Chapter 11: Empirical Case Studies: Green Technology Innovation and Venture Finance

In 2023, a subnational green fund in a middle-income province reported a gross multiple of 0.6x on a portfolio of 58 companies. The number matters less than how the fund got there: 31 of its 58 investments had been declined by at least one private venture fund in the twelve months before the public fund wrote its check. The investment committee read those declines as evidence of unmet need. By 2023, 22 portfolio companies had ceased operations. The provincial government recapitalized the fund in 2022 and attached a governance reform to the money—a private majority on the investment committee and a requirement that every deal carry a private lead investor.

That episode frames the problem this chapter takes up. In agriculture and forestry, the previous chapter found a two-channel mechanism in which rating and insurance interact and catalysis is missing because the sector does not attract venture finance. Move to firms developing solid-state batteries, electrolyzers, industrial heat pumps, or microbial crop protection, and the financing problem changes shape. The assets are intangible-heavy, revenue is back-loaded, technology risk is real but diversifiable across a portfolio, and the returns, when they arrive, can be large. Venture capital and sci-tech finance are the instruments built for this terrain. Whether they actually reach green technology firms, and under what conditions, is an empirical question.

Three case clusters organize the evidence. Private venture capital in green technology: who invests, at what stage, in what instruments, with what results. Government guiding funds, the policy instrument through which many governments attempt to seed or accelerate a green venture market. And the cross-cutting question of what separates the ventures and instruments that work from those that do not. The argument is that venture finance is the catalytic channel in the triad, but catalysis is conditional. It operates where a verifiable technical milestone structure exists, where a risk-buffering layer absorbs early-stage downside, and where a policy fund is designed to crowd in private capital. Remove any one and the channel stalls; the firm falls back on the credit channel, which fits its asset structure badly.

11.1 Venture Capital in Green Technology: The Shape of the Market

11.1.1 What Makes Green Tech a Distinct Venture Asset Class

A software firm can reach a working prototype on a few million dollars. A battery-chemistry firm typically needs $20–50 million before a pilot line produces cells at spec. That gap in capital intensity per unit of technical progress is the first thing that separates green technology from conventional venture categories, and it drives most of what follows.

The second is time. In a sample of 74 deep-tech green ventures drawn from Crunchbase and PitchBook records for 2015–2023 and cross-checked against company filings, the median firm took 6.2 years from first institutional round to first material commercial revenue. A matched sample of enterprise software firms took 2.8 years.

Technology risk in this sector does not resolve monotonically. A chemistry that works in the lab can fail at scale for reasons that appear only at scale, and the failure is often uninformative about adjacent approaches—which means a missed milestone cannot be read as a verdict on the underlying science. Exits, finally, are shaped by regulation and by incumbents. Acquirers are frequently utilities, chemical majors, or automakers whose willingness to pay depends on compliance timelines they do not control.

These features push the financing structure in a particular direction. Capital requirement and long horizon together imply that no single round carries a firm to commercial viability; the firm needs a sequence of rounds, each conditioned on a technical milestone. Non-monotonic risk implies that milestone contracts have to be written with care, because a failed milestone is not always a signal of a failed approach. Regulatory dependence of exits implies that investors underwrite policy scenarios alongside technology scenarios.

Table 11-1 sets out the stage structure, with typical check sizes and the milestone each stage is designed to test.

Table 11-1: Stage Structure of Green Technology Venture Financing

The table shows where the catalytic channel breaks. Between Series A and Series C the capital requirement jumps by an order of magnitude and the diligence focus shifts from technical to manufacturing—precisely where specialist venture capital runs out of mandate. A $300 million deep-tech fund cannot write a $60 million Series C without concentrating its portfolio beyond prudence. The green finance literature calls this the valley of death. It is a structural gap in the stage structure, not a temporary shortage of risk appetite.

11.1.2 The Investment Record: What the Data Show

Aggregate figures on green-tech venture investment get quoted widely and reconciled rarely. The series below was assembled from PitchBook deal-level records, filtered to firms whose primary product is a climate or environmental technology, and cross-checked against the Cleantech Group i3 database for the earlier years. Table 11-2 reports global green-tech venture investment by stage band for 2013–2023, in billions of USD.

Table 11-2: Global Green Technology Venture Investment by Stage Band, 2013–2023 (USD billions, annual average within period)

Three patterns stand out. Early-stage investment was resilient through the 2022–2023 correction: seed and Series A volumes grew slightly in deal count while falling in average size. Growth-stage investment was cyclical: Series C and later rounds contracted by roughly 40 percent from the 2021 peak, and the contraction was concentrated in capital-intensive hardware—the segment where the financing gap is widest. And the market is geographically concentrated. The United States, China, and the European Union accounted for approximately 88 percent of green-tech venture investment over the period; the remainder split among Israel, Japan, India, and a handful of others.

Aggregates conceal more than they reveal, so the rest of this section works through firm-level cases spanning the stage structure and four technology categories: battery materials, grid software, industrial heat, and biological inputs. Each case traces the financing sequence, the milestone structure, and the outcome to date.

11.1.3 Case A: A Battery-Materials Firm and the Milestone Contract

Novacell (a pseudonym; the underlying facts come from public filings and two interviews with the founding team) was founded in 2014 to commercialize a cathode chemistry developed at a European university. The technical claim: a 20 percent reduction in cobalt content at equal energy density, with a synthesis route that could run on existing cathode plants. The claim mattered because cobalt supply is concentrated and price-volatile, and because retrofitting a cathode plant costs less than building one.

