Chapter 10: Sci-Tech Finance and the Catalysis of Green Innovation

The Jiangsu furnace upgrade stalled for eleven months. The manufacturer had the collateral, the cash flow, and a supplier quote for RMB 40 million in new diffusion equipment. What it lacked was a document that could tell a lender the upgrade would cut 11,000 tons of CO2 a year and lift conversion efficiency by 0.4 percentage points. The bank's credit model read a capex line and a depreciation schedule. It did not read carbon. The loan closed only after a provincial sci-tech finance platform rebuilt the application around a technology-readiness score, a third-party environmental verification, and a guarantee from a RMB 200 million risk compensation fund. The rate came in 80 basis points below the bank's standard SME pricing.

That eleven-month delay is the subject of this chapter. Rating identifies environmental performance and converts it into a credit-relevant signal. Insurance buffers the risks that make green cash flows volatile. Neither moves capital into early-stage, technology-heavy segments at the scale the transition requires. The third function is catalysis: the institutional arrangements that lower the cost of connecting a green technology firm to a financier who can price it. Venture capital, government guiding funds, and sci-tech financial intermediaries each perform part of this function. The question is how the parts fit, where each fails, and what a policymaker or a fund manager can do about the failures.

Catalysis, in the chemical sense, lowers activation energy without being consumed. In finance the activation energy is the transaction cost of matching a firm with capital. A four-month technical due diligence on a novel electrolyzer membrane is one example; the consulting fee alone ran to RMB 400,000, and the VC bore it before knowing whether the round would close. A sci-tech finance mechanism is catalytic if it reduces that cost for green technology firms specifically, not for firms in general.

10.1 Venture Capital and Green Tech

10.1.1 The Fit and the Mismatch

Venture capital matches early-stage technology risk better than any other instrument. It accepts equity risk, tolerates long horizons, prices failure across a portfolio, and supplies governance and network support alongside money. Green technology firms share the features that make VC useful: high uncertainty, intangible assets, a long path to profitability, and a founder team that needs more than a loan.

The mismatch sits in the return distribution. A software investment can return 10x within five to seven years because the marginal cost of scaling is near zero. A new battery chemistry or a more efficient electrolyzer must pass through pilot, demonstration, and commercial-scale deployment before it generates revenue at scale. Each stage consumes capital and time. The marginal cost of scaling is the cost of a factory. Time to exit lengthens, capital intensity rises, and the probability of total loss at the demonstration stage is higher than in software because the failure modes are physical.

The arithmetic is unforgiving. Let a fund hold investments \(i = 1, \dots, N\), each with expected exit value \(V_i\), investment cost \(C_i\), time to exit \(T_i\), and probability of total loss \(p_i\). The net multiple on invested capital is

and the internal rate of return solves

The critical parameter is \(T_i\). A software investment might exit in four years; a demonstration-stage green technology investment in eight. Holding \(V_i\) and \(p_i\) fixed, the increase in \(T_i\) alone lowers \(r\) substantially. A fund that must return 20 percent net to its limited partners cannot hold an eight-year position with the same expected multiple as a four-year position and clear the hurdle. The allocation follows the arithmetic, not the ideology.

A second structural issue compounds it. Green technology investments often require follow-on capital at the demonstration and early-commercial stages, and the amount required at those stages is large relative to the initial venture round. A fund that writes a $3 million Series A check may face a $30 million Series C requirement two years later. If the fund cannot lead or even participate at that scale, its ownership is diluted, and the return accrues to later-stage investors with different risk tolerance. The early-stage VC bears the technology risk and captures a smaller share of the deployment return.

10.1.2 What the Aggregate Numbers Hide

Global venture investment in climate technology reached roughly $70 billion in 2022, according to BloombergNEF, up from about $2 billion a decade earlier. The composition matters more than the total. A substantial share went to electric-vehicle manufacturers, battery cell producers, and solar module manufacturers — capital-intensive but relatively mature segments with visible revenue and collateralizable assets. These are not the segments where the catalysis problem is acute. The acute problem is in earlier-stage, deep-technology segments: green hydrogen, long-duration storage, carbon capture, industrial process heat, sustainable aviation fuel. These received a small fraction of the total.

Capital flows to the segments where time-to-exit is short and the path to revenue is visible. It does not flow to the segments where the physics is hard and the deployment timeline is long. If the transition requires capital in the long-duration segments, the fund model must be supplemented.

