that no one cites.
What Works & What Breaks
Healthy vs. breaking
What healthy looks like
What breaks it
The Profile
Profile
Best Fit Founder
This suits a visionary scientist or engineer who's comfortable with risk and does their best work experimenting. They handle ambiguity well, but they struggle with the tedious operational side of running a company.
Go-To-Market
A Lab reaches its market through early adopters and pilot programs, and through the academic route of conferences, peer-reviewed papers and patent filings. It also takes on a few selective partnerships with strategics (large companies in the space) or grant programs. Demand comes from how credible the science and the team are, so performance marketing (paid ads judged on direct sales) doesn't do much here.
Ideal Focus
We'd keep a Lab narrowly focused on patents and defensible IP, and on long-term innovation bets inside one problem domain where solving it would have civilization-scale upside. Spreading out kills Labs faster than it kills any other archetype, because each new direction splits up the talent and the IP portfolio, and it muddies the story investors bought into. The right focus sounds like "we are the world's best at X, and we will spend a decade compounding that lead." Going broader waters down the one thing that actually defends the company, which is depth of expertise no one else has.
Revenue Pattern
Revenue comes late and in lumps, and early on there often isn't any. The standard Lab P&L shows three to five years of essentially zero revenue. After that come grant or licensing payments that are hard to predict, and then one big step up when the product is commercialized. That's why standard SaaS-style forecasting (the steady subscription growth curves software companies project) is meaningless for a Lab.
Unhealthy Signs
Risk, Speed & Burnout
Sales Cycle
Material revenue is 3-10+ years away, and sometimes a Lab never gets there. The first 3-5 years typically bring in zero commercial revenue, and after that the first money is small grant or pilot revenue. We'd plan capital reserves on the assumption that revenue is absent for the whole first half-decade.
Hiring Sequence (first 5 after founder)
- Co-founding researcher: works as a technical peer to the founder
- Second senior researcher: proves the lab's depth and gets the recruiting flywheel going, since strong researchers attract more
- Lab / operations manager: takes logistics and grant admin off the researchers' plates
- Head of partnerships / BD: runs business development, which means bridging to the commercial world and owning grants and strategic deals
- CFO: owns capital strategy before the lab scales, because for a Lab that strategy decides whether the company survives
Lab tech is specialized to the research domain, so it can mean wet labs, GPU clusters, simulation platforms or custom hardware, and it's often the single largest cost item.
Most of the work is deep and quiet and runs over long stretches, with sprints around grant deadlines and demo days. The thinking is intense, even though day-to-day operations move at a slow tempo.
Leading KPIs
- Months of runway remaining: this is the one metric every Lab shares, and if it hits zero, nothing else matters.
- Burn rate trend: a burn that rises without the milestones rising with it is the earliest warning that the organization is drifting.
- R&D spend % of revenue (or of capital deployed pre-revenue): this shows whether the capital is actually going to the science or to overhead.
- Patents filed per year: this is one of the few hard outputs that survives milestone-shifting, and a Lab that stops filing has stopped producing.
- R&D efficiency (% of projects reaching market or measurable milestones): this is how often experiments turn into outputs, and it's a leading indicator of whether the team has a real thesis.
- Product development cycle time: this shows whether the team is converging on an answer or wandering.
- Proof-of-concept adoption rate: this is the first commercial validation signal, because early customers willing to pay anything show that real demand exists.
- % of prototypes commercialized: this is the number that separates "interesting science" from "value-creating science" in a Lab.
- Top-quartile researcher retention rate: the team is the hardest moat a Lab has, and losing top researchers can't be undone.
- Strategic partnership / grant pipeline value: this is the leading indicator of whether capital and commercial interest will be there for the next round.
Feature Focus
Experiments can produce breakthroughs that define entire categories.
Failed experiments earn a low ROI (return on investment), and the team lives with chronic anxiety about the burn.
CFO Watchpoints
What to watch
Defense
Defensibility
For a Lab this risk is inverted into capital concentration risk, because before it's commercial the funder is the customer. One or two patient backers often account for 80%+ of whether the company stays viable.
There are few direct competitors. The threat comes from adjacent labs and corporate R&D arms (Google Research, Meta AI, pharma), which can leapfrog a Lab with far more capital and talent. It isn't an even fight, since corporate R&D doesn't have to commercialize. The only response that holds up over time is to go deeper than they do.
The moat is IP, tacit knowledge and talent density (a lot of top people on a small team) working together. Patents are the part you can see, but the deeper moat is a team that knows things no one else knows, with methods and judgment built up over years. The risk is losing that moat when key researchers leave.
Before a Lab is commercial, switching cost doesn't apply. Once licensing begins, it's high if the IP is foundational.
Financial Impact
The money
Impact
A Lab's economics are all-or-nothing. Early on, cash burn is bigger than revenue and R&D CapEx (spending on equipment and other long-lived assets) runs high, so you need a long runway. In the end a Lab either produces a breakthrough that justifies decades of burn, or it ends in a write-down.
Target P&L Allocation
| Category | Target |
|---|---|
| Client Acquisition / Ad Spend | 32% |
| Total Labor | 23% |
| Fixed Overheads & G&A | 22% |
| Legal & Professional | 5% |
| Operating Income (EBITDA) | 10-18% |
Negative early; must raise or pre-sell IP. Positive only at commercialization.
Source: Compass archetype one-pager, target allocation as a percent of revenue
Engineers and scientists are on salary, and milestone-based equity is where most of their upside comes from. Cash pay is set high enough to take away financial anxiety, even if it's below FAANG rates, while the equity vests against scientific outputs (patents, prototypes, papers) rather than time alone. Senior researchers often hold 1-5% equity stakes, because they are the moat.
Strategic BD leads get a high base salary, plus long-tail bonuses tied to commercialized patents and licensing deals. Commission doesn't work here, because the deals are infrequent and multi-year and each one is bespoke, so a rep on commission would push to close too early. We'd pay the bonuses out over the lifetime of the royalty stream instead of at deal close.
A Lab's revenue is mostly grants and licensing advances, or partnership payments from sophisticated counterparties, so refund and chargeback risk is negligible.
Labs run on multi-year capital instead of operating cash flow, so the working capital need is massive and chronic.
Revenue Quality Score
2 / 10. When there is revenue, it's lumpy and back-loaded and often subsidized, so it isn't a meaningful operating metric until commercialization.
Trajectory
Where this goes
The amount of capital available and the depth of the research talent pool cap growth, and both are global rather than set by the market. A Lab either produces a breakthrough that scales or it doesn't, and the ceiling plateaus hard at around 50-150 employees.
A Lab has low optionality (few other directions it can go), and it's hard to pivot. The specialized equipment, the researchers it recruited and its IP focus were all built for one purpose, so pivoting usually means starting over.
The most common outcome is acquisition by a strategic. The other paths are a spin-out around a specific breakthrough or a wind-down, and an IPO is rare. Exit values can be enormous when an asset commercializes, but most Labs return zero.
Common Founder Failure Modes