The financing sequence ran as follows. A seed round of €2.1 million in 2015, led by a university-affiliated fund with participation from a chemical company's venture arm, funded the reproducibility work. A Series A of €9 million in 2017, led by a specialist deep-tech fund, funded a bench-scale pilot and first customer qualification with two cell makers. A Series B of €28 million in 2019, led by a syndicate of two deep-tech funds and a strategic investor, funded a 2,000-tonne-per-year pilot line. A Series C of €75 million in 2021, led by a growth fund with a sovereign wealth fund as co-investor, funded the first commercial line.

The milestone contracts repay close reading. In the Series A term sheet, tranches were conditioned on three technical milestones: reproduction of the energy-density claim by an independent laboratory, 100 cycles at 80 percent capacity retention in a full cell, and a signed qualification agreement with at least one cell maker. The first two are technical; the third is commercial. The Series B term sheet added a manufacturing milestone: the pilot line must reach 90 percent of nameplate capacity with first-pass yield above 85 percent within 18 months. The Series C term sheet replaced technical milestones with offtake milestones: at least two binding offtake agreements covering 60 percent of first-line capacity.

This staging maps onto what Kaplan and Strömberg (2003) found in their analysis of 213 venture capital contracts: financing is staged, and each stage is conditioned on observable proxies for progress. What is distinctive in the green-tech case is the weight placed on third-party verification. The independent laboratory test and the qualification agreement in the Series A are both devices for converting a technical claim into a signal a non-specialist investor can underwrite. In software the analogous signal is a usage metric; in green hardware it is a test report.

The outcome to date is partial success. The pilot line reached 87 percent of nameplate capacity, just short of the Series B milestone, and the tranche was renegotiated rather than called. The first commercial line started production in 2023, 14 months behind the Series C plan. The firm is not yet profitable. Its most recent round, in 2024, was a down round at roughly 0.8x the 2021 valuation. The milestone structure absorbed a shortfall without triggering liquidation because investors and firm could renegotiate against a shared technical model. Where that shared model is absent, the same shortfall produces a write-off—a pattern that recurs in the failure analysis in section 11.3.5.

11.1.4 Case B: Grid Software and the Faster Path

GridSense (a pseudonym) was founded in 2016 to sell software that forecasts distribution-network constraints and schedules flexible demand around them. The technology is software, the capital requirement is modest, and the revenue model is a subscription per substation. The financing sequence was correspondingly short: a seed round of $1.5 million in 2017, a Series A of $8 million in 2019, and a Series B of $22 million in 2021. The Series B was led by a growth fund, and the firm reached cash-flow breakeven in 2023.

GridSense's milestones were commercial rather than technical: substations under contract, annual recurring revenue, net revenue retention. Technical risk was low enough that investors could underwrite the firm on standard SaaS metrics. The green attribute of the product—it enables higher renewable penetration on constrained networks—was relevant to customers (distribution utilities under renewable-integration mandates) but not to the structure of the financing.

The pattern generalizes. Green-tech ventures whose product is software or a service layered on existing assets finance like conventional ventures. Ventures whose product is a physical material, a device, or a process finance like deep-tech ventures, with longer timelines, larger rounds, and more third-party verification. The catalytic channel is not one channel. Software green ventures need the standard venture ecosystem; hardware green ventures need a deep-tech ecosystem with a public risk-buffering layer. Conflating the two produces policy that is too slow for the first and too thin for the second.

11.1.5 Case C: Industrial Heat and the Strategic Investor

HeatCore (a pseudonym) was founded in 2015 to build high-temperature heat pumps for industrial process heat. Industrial process heat accounts for roughly 20 percent of global final energy consumption, and much of it falls below 200°C—within the range of heat-pump technology. The capital requirement is moderate: a heat-pump assembly line costs less than a battery line. The financing sequence was seed ($3 million, 2016), Series A ($12 million, 2018), Series B ($35 million, 2020), and a strategic round ($50 million, 2022) led by an industrial conglomerate that also became the firm's largest customer.

The strategic round carried a valuation implying a 1.4x step-up from the Series B—below what a purely financial round would have commanded in 2022—in exchange for a supply agreement and a joint-development program. The founders accepted the lower valuation because the strategic investor supplied what the financial investors could not: a reference customer, a service network, and a manufacturing partner. In the venture literature this is complementary assets, and the lower valuation was the price of them.

Green hardware ventures often need complementary assets—manufacturing, service, offtake—that financial investors do not own. Corporate venture capital and strategic investment are the standard route to those assets. Whether public instruments should support strategic investment depends on whether the strategic investor's interests align with the public objective. At HeatCore they did: the conglomerate's decarbonization commitments made the heat pump a strategic product rather than a hedge. In other cases strategic investment can slow a venture by tying it to a single customer's roadmap. Dushnitsky and Lenox (2005), analyzing a panel of US public firms from 1990 to 1999, found that corporate venture capital is associated with higher firm value only where the investing corporation has a strategic—not purely financial—rationale; purely financial CVC programs showed no such association. The alignment question is therefore not a side issue. It determines whether the strategic route builds the venture or captures it.

11.1.6 Case D: Biological Inputs and the Regulatory Clock

BioNova (a pseudonym) develops microbial crop-protection products—a bacterial strain that suppresses a fungal pathogen in cereals. The capital requirement is moderate; the regulatory pathway is long. Registration of a biological pesticide in the European Union takes 3–5 years and costs $3–8 million per active substance. The EPA's biopesticide pathway in the United States is faster but still takes 2–3 years.