10.1.3 Corporate Venture and Strategic Investors

One supplement is corporate venture capital. Utilities, oil and gas majors, and equipment manufacturers have investment horizons longer than a ten-year fund and strategic reasons to hold a position through demonstration. A utility that invests in a long-duration storage company is buying a supply relationship, an option on a technology it may need to deploy, and a seat at the table where the technology's specifications are set. These non-financial returns can justify a lower financial return, which extends the set of investable projects.

The record is mixed. Strategic investors can provide pilot sites, offtake agreements, and engineering support that a financial VC cannot. They can also distort the investee's incentives, pushing the technology toward the parent's incumbent assets and away from more disruptive applications. Corporate venture budgets are procyclical: they expand when the parent's cash flow is strong and contract when it is not, which is the opposite of the countercyclical support early-stage firms need during a downturn. Corporate venture fills part of the gap for firms whose technology aligns with an incumbent's asset base. It does not fill the gap for firms whose technology threatens that base, which is often where the largest environmental gains lie.

10.1.4 The Exit Problem

VC returns are realized at exit, and each exit route has frictions for green technology firms.

Public listing requires a revenue profile and a disclosure package that early-stage hardware firms rarely have. The special-purpose acquisition company route, which briefly offered an alternative, proved to have its own problems: disclosure quality was uneven, sponsor economics created adverse incentives, and the post-merger performance of many green SPACs was poor. The route has narrowed.

Trade sale requires a buyer. The natural buyers are incumbent industrial firms, which brings back the strategic-investor dynamic: the buyer values the target partly for its fit with existing assets, which caps the price for technologies that do not fit.

Secondary sales have grown as a share of VC exits, and this is a genuine development. A secondary market allows early-stage investors to realize returns before the firm reaches commercial scale, which shortens the effective holding period and improves the fund arithmetic. The growth of specialist secondary funds that buy positions in green technology portfolios is one of the more consequential institutional innovations of the past decade. It does not solve the duration problem, but it redistributes it to investors with longer horizons.

The exit problem also has a geographic dimension that rarely gets discussed. A Chinese green technology firm that lists domestically faces a retail-dominated market with limited analyst coverage of deep technology; a firm that lists in Hong Kong or on a US exchange faces valuation discounts and geopolitical risk premia that have widened since 2021. The choice of exit venue changes the expected multiple by a factor that can exceed the difference between a good and a mediocre technology. Founders and VCs price this in, and the pricing feeds back into which technologies get funded at the seed stage.

10.1.5 The Catalytic Function of VC, Properly Specified

VC catalyzes in three ways a bank cannot. It supplies equity risk capital that requires neither collateral nor current cash flow. It supplies governance and network access that improves the firm's probability of survival. And it supplies a screening function: the decision to invest is a costly signal that other investors can use.

The third function matters most for the integrated mechanism this book describes. A VC investment is a rating in a loose sense: an assessment, backed by money, that the firm's technology and team are worth a continued bet. The problem is that this rating is private, idiosyncratic, and not comparable across firms. Two VCs may reach opposite conclusions about the same firm, and neither conclusion is expressed in a form a lender can use.

A standardized environmental rating would let a VC's portfolio companies be compared on environmental performance, and it would let later-stage investors, lenders, and government funds use the VC's screening as an input without replicating the entire due-diligence process. The rating does not replace the VC's judgment; it makes the judgment legible. A VC that can point to a standardized environmental rating for its portfolio companies can raise follow-on capital more cheaply, because the follow-on investors face lower information asymmetry. The rating lowers the activation energy for the next round.

10.1.6 Two Firms, One Round

Consider two firms, both seeking a Series B of $15 million.

Firm A is a software platform that optimizes building energy use. It has 200 commercial customers, $4 million in annual recurring revenue, gross margins of 80 percent, and a clear path to profitability in three years. Its environmental impact is measurable: the platform reduces customer energy consumption by an average of 12 percent, which the firm can document. The VC's due diligence takes six weeks. The round closes at a $60 million pre-money valuation.

Firm B is a developer of a novel membrane for green hydrogen electrolysis. It has a pilot line, two industrial partners testing the membrane, and no revenue. Its environmental impact, if the technology scales, is large: a 15 percent improvement in electrolyzer efficiency translates into a meaningful reduction in the levelized cost of hydrogen. But the impact is contingent on a demonstration plant that will cost $40 million and take three years to build. The VC's due diligence takes four months and requires a technical consultant. The round closes at a $35 million pre-money valuation, if it closes at all.

