Why CVC-Backed Start-Ups Look Better Before the Deal
This chapter looks at start-ups before a corporate venture capital investment is made. The specific question is why CVC-backed companies appear more innovative than other VC-backed companies, and the practical value is learning not to mistake that pre-existing difference for an effect caused by corporate money. The comparison is the foundation for everything that follows: before asking what CVC changed, we need to know what CVC investors were already choosing.
The baseline is not an even playing field
The OECD study compares start-ups that received CVC with start-ups financed by other kinds of venture capital. That comparison is more useful than comparing CVC-backed companies with all young companies, because both groups have already entered a financing market. They are innovation- and growth-oriented firms seeking external capital, not a random sample of businesses with no common need for investment.
Even so, the two groups are not interchangeable. A corporate investor may be looking for a technology that can complement its own capabilities, reduce uncertainty about a new market, or fit a broader strategic direction. An independent VC may also care about technology and growth, but the corporate investor has an additional reason to inspect the start-up’s relationship to the parent company. Selection begins before the investment event, through screening, networks, signals, and the perceived usefulness of a particular technology.
The descriptive differences are substantial. In the OECD sample, 31% of CVC-financed start-ups filed at least one patent during the observation period, compared with 15% of start-ups financed by other types of VC. Among companies that filed patents, CVC-backed firms filed an average of 2.43 new patents per year, compared with 2.02 for other VC-backed firms. They also raised more money: average yearly VC funding was about USD 22.39 million for CVC-financed start-ups and USD 12.55 million for other VC-financed start-ups.
These figures do not say that CVC produced the difference. They say that the companies arriving at the CVC deal are already different on observable measures. They are more likely to have a technological record and more capital behind them. A later comparison that ignores this baseline would attribute the start-up’s prior trajectory to the corporate investor.
There is also a timing clue. The average CVC-financed start-up was 3.74 years old at its first CVC investment, while its first VC investment occurred at an average age of 2.64 years. More than half—57%—received CVC in their first funding event, but the remainder had other VC financing first. This matters because CVC can enter at different stages of development. A company receiving a corporate investor after it has already built a patent portfolio is not in the same position as a company receiving corporate money to develop its first prototype.
The OECD analysis addresses this by comparing CVC and other VC deals at similar funding rounds and controlling for factors such as country, industry, year, founding cohort, and funding stage. Controls do not make the two groups identical. They make the comparison more disciplined by reducing some obvious differences. The remaining question is whether a measurable difference in prior innovation predicts the probability of receiving CVC.
The comparison also needs careful denominators. The 31% and 15% figures describe the share of each financing group that filed at least one patent over the observation period; they do not say that a CVC-backed company filed twice as many patents, nor that every CVC investment went to a patent holder. The 2.43 and 2.02 figures apply only after conditioning on patenting. Keeping those populations separate prevents a common statistical shortcut: treating a higher share of patenting firms as if it were a uniform increase in every firm’s output.
The timing of the first investment creates another selection issue. A start-up that receives CVC in its first round is exposed to a corporate investor at a different point from one that first develops with independent VC and later adds CVC. The latter may have already learned which technology to pursue, built relationships, and passed several market tests. Any claim about CVC’s effect must therefore say which entry point it studies, rather than treating all CVC events as one identical treatment.
Patents make a start-up legible to a corporate investor
The study’s deal-level analysis finds that patent-related signals are associated with a higher likelihood of CVC financing. In the broad sample, having a patent was associated with a 2.01 percentage-point increase in the probability of receiving CVC. The unconditional probability in that specification was about 4%, so the association represented roughly a 50% premium relative to the baseline probability. Among start-ups that patented, a larger patent stock was also associated with a higher probability of CVC, as was a citation-weighted patent stock.
A patent is useful here not because it is a complete definition of innovation, but because it can reduce uncertainty. A corporate investor often evaluates a technology from outside its existing operating structure. A patent portfolio provides an observable signal that a start-up has generated, documented, and claimed inventions. Citations add a rough indication that earlier inventions have been noticed or used in subsequent patenting. Funding history supplies another signal: a start-up that has already attracted investors and raised substantial capital may appear more capable of reaching a market.
Signals are especially valuable when the parties know less about one another. The study finds that the association between patenting and CVC financing is stronger for cross-industry and cross-country investments than for deals within the same industry or country. In a different industry, the corporate investor may have less shared language, fewer common contacts, and less direct experience with the start-up’s market. Across borders, distance and unfamiliar institutions can add to the uncertainty. A documented technological asset helps make the start-up assessable when personal networks and local knowledge are weaker.
This does not mean that patents are always the best investments, or that a start-up without patents lacks innovation. The study itself treats patents as an imperfect proxy. Some technologies are protected through secrecy, speed, data, customer relationships, or execution rather than patents. Software patenting also differs across jurisdictions and sectors. A patent signal is therefore a tool used in an investment decision, not a final measurement of the start-up’s inventive capacity.
There is a second reason not to read the signal too literally: a patent can reflect both the underlying technology and the decision to disclose it. Two equally capable start-ups may follow different intellectual-property strategies. One may build a large public portfolio because patents help attract financing; another may protect the same kind of advantage through secrecy or rapid product cycles. The OECD result remains informative about how CVC investors screen visible evidence, but it is not a complete census of technological quality.
The reported 2.01 percentage-point association is meaningful relative to a 4% average probability, yet it should not be mistaken for a guarantee. A start-up with a patent still has a low absolute probability of receiving CVC in the pooled deal sample. The estimate describes a change in likelihood after the model accounts for its listed controls; it does not identify every unmeasured feature that makes a company attractive, such as founder reputation, private demonstrations, customer commitments, or a relationship developed through a corporate network.