The financing sequence was seed ($2.5 million, 2017), Series A ($10 million, 2019), Series B ($30 million, 2021), and a bridge round ($15 million, 2023) while EU registration was pending. The bridge is the instructive part. The firm had met its technical milestones, but the registration clock had not run out, and investors were unwilling to price a Series C before the registration decision. The bridge was led by existing investors at a flat valuation, with warrant cover that would trigger on registration.

The regulatory clock is a systematic feature of green-tech venture finance in biological and chemical categories. It introduces a delay uncorrelated with technical progress, and no amount of capital compresses it. The financing structure that emerges is a bridge or extension round, priced flat or with warrant cover, that carries the firm through the wait. Public instruments can play a role here: a guarantee or matching facility that shares the regulatory-delay risk would reduce the cost of the bridge and prevent the dilution that otherwise falls on founders at the worst moment.

11.1.7 What the Four Cases Establish

Read together, the cases point to three propositions the rest of the chapter builds on.

The milestone contract is the core financing technology in green-tech venture capital. It converts a technical claim into a sequence of observable tests and allocates the risk of failure across rounds. It works when the milestones are third-party verifiable and when investors and firm share a technical model that allows renegotiation after a shortfall—the Novacell renegotiation and the BioNova bridge are the same mechanism applied to different shortfalls.

The stage structure has a gap at Series B and Series C, where the capital requirement jumps and the diligence focus shifts from technical to manufacturing. Specialist venture funds cannot fill the gap without concentrating their portfolios. Where it is filled, it is filled by growth funds, sovereign wealth funds, strategic investors, and public instruments.

The regulatory clock is a distinct risk factor in several green-tech categories, and capital cannot diversify it away. Financing structures that share the risk—bridge rounds with warrant cover, public guarantees—reduce the cost of carrying a firm through the wait.

These propositions frame the analysis of government guiding funds, the policy instrument through which many governments attempt to fill the Series B/C gap and to seed the earlier stages.

11.2 Government Guiding Funds in Action

11.2.1 The Logic and the Risk

A government guiding fund is a public or quasi-public fund that invests in venture capital funds or directly in firms, with the objective of developing a venture market or directing capital toward a policy priority. The instrument is widely used: the European Investment Fund's venture programs, Israel's Yozma, China's national and local guidance funds, Singapore's Temasek-backed venture vehicles, and a long list of national and subnational programs. The green variant directs it toward climate and environmental technologies.

The economic logic rests on two claims. The first is that the private venture market underinvests in early-stage green technology because the social return exceeds the private return—knowledge spillovers benefit competitors, some environmental outcomes have public-good characteristics, and building a deep-tech ecosystem is a coordination problem in which each investor waits for the others. The second claim is that a public fund can certify quality and thereby crowd in private capital. If the public fund conducts diligence and invests, private investors update their beliefs and follow.

The risk is that the second claim fails. A public fund that invests at a valuation private investors regard as too high, or on terms that subordinate private capital, or in firms private investors have already rejected, crowds out rather than crowds in. The empirical record on this question is mixed and worth stating precisely. Leleux and Surlemont (2003), analyzing European public venture programs through the 1990s, found crowding out in some countries and crowding in in others, with the difference traceable to fund design. Cumming and MacIntosh (2006), using Canadian data, found that public funds crowded out private equity in some market segments. Da Rin, Nicodano, and Sembenelli (2006), using cross-country data for 14 European countries, found that public venture capital was associated with the creation of active venture markets when the public funds were designed to be temporary and to co-invest with private lead investors. The design features that separate the two outcomes are summarized in Table 11-3.

Table 11-3: Design Features Associated with Crowding In versus Crowding Out

The green guiding funds that have worked conform to the left column. The ones that have not conform to the right. The cases below illustrate both.

11.2.2 Case E: A National Green Guiding Fund with a Sunset Clause

A Northern European country established a national green technology fund in 2013 with an initial capitalization of €1.2 billion and a mandate to invest in venture funds and directly in growth-stage green-tech firms. The fund was structured as a fund-of-funds with a co-investment facility, and it carried a sunset clause: no new commitments after 2023, full divestment by 2030.

Over its ten-year commitment period the fund committed €740 million to 23 venture funds and €410 million to 41 direct co-investments. In every direct co-investment a private investor led the round and set the terms; the public fund took the same instrument at the same price. In the fund commitments the public fund took a capped return—a 4 percent preferred return, with the excess carried by the private fund manager—which reduced the public fund's cost of capital without subsidizing the private investors' downside.

The most recent published report (2024) gives the following outcomes. The fund-of-funds portfolio had a gross multiple of 1.6x and an IRR of 11.2 percent, slightly below the private venture benchmark but positive. The direct co-investment portfolio had a gross multiple of 2.1x and an IRR of 14.8 percent, in line with the private benchmark. Portfolio companies attracted €4.3 of private capital for every €1 of public capital.

That leverage ratio is the number that matters for the crowding-in question, but it has to be read with care. A high ratio can reflect genuine certification—private investors following the public fund's diligence—or a rising market in which private capital was abundant and would have invested anyway. The evaluators addressed this by comparing the fund's co-investment deals to a matched sample of deals the fund considered and declined. The declined deals attracted less private capital and had lower survival rates, which is consistent with certification rather than pure market timing.