The valuation gap is not a judgment about the relative importance of environmental impact. It is a judgment about the time and certainty of the return. Firm B's technology may matter more for the transition. Firm A's technology is easier to finance. The gap is the catalysis problem, and exhortation does not close it. Instruments do: grants that fund the demonstration, offtake agreements that de-risk revenue, insurance that covers performance risk, and government funds that accept a lower return in exchange for the environmental outcome.

10.2 Government Guiding Funds

10.2.1 The Logic and the Governance Trap

The case for public intervention in early-stage green capital rests on a specific market failure. The social return to a green technology investment exceeds the private return because the environmental benefit is a public good: the firm that develops a lower-carbon process cannot capture the full value of the carbon avoided, because the carbon price, where one exists, is below the social cost of carbon, and where one does not exist, the benefit is unpriced. The gap between social and private return justifies public co-investment.

The argument justifies public support for green technology in general. It does not specify which instrument — grant, equity, loan guarantee, tax credit — fits which stage, and it does not address the governance problems that have plagued public investment vehicles. Public capital is exposed to political capture, to the soft budget constraint, and to the temptation to substitute for private capital rather than to catalyze it. A government fund that invests in firms private investors would have funded anyway adds little; a fund that invests only in firms no private investor will touch may be funding the unviable. The design challenge is to locate the narrow band where public capital changes the private investor's decision without displacing it.

10.2.2 The Chinese Guiding Fund Model

China's government guiding funds (政府引导基金) offer the most extensive laboratory for this design problem. The model originated in the early 2000s and expanded rapidly after 2015. By the end of 2022, according to the Zero2IPO Research Center, the number of government guiding funds at the provincial level and below exceeded 1,500, with total committed capital in the trillions of RMB. The funds are typically structured as funds of funds: the government contributes a share of a venture fund's capital, the rest is raised from private limited partners, and the fund is managed by a professional GP selected through a competitive process.

The structure addresses the catalysis problem directly. The government's contribution lowers the private investor's cost of capital and absorbs part of the downside, which changes the arithmetic for investments in the long-duration segments. The GP selection process is meant to preserve professional judgment. The fund-of-funds structure is meant to diversify risk and to leverage the GP's deal-sourcing network.

The record is mixed, and the reasons are instructive.

Capital has concentrated in the later stages. A guiding fund evaluated on financial return, or on the number of firms it supports, has an incentive to invest in firms close to listing. The early-stage mandate is often stated but weakly enforced. A 2021 study by the Chinese Academy of Social Sciences found that a majority of guiding fund capital flowed to expansion-stage and pre-IPO rounds, with a minority to seed and early-stage rounds. The funds were catalyzing, but not at the stage where the catalysis was most needed.

Geographic mandates have distorted allocation. Many guiding funds require the GP to invest a specified share of capital within the fund's home jurisdiction. The requirement is politically understandable — local governments want the jobs and tax base — but it fragments the capital market and forces GPs to invest in firms that happen to be located in the right place rather than the firms with the best technology. A solar firm in a province with a strong guiding fund may be funded; a better solar firm in a province without one may not.

Return expectations have been unrealistic. Guiding funds often set a financial return target close to the private market's, which defeats the purpose. If the fund must earn a market return, it cannot invest in the segments where the social return exceeds the private return, because those are precisely the segments where the private return is below market. The funds that have been most effective at reaching early-stage green technology are those that explicitly accept a lower financial return in exchange for a measured environmental outcome.

10.2.3 Design Principles

Five principles follow from the Chinese experience and from smaller-scale experiments in Europe and the United States.

Stage specificity. The fund's mandate should specify the stage at which it invests, and the evaluation should use the stage-appropriate metric. A seed-stage fund should be evaluated on the number and quality of firms it seeds, not on its IRR. A growth-stage fund can be evaluated on financial return with an environmental overlay.

Additionality testing. The fund should be required to demonstrate additionality: that its investment changed the private investor's decision. The test can be operationalized by requiring the GP to document the counterfactual — what would have happened without the public capital. A simple version: the fund invests only when the private co-investors' committed share is below a threshold that indicates genuine reluctance.

Return tolerance with environmental conditionality. The fund should accept a below-market financial return, but the concession should be conditional on a verified environmental outcome. The conditionality converts the subsidy into a payment for a public good, which is defensible, rather than a transfer to a private investor, which is not.

Geographic flexibility. The local-content requirement should be relaxed or replaced with a requirement that the funded firm locate a specified activity — a pilot plant, a manufacturing line — in the jurisdiction. This captures the local benefit without distorting the selection of firms.

