A machine can do something extraordinary in a laboratory. Whether a customer can install it, rely on it, and pay enough for it to support a growing company is a different question.
That distinction sits at the heart of deep-tech investing. Proving a technology means showing that it can perform as intended. Proving a business means demonstrating that customers need the solution, will adopt and pay for it, and can be served on terms that support a sustainable company. Both matter, and evidence for each develops over time. [1]
What the evidence can tell you
Laboratory results can show that an idea works under controlled conditions. A prototype can demonstrate how components work together. Independent testing can strengthen confidence in a specific performance claim, especially when the test conditions and comparison are clear.
A customer pilot puts the technology to work in a setting closer to its intended use. Repeated deployments can show whether it performs across installations. Each step adds context, but none answers every question. A successful test on one machine, for example, does not establish performance across an entire factory. [2]
Technical and commercial progress also overlap. A company might speak with customers before building its first prototype. A pilot can reveal both a technical problem and a feature customers need. The team may return to the design, try again, and discover a better application along the way. There is no universal sequence or timetable.
What changes in the real world
A product has to fit a customer’s operation, not just a demonstration. An industrial sensor may need to work around heat and vibration, connect to existing equipment, and produce information maintenance teams can use. If installation interrupts production, the customer must weigh that disruption against the expected benefit.
The company also has to deliver more than one working unit. Depending on the product, it may need consistent manufacturing, dependable suppliers, customer-specific qualification—testing that it meets required specifications—or regulatory approval, and a plan to install, maintain, and repair equipment. Software may require integration with existing systems and continued support. Each requirement can affect cost and adoption. NIST identifies consistency and qualification as distinct challenges in additive manufacturing, even where individual high-quality parts can be produced. [3]
The relevant question is not simply whether a product performs. It is whether it performs reliably enough, at an acceptable total cost, for the people who must use and maintain it.
Interest is not the same as a purchase
A potential customer expressing interest shows that a problem may be worth exploring. A pilot shows a willingness to test a solution; it may be paid or unpaid and may have no commitment beyond the trial. A paid pilot shows willingness to pay for the test, but it does not necessarily mean the customer has agreed to buy the product afterward.
An agreement for a paid deployment can be stronger evidence of a purchasing decision, but its terms matter: Is the order conditional, cancellable, discounted, or dependent on meeting performance targets? Has the customer actually used and paid for the product?
Repeat purchases, renewals, and orders from additional customers can provide stronger evidence of sustained demand. Even then, investors need to understand margins, service costs, and concentration. One enthusiastic customer can be valuable evidence without proving a broad market.
A hypothetical example: A factory sensor
Imagine a company developing a sensor and software to identify signs of motor failure before a factory experiences an unexpected shutdown. The company believes earlier warnings could help maintenance teams plan repairs.
Laboratory and prototype. Its prototype identifies fault patterns on a test bench, and an independent lab reproduces the result under specified conditions. The investor’s questions shift from “Can the approach detect these signals?” to “Will it work on different motors, and does it provide enough warning to be useful?”
Customer pilot. A factory agrees to a trial. The team compares alerts with the factory’s existing maintenance approach and tracks installation time and false alarms. Vibration from nearby machinery triggers misleading alerts, so the company revises its software. The question becomes: “Can the improved system perform reliably in this customer’s environment, and does it deliver enough value to justify adoption?”
Paid deployment. The factory signs a contract to install the system on additional motors, subject to agreed conditions. That is evidence of a purchasing decision. The investor now examines the contract, installation and support costs, and whether the company can manufacture the devices consistently.
Repeat use. The first customer buys an installation at a second site, and another manufacturer purchases the system after its own evaluation. Demand looks more promising, but questions remain: Can the company serve different sites without extensive customization? Can it maintain performance and earn attractive margins as it grows?
This example is illustrative. Another company might secure a paid agreement earlier, spend longer in testing, or find that a setback changes its target market.
A framework for assessing progress
The stages below are a guide to the types of evidence investors may encounter, not a fixed pathway. Technical and commercial progress can advance at different speeds.
| Illustrative stage | Technical evidence | Commercial evidence | Remaining risks |
|---|---|---|---|
| Lab research | A result under controlled conditions | Customer discovery and problem interviews | Reproducibility; relevance to a real need |
| Prototype and testing | Components work together; claims may be independently tested | Defined use case; potential customer interest | Performance in operating conditions; adoption costs |
| Customer pilot | Performance measured in a user’s environment | A customer commits time or money to a trial | Integration, reliability, conversion to a broader purchase |
| Repeat deployments | Performance across installations and over time | Contracts, renewals, or additional orders | Margins, manufacturing and support capacity, competition |
Connect milestones to capital
A milestone is useful when it answers a consequential question. For a company preparing its first customer deployment, that might mean completing a defined operating test with agreed measures of performance, cost, and reliability—not simply installing the product.
An investor can then ask: How much time and capital are needed to reach a decision-useful result? Does the budget cover manufacturing, installation, qualification, support, and a reasonable allowance for delays or redesign? What funding is already available, and will the company need to raise more before the results are known? If so, future financing terms and dilution matter.
A milestone can resolve one uncertainty while revealing another. A successful pilot may show that the product works, then expose a costly installation process. A signed contract may confirm demand, then reveal a supply constraint. Progress makes the next decision better informed; it does not make the investment safe.
At BlueLake.vc, our focus is applied deep tech: companies working to turn scientific and engineering advances into commercial possibility. Understanding the distance between a promising result and a durable business is part of understanding the opportunity. [4]
Sources & further reading
- U.S. Department of Energy: Technology Readiness Levels
- U.S. Department of Energy: Adoption Readiness Levels
- NIST: Additive manufacturing FAQs
- BlueLake.vc: Investment focus
Hypothetical examples are illustrative. Technical and commercial progress can follow different paths, and investment outcomes remain uncertain.
Continue learning
Deep Tech Investing: The Opportunities and the RisksFor educational purposes. This material is not investment advice or an offer to invest.