11.2.3 Case F: A Subnational Green Fund and the Selection Problem

The fund that opened this chapter deserves a closer look, because it isolates the selection mechanism. Established in 2016 in a large middle-income country with a capitalization of approximately $600 million, it invested directly, without a private lead, and its investment committee was chaired by a provincial official. Of its 58 direct investments, 31 were in firms declined by at least one private venture fund in the preceding twelve months. The committee treated those declines as evidence that the firms needed public support—the inverse of the certification logic. By 2023, 22 of the 58 portfolio companies had ceased operations and the fund's gross multiple was 0.6x.

The 2022 recapitalization came with a governance reform: a private majority on the investment committee and a requirement that every deal have a private lead investor. The reform is recent enough that its effects cannot yet be assessed. The fund's failure was not a failure of capital or of deal flow. It was a failure to use the private market's screening information. A public fund that treats private declines as a buy signal is not certifying quality; it is adverse-selecting.

11.2.4 Case G: China's National Guidance Fund and the Green Tech Allocation

China's national guidance fund system is the largest public venture instrument in the world by capitalization. It operates at national, provincial, and municipal levels, with total commitments the Ministry of Finance reported at approximately RMB 2.1 trillion by 2023. The green-tech allocation is not separately reported, but it can be estimated from the disclosed portfolios of the major funds.

The system's design has evolved. The early funds (2015–2017) were structured as direct investors with government-appointed management, and they exhibited the selection problem described in Case F. The later funds (2019 onward) increasingly adopted a fund-of-funds structure with private managers, a co-investment requirement, and a return-sharing mechanism giving the private manager a larger share of the upside. Green-tech funds within the system followed this evolution.

The evidence on green-tech performance is partial but informative. A 2023 evaluation by researchers at Tsinghua University's National Institute of Financial Research, using disclosed portfolio data for 14 provincial green funds, found that funds with private managers and co-investment requirements had portfolio survival rates comparable to the private benchmark, while funds with government-appointed management had survival rates roughly 20 percentage points lower. The same evaluation found the later funds were more likely to invest in hardware (batteries, solar, hydrogen) and less likely to invest in software, reflecting the provincial industrial base rather than a deliberate portfolio strategy.

The Chinese case matters beyond its size. It shows the guiding-fund instrument can be reformed within a system that retains a strong public role. The reform did not privatize the funds; it changed the governance and the deal structure. The crowding-in design features are not tied to a particular political economy. They are tied to the incentive structure of the fund manager and the investment committee.

11.2.5 Case H: The European Innovation Council Accelerator

The European Innovation Council (EIC) Accelerator, established in 2021 under Horizon Europe, provides blended finance: a grant component for technology development and an equity component for scale-up, with the equity co-invested alongside private investors. The equity component follows a private lead; the grant component is disbursed against technical milestones.

The instrument is too recent for a full outcome assessment, but the early data are informative. In its first two full calls (2021–2022) the Accelerator received approximately 4,000 proposals and funded 249, of which 71 were in green-tech categories. Funded green-tech firms raised a median of €12 million in private capital within 18 months of the award, against a median of €4 million for a matched sample of unfunded applicants. The grant component was associated with a higher probability of reaching the next technical milestone—consistent with the grant relaxing the financing constraint on pre-revenue technical work.

The design addresses the Series B/C gap directly. The grant carries the firm through pre-revenue technical work; the equity component, co-invested with a private lead, carries it into scale-up. The blended structure is more complex to administer than a pure fund, and the administrative cost is a real drawback. But the structure matches the risk profile of deep-tech green ventures, where early technical work has public-good characteristics and later scale-up has private returns.

11.2.6 What the Guiding-Fund Cases Establish

The crowding-in versus crowding-out distinction is real, and it is traceable to design features. The features that matter most are the lead/follow rule, the pricing rule, the selection rule, and the governance of the investment committee. A public fund that follows a private lead, prices on the same terms, selects from the private pipeline, and delegates deal approval to an independent committee with a private majority is likely to crowd in. A fund that inverts any of these rules is likely to crowd out.

Certification is the main channel through which crowding in operates. The public fund's diligence and investment signal quality to private investors, who then follow. The signal is credible only if the public fund's selection is informative, which requires that it not treat private declines as a buy signal.

The fund-of-funds structure with a co-investment facility is a workable template for green-tech guiding funds. It separates fund manager selection, which requires venture expertise, from deal selection, which requires technical expertise, and it lets the public fund leverage private diligence.

The blended grant-plus-equity structure addresses the deep-tech risk profile more directly than a pure equity fund. The grant component funds pre-revenue technical work, where the social return is highest and the private return most uncertain; the equity component funds scale-up, where the private return is clearer.

These conclusions set up the third section, which examines the success factors and challenges that cut across the cases.

11.3 Success Factors and Challenges

11.3.1 The Success Factors

Five factors recur across the cases. They are not independent; a venture that has four and lacks the fifth often fails.

A verifiable technical milestone structure. The milestone contract is the core financing technology, and it works only when milestones are third-party verifiable. Novacell's independent laboratory test and GridSense's substation-count metric are both verifiable. BioNova's registration clock is verifiable but not compressible, which is why the bridge round was needed. Public instruments should fund testing laboratories and certification bodies, not only the ventures.

A risk-buffering layer that absorbs early-stage downside. Venture investors in the cases were willing to take technical risk but not regulatory risk, policy risk, or the risk of an uninformative milestone shortfall. The instruments that absorbed those risks were the EIC Accelerator's grant component, the warrant cover in BioNova's bridge, and the renegotiation of Novacell's Series B tranche. Public instruments should target the risks private investors cannot diversify, not the risks they can.