Exit discipline. The fund should have a clear exit strategy and a sunset date. Public capital that remains in a firm indefinitely becomes a form of state ownership, with the governance problems that entails.

10.2.4 The Leverage Ratio and the Catalysis Multiplier

The central metric for a guiding fund is the leverage ratio: private capital mobilized per unit of public capital. The leverage ratio alone is incomplete, because it does not distinguish between public capital that displaced private capital and public capital that mobilized additional private capital.

Let \(I_{pub}\) be the public investment and \(I_{priv}\) the private investment in a firm. The observed total is \(I = I_{pub} + I_{priv}\). The counterfactual private investment, absent the public capital, is \(I_{priv}^0\). The catalysis effect is

If \(\Delta > 0\), the public capital crowded in private capital. If \(\Delta = 0\), it displaced private capital one-for-one. If \(\Delta < 0\), it displaced more than one-for-one, which can happen if the public capital signals a political commitment that discourages other investors or distorts the firm's incentives. The catalysis multiplier is

A multiplier above one means each unit of public capital mobilized more than one unit of additional private capital. A multiplier between zero and one means partial crowding in. A negative multiplier means crowding out.

Estimating \(m\) requires an estimate of \(I_{priv}^0\), which is unobservable. The practical approaches are comparison with matched firms that did not receive public capital, and structured interviews with the private co-investors about their decision process. Neither is perfect. The matched comparison is contaminated by selection: firms that receive public capital may differ systematically from those that do not. The interviews are subject to ex post rationalization. An imperfect estimate of \(m\) still beats the leverage ratio, which is not an estimate of catalysis at all.

The measured multipliers vary widely by fund design. The Suzhou Industrial Park guiding fund, which imposes a stage mandate and a professional GP selection process, reported a leverage ratio of roughly 1:4.2 in its 2021 annual report — RMB 4.2 of private capital per RMB 1 of public capital — but the fund does not publish a counterfactual estimate, so the crowding-in component cannot be separated from the displacement component. At the other end, several prefecture-level funds with strict local-content requirements have reported leverage ratios below 1:2, and interviews with co-investors in two such funds (conducted for this study in 2022) suggest that a substantial share of the private capital would have been invested in the same firms regardless of the public contribution. The honest summary is that the design features predict the multiplier, and the measured range across well-designed and poorly designed funds is wide enough that the average is not a useful policy target.

10.2.5 The Interaction with the Rating Infrastructure

A guiding fund that invests in green technology firms faces the same identification problem as a lender. It must assess the firm's environmental performance to determine whether the investment qualifies for the fund's mandate. The environmental performance rating developed in Chapters 3 and 4 can serve this function, and its use has a second-order benefit: it makes the fund's portfolio comparable across funds and over time, which is a precondition for any serious evaluation of environmental additionality.

The practical arrangement is straightforward. The fund requires each portfolio company to obtain an environmental performance rating from an approved provider. The rating is updated annually. The fund's own reporting aggregates the ratings of its portfolio, weighted by investment, to produce a portfolio-level environmental score. The score is reported alongside the financial return, and both are used in the fund's evaluation.

This arrangement does not solve the deeper problem — the rating measures the firm's environmental performance, not the fund's contribution to it — but it makes the fund's environmental claim auditable. An auditor can check whether the portfolio's weighted rating improved over the fund's holding period, and whether the improvement is attributable to the fund's governance and capital or to exogenous factors. The attribution question is hard, but it is at least posed.

10.2.6 The Missing Demonstration Capital

The most persistent gap in the green technology capital stack is at the demonstration stage. A firm that has proven its technology at laboratory scale and raised a seed round needs capital to build a demonstration plant. The amount is too large for VC (tens of millions), too risky for bank debt (no revenue, no collateral), and too early for infrastructure funds (no operating asset). Government grants can fill part of the gap, but grant programs are typically small and slow. Government guiding funds can fill part of the gap, but their mandates often skip the demonstration stage.

The gap is structural, and it is the single largest obstacle to the deployment of deep-technology green innovations. A firm that cannot build a demonstration plant cannot prove commercial viability, and a firm that cannot prove commercial viability cannot raise the capital to build the plant. The circularity is the trap.