The distinction is important for interpreting the 31% versus 15% contrast. The result can be read in two ways at once. It shows that CVC investors are more likely to choose start-ups with observable technological signals. It also shows that the group of CVC-backed start-ups contains a higher concentration of companies that were already patenting. It does not show that CVC caused them to patent more.
Selection is also about the pair, not only the start-up
CVC is not simply a contest in which the strongest start-up wins money from any available corporate investor. The identity of the investor matters. The OECD study therefore examines technological proximity between a start-up and a CVC parent. Proximity is estimated from overlap in the technological fields represented in their patent stocks. In plain terms, the question is whether the start-up and the parent work in related technological territory.
To test this, the study compares realised deals with fictitious pairings. Each actual deal is matched with several counterfactual combinations: the same corporate parent paired with other VC-backed patenting start-ups, and the same start-up paired with other CVC parents. The exercise is not a complete model of negotiation. It asks whether the characteristics of the actual pair are associated with deal formation more often than the characteristics of plausible alternatives.
The answer is yes for technological proximity. The reported coefficient is positive and statistically significant, meaning that CVC arms are more likely to invest in start-ups technologically closer to them than the comparison start-ups are. Larger patent stocks on both sides also increase the likelihood of a realised CVC pair. This supports a strategic interpretation of CVC: corporate investors are not only buying exposure to generic growth; they are often looking for a relationship in which the start-up’s technology can connect to the parent’s knowledge, products, or ecosystem.
The pattern is not simply “similarity is always better.” Deals across countries are less likely than domestic deals, but technological closeness makes cross-border pairing more plausible. Across industries, the raw probability is also lower, yet technological proximity softens part of that difficulty. A start-up and a corporation can be in different formal industries while still sharing a technological problem or complementary patent space. Conversely, companies in the same broad sector may be poor partners if their technologies, incentives, or markets do not connect.
This pair-level view introduces an important tension. Technological proximity can help the parties understand one another and realize synergies. It can also increase the chance that the start-up will be drawn toward the parent’s technological priorities. At the selection stage, proximity makes a deal easier to justify. After the deal, the same proximity may shape which experiments look useful and which appear too far from the corporate ecosystem. The selection evidence does not prove that this narrowing occurs, but it identifies the relationship in which the question becomes most important.
It also explains why “strategic fit” should not be reduced to industry labels. The OECD analysis finds that cross-industry deals are common in the broader CVC activity, while the technological relationship between the parties still matters. A corporate investor may operate in one formal sector and invest in a start-up classified in another because the underlying technology is transferable or complementary. Conversely, being in the same sector does not guarantee that the parent can help: the start-up may be a direct substitute, compete for the same talent, or have a product that cannot be integrated without surrendering control.
For an investor, the pair is attractive when the start-up’s technology is close enough to understand and useful enough to matter, but not so redundant that the investment adds little. For a start-up, the same proximity raises a bargaining question. Does the parent understand the technology well enough to accelerate it, or does it understand it well enough to constrain it? The selection data cannot answer that question by themselves. They show why the parties meet; governance and post-investment outcomes must show what the meeting does.
What the comparison can—and cannot—tell us
The selection evidence changes the wording of the central claim. It is accurate to say that CVC-backed start-ups in this sample had stronger technological signals than other VC-backed start-ups before investment. It is not accurate to turn that into “CVC-backed start-ups are more innovative because CVC makes them so.” The first statement concerns who received the money. The second concerns what the money caused.
A credible post-CVC claim therefore needs a counterfactual: what would have happened to a similar start-up, at a similar stage and time, if it had received other VC rather than CVC? The comparison does not need to be perfect to be informative, but the analysis must address why the chosen comparison is relevant. That is why the OECD study aligns funding cohorts and treats the first funding event carefully in its later event-study analysis.
The distinction also protects against a subtler error. Suppose CVC-backed start-ups receive more citations after investment. That could mean the corporate relationship increased the use or visibility of their technology. But it could also partly reflect the fact that CVC chose firms whose technologies were already promising. The post-investment design must examine timing and pre-existing trends before assigning the change to the investment.
The same caution applies to funding amounts. CVC-backed start-ups raised more money in the descriptive data. That may indicate that the market sees them as promising, or that corporate participation helps mobilize further capital. It does not, by itself, identify which actor created the advantage or whether the additional capital altered the innovation path.
For founders, the practical lesson is not to reject corporate investors because selection exists. It is to recognize what the investor is selecting for. A CVC that wants a technology close to its current capabilities may be well suited to a start-up seeking technical integration or commercial reach. It may be less suited to a start-up whose value depends on remaining independent from the incumbent’s roadmap. The start-up should ask whether the corporate investor is attracted to the company’s existing strengths, its possible complementarity, or a future direction that the investor expects to influence.
For readers evaluating evidence, the lesson is equally concrete. Start with three questions: What was true before the investment? What is the comparison group? What changed after the investment, and when? Without those questions, a corporate investment can receive credit for selecting a strong company, and a strong company can be judged as proof of a universal financing effect.
The baseline is now clear: CVC-backed start-ups look better before the deal because corporate investors select for technological and financial signals, and because particular start-up–parent pairs appear strategically compatible. The next question is harder and more revealing. Once the investment is made, do these promising companies file more patents, produce more influential technology, or redirect their effort toward a different kind of value?