A private lead investor with sector expertise. Every successful case had one, and that investor conducted technical diligence and set the terms while the public fund followed, certified, and filled the gap. Public instruments should require a private lead as a condition for the certification mechanism to operate, not as a procedural box.

Complementary assets from a strategic partner. HeatCore's strategic round supplied manufacturing, service, and offtake. Green hardware ventures that succeeded without a strategic partner built those assets in-house, which is slower and more capital-intensive. Public instruments should not penalize strategic investment, but should screen for alignment between the strategic investor's interests and the public objective.

A policy environment that makes the green attribute valuable. GridSense's customers bought the product because renewable-integration mandates made network constraints costly. Novacell's customers bought the cathode because cobalt price volatility and supply-concentration rules made cobalt reduction valuable. A green attribute is valuable to a venture only when a policy or a market price makes it valuable to the customer. Venture finance policy and environmental policy have to be coordinated; a green venture fund operating without carbon pricing or renewable mandates is swimming against the current.

11.3.2 The Challenges

The challenges mirror the success factors, plus two that are structural.

The Series B/C gap is the central structural challenge. The capital requirement jumps by an order of magnitude between Series A and Series C, and specialist venture funds cannot fill the gap without concentrating their portfolios. Where it is filled, it is filled by growth funds, sovereign wealth funds, strategic investors, and public instruments. The gap is widest in capital-intensive hardware, which is also where the emissions-reduction potential is largest.

The regulatory clock is the second. In biological and chemical categories the registration pathway introduces a delay uncorrelated with technical progress and incompressible by capital. The financing structure that emerges is a bridge round, which dilutes founders and strains investor patience. Public instruments that could share this risk—guarantees, matching facilities, regulatory-delay insurance—are underdeveloped.

The exit landscape is the third. Green hardware ventures are acquired by utilities, chemical majors, and automakers whose willingness to pay depends on compliance timelines they do not control. The IPO route is available for some categories (solar, batteries) but not others (biological inputs, industrial heat). Exit uncertainty raises the required return and narrows the set of fundable ventures. Public instruments should support exit routes—strategic buyer development, secondary markets—as well as entry.

The measurement problem is the fourth. The environmental impact of a green-tech venture is often uncertain ex ante and realized ex post. A battery-materials firm's impact depends on the emissions intensity of the grid charging the batteries; a heat-pump firm's impact depends on the emissions intensity of the heat it displaces. Tying public instruments to impact is therefore difficult, and there is a risk that public funds subsidize ventures green in label but not in effect. Public instruments should require impact measurement but should not use impact as a selection criterion in early stages, when the measurement is too noisy to be informative.

The policy-finance coordination problem is the fifth. The green attribute of a product is valuable only when environmental policy makes it valuable. A venture fund operating in a jurisdiction without carbon pricing, renewable mandates, or efficiency standards is financing products whose value depends on policy developments the fund cannot control. Green-tech venture markets cluster in jurisdictions with strong environmental policy for this reason. Venture finance policy should be sequenced with environmental policy, not treated as a substitute for it.

11.3.3 A Framework for Assessing Catalytic Effectiveness

The success factors and challenges can be assembled into a diagnostic sequence for assessing whether a public instrument is catalytically effective. Three tests, applied in order.

Additionality: does the instrument cause investment that would not have occurred otherwise? The leverage ratio is the standard metric, but it must be corrected for market timing, as in Case E. A better metric is the difference in private capital raised between funded and unfunded but comparable ventures, as in the EIC evaluation.

Certification: does the instrument's investment signal quality to private investors? The test is whether the instrument's portfolio outperforms a matched sample of declined deals, as in Case E, and whether the instrument's selection is informative—which requires that it not treat private declines as a buy signal, as in Case F.

Impact: does the instrument's portfolio deliver environmental outcomes that would not have occurred otherwise? The measurement problem makes this difficult, but it can be approached through ex post impact measurement combined with a counterfactual based on the emissions intensity of the displaced technology.

Figure 11-1 shows the sequence. An instrument that fails additionality is crowding out, and no amount of impact measurement fixes that. An instrument that passes additionality but fails certification is selecting ventures the private market would have funded anyway—public capital spent for no additional investment. An instrument that passes both but fails impact is subsidizing ventures green in label but not in effect, a reputational risk for the whole green finance agenda.

flowchart TD
    A[Public instrument] --> B{Additionality test}
    B -->|Fail| X[Crowding out]
    B -->|Pass| C{Certification test}
    C -->|Fail| Y[Adverse selection]
    C -->|Pass| D{Impact test}
    D -->|Fail| W[Green label without green effect]
    D -->|Pass| Z[Catalytic]

11.3.4 The Interaction with Rating and Insurance

The cases so far have treated the catalytic channel in isolation. The triad argument requires examining its interaction with the other two channels, and in green-tech venture finance that interaction is weaker than in agriculture and forestry—but not absent.

Rating interacts with venture finance through the debt component of a venture's capital structure. A green-tech venture that has reached commercial production typically carries a debt layer—equipment loans, working capital lines, project finance—alongside its equity. The rating of that debt depends on the venture's environmental performance, and a credible environmental rating can lower the cost of the debt layer, which improves the equity return and makes the venture more attractive to venture investors. Novacell's first commercial line was financed with a debt layer that included a sustainability-linked loan with a margin step-down tied to the emissions intensity of the production process. The step-down was 15 basis points—small, but a signal that the rating channel was operating.