Three instruments have broken the trap in practice, and all three share a common feature: they transfer part of the demonstration risk to a party that can bear it. The first-of-a-kind (FOAK) financing model, developed by the European Investment Bank and replicated by several national development banks, combines a grant, a subordinated loan, and a commercial loan in a single package, with the grant absorbing the technology risk and the commercial loan priced against the demonstrated cash flow. The model has been used for offshore wind, advanced biofuels, and carbon capture. Its limitation is scale: transaction costs are high relative to the ticket size, and the model has not been standardized.

A second instrument is the advance market commitment (AMC), in which a public buyer commits to purchase a specified volume of the technology at a specified price, conditional on performance. The commitment de-risks the revenue side, which makes the demonstration plant bankable. The AMC has been used successfully for vaccines and is being piloted for green hydrogen and sustainable aviation fuel. The design challenge is the price: set it too high and the public overpays; set it too low and the commitment does not change the investment decision.

A third instrument is the blended finance structure, in which concessional capital from a development finance institution is combined with commercial capital in a layered structure. The concessional tranche absorbs first losses, which improves the risk-return profile of the commercial tranche. The structure is flexible and has been used for renewable energy in emerging markets. Its limitation is that the concessional capital is scarce and the structuring costs are high.

None of these instruments is a complete solution. Each addresses a part of the demonstration gap. The catalysis function, in this segment, is performed by a portfolio of instruments, and the portfolio is still being assembled.

10.3 Intermediary Systems and Platforms

10.3.1 Why Intermediaries Matter

The third catalysis mechanism is the intermediary layer: the institutions and platforms that sit between the green technology firm and the capital provider and reduce the cost of the transaction. The category is broad. It includes technology evaluation firms that assess readiness and performance; intellectual property (IP) valuers that price a patent portfolio; testing and certification bodies that verify performance claims; data platforms that aggregate firm-level environmental data; and specialist financial intermediaries — green investment banks, sci-tech finance divisions of commercial banks, and technology finance platforms — that structure and place the transactions.

The economic function of the intermediary is to produce and certify information. The green technology firm knows more about its technology than any outside investor. The investor knows more about the capital markets than the firm. The intermediary reduces the information asymmetry in both directions, and in doing so lowers the cost of the transaction. This is the same function credit rating agencies perform for bond issuers and auditors perform for listed companies. The sci-tech finance intermediary performs it for the segment of the market where the information asymmetry is most severe: early-stage technology firms whose assets are intangible and whose performance is uncertain.

Markets with a denser intermediary layer — Silicon Valley, Cambridge in the UK, Shenzhen — exhibit higher rates of technology firm formation and faster capital deployment. The causality is contested: intermediaries may follow firm formation rather than cause it. The policy experiments that have built intermediary capacity tell a more specific story. The UK's Catapult centres, established from 2011, reported that their client firms raised £1.2 billion in follow-on capital between 2015 and 2020, against £300 million in public funding over the same period. The attribution problem remains — the firms might have raised the capital anyway — but the Catapult evaluation used matched comparison firms and found a follow-on capital rate roughly 40 percent higher for Catapult clients. That is one of the few direct estimates available.

10.3.2 The Sci-Tech Finance Platform: A Taxonomy

The platforms that have emerged in China over the past decade illustrate the range of intermediary models. They can be grouped into four types.

The government-backed technology finance platform. These are typically operated by a local science and technology bureau or a state-owned investment company. They maintain a database of technology firms, provide matching services between firms and investors, and often administer government subsidy and guarantee programs. The Suzhou Science and Technology Finance Service Platform is a representative example: it aggregates data on several thousand firms, provides a standardized evaluation, and connects firms to a network of banks, VCs, and guarantee companies.

The bank-affiliated sci-tech finance division. Several Chinese banks, including the Bank of Beijing and the Shanghai Pudong Development Bank, have established dedicated sci-tech finance divisions that lend to technology firms using evaluation methods that differ from the standard corporate credit model. The methods incorporate IP value, R&D intensity, and government certification, and the loans are often structured with a guarantee from a government risk compensation fund.

The IP exchange and valuation platform. These platforms provide IP valuation, trading, and pledge financing services. The valuation methods combine cost, income, and market approaches, and the platform certifies the valuation for use in loan collateral. The effectiveness of IP pledge financing has been limited by the thinness of the secondary market for IP, which makes the collateral hard to liquidate. The platforms address this partly by providing a trading venue, but the liquidity remains thin.

The industry-specific technology platform. These platforms focus on a single technology domain — new energy, biomedicine, advanced materials — and provide specialized evaluation, testing, and pilot-scale services. They are closer to the technology than the general platforms, which improves the quality of the evaluation, but they are also more expensive to operate and harder to scale.

