The Valley of Death Between R&D and Sales
For SBIR-funded technologies, the important problem after a promising research result is often not a single technical failure. It is the handoff from federally supported R&D to the outside capital, customer, production, or procurement decision that must carry the work forward. This chapter explains why that handoff can break down, so readers can diagnose an award-to-market gap as a combination of timing, capability, financing, and demand constraints rather than as simple evidence that the underlying technology failed.
Why technical promise and commercial viability diverge
A technology can be promising in a laboratory or agency program and still be unready for a buyer. Technical feasibility asks whether an approach can work. Commercial viability asks whether someone can obtain, integrate, pay for, and continue using the result under real conditions. Those questions overlap, but they are not interchangeable.
The difference begins with the conditions of proof. Research can be conducted with a carefully controlled setup, a small number of units, specialized personnel, or a flexible schedule. A customer may need the technology to work repeatedly, fit an existing process, arrive on time, and be supported after installation. A government acquisition office may need a clear procurement path and confidence that the technology can be integrated into an operational program. An investor may need to understand how much additional capital is required before revenue and what happens if development takes longer than expected. The same technical result can look encouraging to one decision maker and insufficient to another because each is responsible for a different risk.
This is why moving from Phase I to Phase II, or completing Phase II work, does not close the commercial question. Phase I can reduce uncertainty about whether an idea is feasible. Phase II can make the design more concrete through continued R&D or prototyping. But the next party may still need evidence about repeatability, scale-up, operating performance, cost, customer fit, or long-term support. A project may be technically ahead of its financing and market position.
The distinction is not an argument for demanding a finished product before funding research. If every uncertainty had to be resolved before public support began, much of the work SBIR is meant to stimulate would never start. The point is to identify which uncertainties the award addresses and which it leaves for later. An award is useful partly because it moves a project to the next question. It is not proof that all subsequent questions have already been answered.
How the handoff becomes a funding and capability gap
SBIR funding can end before a promising technology has attracted significant non-SBIR government contracts or private capital to license and commercialize it. At that point, the firm may have a result worth pursuing but no immediate source of money for the work between the result and use. That work can include additional engineering, performance testing, production preparation, customer-specific adaptation, or the practical steps needed to turn a demonstration into something another organization can depend on.
Timing makes the gap sharper. Research funding is often committed for a defined period, while customer and investor decisions may occur only after the technology has been tested in a setting that the award did not fully cover. A potential customer may want a demonstration before signing a contract. An investor may want evidence of customer interest before committing capital. The company may need the investment to build the demonstration. Each party can be acting rationally while the project remains stuck between their requirements.
The problem can become a coordination loop. Without production capacity, the firm cannot deliver enough units to test demand. Without a customer commitment, it may be difficult to finance that capacity. Without an operating installation, the customer cannot judge reliability or integration cost. Without evidence of reliable demand, a private investor may regard the remaining technical and market risks as too high. Public R&D money can interrupt this loop, but a normal research award is not designed to supply every form of transition capital or to compel the later parties to participate.
Small businesses are particularly exposed because they may have limited reserves, few parallel products, and little ability to keep a specialized team in place while waiting for the next commitment. A delay can mean losing employees who hold critical technical knowledge, pausing experiments, or shifting the project toward work that can be funded more immediately. The consequence is not necessarily that the technology was unsound. It may be that the firm could not survive the interval between one source of support and the next.
GAO uses the phrase “valley of death” to describe this vulnerability. The phrase should be understood as a financing and transition condition, not as a claim that every project follows the same failure pattern. Some projects may need more technical work; others may need a customer, an investor, a government contract, or production capacity. Some may find a different application or source of support. The common feature is that technical promise alone does not supply the resources and commitments required for the next stage.
Why long timelines and niche markets change the judgment
The length of the development cycle matters because commercial evidence arrives on a different schedule from technical evidence. GAO reports that achieving commercialization, especially for high-tech products, may take many years; one awardee said full development of its product was expected to take 20 years. A project observed early in that timeline should not be judged as though a mature product ought already to have generated a large revenue stream. At the same time, a long timeline should not be used as a blanket excuse for never testing whether the technology is progressing toward a user.
The right question is whether the project is resolving the next relevant uncertainty at a reasonable stage. Early on, that may mean demonstrating a technical principle. Later, it may mean producing a repeatable prototype, testing an application with potential customers, securing outside investment, or establishing a production method. A long cycle can justify patience about the final outcome while increasing the importance of intermediate evidence. Without that evidence, “commercialization is still ahead” becomes impossible to distinguish from indefinite postponement.