Insurance interacts through the risk-buffering layer. Weather index insurance and forest insurance are irrelevant to a battery-materials firm, but the general principle of risk buffering is not. The instruments that buffered the ventures in the cases were grant components, warrant cover, and milestone renegotiation, not insurance contracts. The insurance industry has not developed products for the specific risks of green-tech ventures—technology performance risk, regulatory-delay risk—and this is a gap. A performance warranty or a regulatory-delay insurance product would be a genuine innovation, and it would strengthen the catalytic channel by absorbing risks venture investors currently bear.

The interaction among the three channels is therefore partial in green-tech venture finance, as it was in agriculture and forestry, but the partiality has a different shape. In agriculture and forestry the catalytic channel was absent and the interaction ran between rating and insurance. In green-tech venture finance the catalytic channel is present and the interaction runs between catalysis and a thin rating layer, with insurance largely absent. The integrated mechanism is a two-channel mechanism in both sectors, but the two channels are different. The integrated mechanism is not a single design that applies across sectors; it is a family of designs, each combining the channels relevant to the sector's asset structure. The policy framework has to accommodate that variation, which means it cannot be a single instrument or a single mandate. It must be a portfolio of instruments, sequenced according to the sector's stage of financial development.

11.3.5 Where the Catalytic Channel Breaks: Evidence from Failed Ventures

The success factors describe what works. Failure modes describe what happens when a factor is missing, and they are more informative for policy because they isolate the binding constraint. A sample of 31 green-tech ventures that raised at least one institutional round and subsequently ceased operations or were acquired below invested capital was assembled from PitchBook records and cross-checked against company announcements. The sample is small and subject to survivorship bias in the opposite direction from the usual one, but the patterns are consistent enough to report.

Four failure modes account for 27 of the 31 cases.

Milestone failure without renegotiation occurred in 9 cases. In each, the venture missed a technical milestone and the lead investor had no technical model to assess whether the miss was informative. The tranche was called, the firm ran out of cash, and the technology was shelved. The distinguishing feature was the absence of third-party technical assessment: the lead investor was a generalist fund that had invested on the strength of a syndicate partner's diligence, and when the syndicate partner withdrew, the generalist had no basis for renegotiation.

The financing gap occurred in 8 cases. These ventures met their technical milestones and had customer interest, but the next round required a check size larger than any fund in their syndicate could write, and no growth fund or strategic investor materialized. They were not technical failures; they were capital-structure failures. The gap was widest in capital-intensive hardware, and the failures clustered in the 2022–2023 correction, when growth-stage capital contracted.

Regulatory delay occurred in 6 cases, all in biological or chemical categories. These ventures had positive technical results and a viable product, but the registration clock outlasted investor patience or bridge financing. In 4 of the 6, the technology was licensed to a larger firm that completed the registration. Licensing is not a total loss,but it transfers value to the acquirer and reduces the return to the original investors, which in turn reduces their willingness to fund the next regulatory-clock venture.

Demand failure occurred in 4 cases. These ventures had a working product and a completed registration, but customers did not buy at the price the business model required. In each case the green attribute was not valuable to the customer because the policy or price that would have made it valuable had not materialized. A heat-pump venture failed when the industrial gas price fell and the carbon price did not rise; a biological-inputs venture failed when the retailer's procurement policy did not pass the premium through to the farmer. Demand failure is the clearest illustration of the policy-finance coordination problem: the venture was financeable, but the market was not there.

The four failure modes map onto the success factors directly. Milestone failure without renegotiation is the absence of verifiable milestones and of a private lead with sector expertise. The financing gap is the absence of a risk-buffering layer at Series B/C. Regulatory delay is the absence of an instrument that shares the regulatory-clock risk. Demand failure is the absence of a policy environment that makes the green attribute valuable.

11.3.6 The Cost of Capital and the Required Return

A quantitative way to see the binding constraint is to decompose the required return venture investors apply to green-tech ventures. The required return is not directly observable, but it can be inferred from the hurdle rates fund managers report in limited partnership agreements and from the discount rates implicit in round valuations.

For a sample of 22 deep-tech green funds with disclosed hurdle rates, the median hurdle rate was 22 percent net of fees, against 15 percent for a matched sample of generalist venture funds. The 7-point premium decomposes roughly as follows: 3 points for technology risk (the probability the core technology fails at scale), 2 points for longer time-to-exit (median holding period for green hardware was 8.4 years against 5.6 years for software), and 2 points for policy and regulatory uncertainty (the probability the policy environment changes in a way that affects the exit).

The decomposition tells us which risks a public instrument should target. A grant that funds pre-revenue technical work reduces the technology-risk premium by lowering cash burn and raising the probability of reaching the next milestone. A guarantee or matching facility that shares regulatory-delay risk reduces the policy-uncertainty premium. An instrument that extends the fund's holding period—a secondary-market facility, a longer-dated fund-of-funds commitment—reduces the time-to-exit premium. An instrument that simply lowers the cost of capital without changing the risk does not reduce the required return; it transfers the risk to the public balance sheet. That may be justified on social-return grounds, but it should be recognized as a subsidy rather than a market fix.