Niche markets introduce another complication. A technology serving nuclear security, advanced scientific instrumentation, or another specialized mission may have substantial value without a large commercial market. The number of customers can be small because the need is specialized, not because the technology has no use. A government buyer may be more important than a broad private market. In such a case, sales volume alone is a poor summary of value, but a credible procurement or deployment relationship remains relevant evidence of use.
This creates a two-part discipline. Do not demand mass-market scale from a technology whose purpose is specialized. Do not call the technology commercialized merely because its purpose is important. Instead, identify the relevant user and the outcome that matters: a government deployment, a license, a completed installation, a recurring order, a capability adopted by a mission program, or another concrete form of use. The standard should be appropriate to the technology, but it should still be a standard.
Production scale-up can expose the same mismatch. A research team may be able to make a small number of high-performing units by relying on custom labor or carefully selected inputs. A buyer may need consistent output at a larger volume. GAO reports that some CAPP awardees said additional funding helped them scale production and increase their capacity to make sales, while others were still seeking funding to advance their technologies enough to make substantial sales. The distinction is revealing: capacity to sell is an important transition step, but it is not the same as evidence that a durable market has arrived.
The possible bridges are not equivalent
A project can cross the gap through several routes, and each route answers a different question. Non-SBIR federal funding may provide time for related R&D or help a government mission continue development. Private investment can supply capital and signal that an outside party sees potential. Customer testing can reveal whether the technology works in a real setting and whether the proposed use matters. A government contract can show that an agency has chosen to procure or integrate the result. Internal adaptation can move the firm toward a different application or product.
None of these bridges should be treated as a universal substitute for the others. New investment is not a sale. A customer trial is not repeat procurement. A government contract is evidence of a particular public use, not automatically of broad private adoption. A technology may produce a patent or expertise without reaching any buyer, while still creating value that later supports another product. The bridge must be named before its strength can be judged.
The alternatives also involve different timing and ownership consequences. A new investor may accelerate development but require an ownership share. A federal program may preserve technical continuity but impose eligibility or reporting conditions. A customer-funded project may provide unusually strong use-case evidence but pull the firm toward one buyer’s requirements. A company may choose a related application because it has a clearer path to revenue, even though that means the original award’s technology is not commercialized exactly as first proposed. These are transition choices, not simple scores on a success scale.
The absence of one bridge does not prove that all bridges are absent. An unsuccessful application for one source of funding may be followed by another federal award, private investment, customer testing, or a change in product direction. GAO found that several unsuccessful CAPP applicants continued working toward commercialization, and some made sales or received non-SBIR investments after their second Phase II awards. But such evidence must remain specific: it shows continued progress for reported applicants, not a general success rate for all SBIR technologies.
Diagnosing what the record really says
“No sales yet” is a symptom, not a diagnosis. It can mean that the technology is still technically immature, that the customer market is narrow, that the production system is not ready, that outside funding arrived too late, that the relevant value is scientific or public rather than commercial, or simply that the available databases do not capture the outcome. Those explanations have different implications and require different evidence.
A disciplined diagnosis asks four linked questions. First, what technical work was completed, and what uncertainty did it reduce? Second, what external party was expected to carry the project forward—an investor, a customer, a procurement office, a licensee, or another public program? Third, what commitment or use actually occurred, and when? Fourth, is the observed result attributable to this project, or could broader market conditions and other firm characteristics explain it?
Evidence that answers only the first question supports a technical-progress claim. Evidence that answers the first and second may support a transition-readiness claim. Evidence of a customer test, outside investment, initial sale, production scale-up, or government contract supports a stronger but still specific claim. Evidence of repeat use, sustained procurement, or economics that keep the business operating supports a more durable commercialization claim. The categories are not a rigid universal ladder, but they prevent a single encouraging event from doing the work of several missing events.
GAO’s evidence is useful precisely because it does not pretend that the diagnosis is complete. Some CAPP awardees reported initial sales, new investments, greater production capacity, and employee gains. Some also said they were still seeking additional funds to make substantial sales. Some unsuccessful applicants continued without CAPP, while others said the missing award delayed testing, specialization for a customer, or employee retention. The reports indicate that additional runway may matter. They do not isolate the effect of CAPP from market conditions, prior differences among firms, or the ordinary passage of time.
The figure’s important point is not that every project follows one fixed sequence. It is that an SBIR awardee eventually depends on a decision outside the original R&D award. That decision may come from a private investor, a government customer, another program, or a market user. Until the relevant party commits and the technology is used, the project remains somewhere in the transition rather than at its destination.
The next policy question follows naturally. If the gap is real, should government simply provide more R&D money, or should it attach conditions that test whether a project has a plausible path to outside support? The Commercialization Assistance Pilot Program was designed around that tension: it offered eligible firms additional continuation funding while requiring signals intended to connect the work to commercialization.