The required return can be written as:

where \(r_f\) is the risk-free rate, the \(\beta\) terms are factor loadings, and the \(\pi\) terms are risk premia. A public instrument's effect is to reduce one or more of the \(\pi\) terms, not \(r_f\). An instrument that reduces \(\pi_{tech}\) through grant funding targets the largest single component, which is why the grant component of the EIC Accelerator was associated with a measurable increase in milestone achievement.

11.3.7 The Portfolio Construction Problem

A venture fund investing in green technology faces a portfolio construction problem more acute than in generalist venture capital. The problem arises from correlated failures. In software, failures are largely independent; a fund with 30 portfolio companies can expect a predictable distribution of outcomes. In deep-tech green hardware, failures are correlated through shared technology platforms, shared supply chains, and shared policy exposures. A fund that invests in five battery-materials ventures is not diversified. It has one bet on battery materials with five draws.

The correlation has a quantitative consequence. In the standard venture portfolio model, the probability that a fund with \(n\) independent bets achieves at least one outcome above the fund's return threshold is \(1 - (1-p)^n\), where \(p\) is the probability a single venture achieves the threshold. For \(p = 0.1\) and \(n = 30\), the probability is 0.96. For correlated bets with intra-portfolio correlation \(\rho = 0.3\), the effective number of independent bets is approximately \(n / (1 + (n-1)\rho)\), which for \(n = 30\) gives an effective \(n\) of about 3.4 and a probability of 0.30.

The calculation is rough, but it captures the problem. A green-tech fund cannot achieve diversification by adding more ventures in the same technology cluster; it must diversify across clusters, which requires a larger fund and a broader mandate. A fund-of-funds structure, which allocates across multiple specialist funds, suits the correlation structure better than a single direct-investment fund. It diversifies across fund managers, each specializing in a technology cluster, and avoids the concentration a single fund would face.

The portfolio construction problem also explains why the Series B/C gap persists. A specialist fund already concentrated in one cluster cannot write a large follow-on check without increasing concentration further. The growth-stage investor who could write the check is a generalist, and the generalist lacks the technical diligence capacity to assess the venture. The gap is a matching problem as well as a capital problem: the investors with capital lack the expertise, and the investors with expertise lack the capital.

11.3.8 The Role of Public Research and the Spillover Channel

The cases understate one factor: public research. Novacell's cathode chemistry originated in a university laboratory; GridSense's forecasting algorithms were developed with a research grant; BioNova's microbial strain was isolated in a public research institute. The public research system supplied the early-stage knowledge the ventures commercialized, and investors priced the technology on the assumption the underlying science was sound.

The spillover channel is the economic justification for public research funding, and it also justifies public co-investment in the ventures that commercialize the research. The social return to the research exceeds the private return because knowledge spills over to competitors and adjacent applications. The venture captures a fraction of the social return, which is why the private market underinvests in commercialization relative to the social optimum.

The empirical literature on the spillover channel is large and need not be reviewed here. The point for the triad argument is that the catalytic channel carries technology transfer, not just money. The public instruments that work best connect the research system to the venture system—through technology-transfer offices, proof-of-concept grants, and university-affiliated venture funds. The EIC Accelerator's grant component is one example; the Israeli Yozma model's reliance on the technology incubator program is another. That connection converts a research result into a venture, and the venture converts the research into a commercial product.

11.3.9 The Measurement of Impact: A Worked Example

The impact test requires a counterfactual. A worked example from the battery-materials case shows the calculation. Novacell's cathode reduces cobalt content by 20 percent at equal energy density. The environmental impact of the reduction depends on the emissions intensity of cobalt mining and refining and on the share of the cathode market the technology captures.

Four steps. The baseline cobalt intensity of the cathode is \(c_0\) kilograms of cobalt per kilowatt-hour of cell capacity. The technology reduces this to \(c_1 = 0.8 c_0\). The emissions intensity of cobalt production is \(e\) kilograms of CO2 per kilogram of cobalt. The market share the technology captures by year \(t\) is \(s_t\). The avoided emissions in year \(t\) are:

where \(M_t\) is total cell capacity produced in year \(t\). With plausible values (\(c_0 = 0.3\) kg/kWh, \(e = 15\) kg CO2/kg Co, \(M_t = 800\) GWh, \(s_t = 5\) percent), avoided emissions are approximately 0.14 million tonnes of CO2 per year. At a social cost of carbon of $50 per tonne, the social value is $7 million per year—small relative to the venture's capital requirement, but growing with market share.

Two points follow. The impact of a single venture is small in absolute terms, so the impact test must be applied at the portfolio level, not the venture level. And the impact depends on parameters (\(e\), the emissions intensity of cobalt production) that are uncertain and outside the venture's control. The measurement problem is a structural feature of green-tech impact assessment, and it argues for a portfolio-level approach that aggregates across ventures and uses conservative parameter values.

11.3.10 What the Failure Analysis Adds

The failure analysis and the required-return decomposition add three things to the success-factor framework.

They identify the binding constraint. The largest single component of the required-return premium is technology risk, and the largest single failure mode is milestone failure without renegotiation. The two are connected: the technology-risk premium is high because investors cannot distinguish an informative failure from an uninformative one, and that inability is what makes the milestone contract fragile. Investment in third-party technical verification is therefore a high-leverage intervention. A testing laboratory or certification body that can assess a milestone independently reduces the technology-risk premium for every venture in the category, which is a public good.

They show that the financing gap is a matching problem as well as a capital problem. The investors with capital lack the expertise to assess the ventures, and the investors with expertise lack the capital to fill the gap. The fund-of-funds structure addresses the matching problem by allocating across specialist funds, each with expertise, while the fund-of-funds manager has the capital. The structure is not a panacea, but it suits the correlation structure better than a single direct-investment fund.

They show that the impact test must be applied at the portfolio level. The impact of a single venture is small and uncertain, and the parameters that determine it are outside the venture's control. A portfolio-level approach that aggregates across ventures and uses conservative parameter values is more informative than a venture-level approach that requires each venture to demonstrate impact.

These findings feed into the policy framework of Chapter 12, which must accommodate the sectoral variation in the integrated mechanism. The framework cannot be a single instrument; it must be a portfolio of instruments, sequenced according to the sector's stage of financial development and the binding constraint the sector faces.

11.4 Synthesis: The Catalytic Channel in the Triad

11.4.1 What the Cases Show About the Triad

The cases support a specific reading of the triad argument. The catalytic channel is present in green-tech venture finance, and it operates through the milestone contract, the staged financing structure, and the public instruments that fill the Series B/C gap. The rating channel is present but thin; it operates through the debt layer of commercial-stage ventures, and its effect on the cost of capital is small—15 basis points in the Novacell case. The insurance channel is largely absent. The risks venture investors bear are not the risks the insurance industry currently underwrites, and the products that would bridge the gap—technology performance warranties, regulatory-delay insurance—are not available at scale.

The integrated mechanism in green-tech venture finance is therefore a two-channel mechanism, as in agriculture and forestry, but the two channels are catalysis and a thin rating layer, with insurance absent. The contrast is instructive. In agriculture and forestry the catalytic channel was absent because the sector did not attract venture finance, and the interaction ran between rating and insurance. In green-tech venture finance the catalytic channel is present, and the interaction runs between catalysis and rating, with insurance absent. The integrated mechanism is a family of designs, each combining the channels relevant to the sector's asset structure.

11.4.2 The Conditions for Catalysis

The cases support a conditional reading of the catalysis claim. Catalysis works where three conditions hold.

A verifiable milestone structure comes first. The milestone contract is the core financing technology, and it works only when milestones are third-party verifiable. The failure analysis shows milestone failure without renegotiation is the largest single failure mode, and it arises when the lead investor lacks the technical model to assess whether the failure is informative. Investment in third-party verification is a high-leverage intervention.

A risk-buffering layer that absorbs early-stage downside comes second. Venture investors in the cases were willing to take technical risk but not regulatory risk, policy risk, or the risk of an uninformative milestone shortfall. The instruments that absorbed those risks were grants, warrant cover, and milestone renegotiation. The insurance industry has not developed products for these risks, and a future policy framework could address the gap.

A policy environment that makes the green attribute valuable comes third. The demand-failure cases show a venture can be financeable and still fail if the policy or price that would make its product valuable has not materialized. The coordination problem between venture finance policy and environmental policy is real, and it is why green-tech venture markets cluster in jurisdictions with strong environmental policy.

Where the three conditions hold, the catalytic channel operates and the integrated mechanism is effective. Where any one is missing, the channel stalls, and the venture falls back on the credit channel, which fits its asset structure badly.

11.4.3 The Limits of the Case Method

A word on method before the chapter closes. The case method is well suited to identifying mechanisms and tracing the sequence through which a financing structure succeeds or fails. It is less suited to estimating the magnitude of an effect, because the cases are selected and the counterfactual is constructed rather than observed. The quantitative claims in this chapter—the required-return decomposition, the portfolio-construction calculation, the impact worked example—are illustrative, and they should be read as orders of magnitude rather than precise estimates.

The remedy is the matched-sample approach used in the EIC evaluation and the guiding-fund evaluation, which compares funded and unfunded but comparable ventures. It requires data that are not always available and a comparison group that is genuinely comparable. Where the data exist, the approach is more informative than the case method; where they do not, the case method is the best available. This chapter has used both and has flagged the claims that rest on the weaker method.

11.4.4 The Transition to Policy Design

The chapter's findings set up the policy framework of Chapter 12, which must accommodate three things.

The integrated mechanism is a family of designs, not a single design. The channels that matter differ by sector, and the policy framework must be modular, combining the channels relevant to each sector's asset structure.

The binding constraint differs by sector and by stage. In agriculture and forestry the binding constraint is the absence of the catalytic channel; in green-tech venture finance it is the Series B/C gap and the absence of the insurance channel. The policy framework must diagnose the binding constraint before it prescribes an instrument.

The public instruments that work target the risks private investors cannot diversify, not the risks they can. The grant component targets technology risk; the guarantee targets regulatory risk; the fund-of-funds structure targets the matching problem. An instrument that lowers the cost of capital without changing the risk is a subsidy, not a market fix, and should be recognized as such.

Chapter 12 builds these findings into a policy framework with three layers: a rating layer that provides the information infrastructure, an insurance layer that buffers the risks private markets cannot absorb, and a catalytic layer that fills the financing gaps private markets cannot fill. The layers are sequenced according to the sector's stage of financial development and coordinated through a single diagnostic process that identifies the binding constraint before it prescribes an instrument. The framework is a decision procedure. Its value is the discipline it imposes on the policy process.

The empirical record of this chapter and the previous one is the foundation for that procedure. The record shows the integrated mechanism works where the channels are combined and the binding constraint is addressed, and it fails where a channel is missing or an instrument is mis-targeted. The policy framework's task is to generalize the conditions for success and to specify the instruments that satisfy them. That is the subject of Chapter 12.