10.3.3 The Evaluation Problem

The central function of any sci-tech finance intermediary is evaluation: the assessment of a technology firm's prospects and risks in a form a financier can use. The evaluation problem is harder than the credit rating problem for established firms because the standard credit metrics — cash flow coverage, leverage, collateral — are weak or absent. The firm's value is in its technology and its team, and both are hard to measure.

The evaluation methods that have been developed fall into three families.

The scorecard approach. A set of weighted indicators — R&D expenditure as a share of revenue, patent count and quality, the founder's track record, the technology's readiness level, the size of the addressable market — is scored and combined into a single rating. The scorecard is transparent and cheap to administer, but the weights are arbitrary and the predictive power is modest. The scorecard works best as a screening tool, not as a pricing tool.

The technology readiness level (TRL) approach. The technology is assessed on a nine-level scale, from basic principles observed (TRL 1) to actual system proven in an operational environment (TRL 9). The TRL is a useful summary of where the technology sits in its development, and it maps reasonably well onto the financing stage: TRL 1–3 corresponds to grant and seed funding, TRL 4–6 to VC, TRL 7–8 to demonstration and project finance, TRL 9 to commercial debt. The limitation is that TRL measures maturity, not commercial viability; a technology can be at TRL 9 and still have no market.

The due-diligence approach. A specialist evaluator conducts a deep technical and commercial review, often with the help of external experts, and produces a narrative assessment with a recommendation. This is the most informative method and the most expensive. It is the method used by VC firms and by the more sophisticated government funds. Its cost is justified only for transactions above a certain size.

The three methods are complementary, and the more developed platforms combine them: the scorecard for initial screening, the TRL for stage classification, and the due diligence for the final decision. The combination is expensive, which is why the intermediary layer has concentrated in regions with a high density of technology firms and a large enough deal flow to amortize the fixed cost.

10.3.4 The Environmental Dimension of the Evaluation

The evaluation methods described above were developed for technology finance in general. The green technology firm requires an additional dimension: the environmental performance of the technology. This is the point where the rating infrastructure of Chapters 3 and 4 connects to the intermediary layer.

An environmental performance rating for a technology firm must capture three things that a conventional technology evaluation does not. First, the environmental impact of the technology if it scales: the tons of carbon avoided per unit of deployment, the reduction in water use, the avoided waste. Second, the additionality of the impact: whether the technology's environmental benefit is additional to what would have happened without it, or whether it is a marginal improvement on an incumbent process that would have improved anyway. Third, the measurement uncertainty: the confidence interval around the impact estimate, which for early-stage technologies can be wide.

The three dimensions can be combined into a single rating, but the combination should not obscure the components. A technology with a large but uncertain impact is different from a technology with a small but certain impact, and the two should be distinguishable in the rating. The practical solution is a two-dimensional rating: an impact score and a confidence score, reported separately. The financier can then apply a weighting that reflects its own risk tolerance.

The rating is produced by the intermediary, using the data the firm provides and the verification the intermediary can perform. The verification is the hard part. A firm's claim about its technology's environmental impact is a projection, and projections are cheap to make and hard to falsify. The intermediary can verify the measurement methodology, the boundary assumptions, and the consistency of the claim with the published literature. It cannot verify the future. The rating should therefore be explicit about what is verified and what is projected, and the financier should price the projection risk accordingly.

10.3.5 A Platform in Operation

To make the intermediary function concrete, consider how a sci-tech finance platform processes a loan application from a green technology firm.

The firm is a developer of an industrial heat pump that can supply process heat at 180°C, replacing a gas-fired boiler. It has 12 employees, two installed units at customer sites, and revenue of RMB 8 million. It needs RMB 25 million to expand its manufacturing line and to fund the working capital for a pipeline of orders.

The platform's process runs in five stages.

Stage 1: Registration and data capture. The firm registers on the platform and submits its financial statements, patent portfolio, customer contracts, and technology documentation. The platform's data system extracts the structured data and flags the missing items.

Stage 2: Technology evaluation. The platform's technical team assesses the heat pump's TRL (7, because the units are installed and operating but the manufacturing process is not yet at scale), reviews the performance data from the installed units, and benchmarks the performance against published alternatives. The team produces a technology score.

Stage 3: Environmental rating. The platform's environmental team estimates the carbon avoided per unit: the heat pump's coefficient of performance (COP) of 3.2 versus the gas boiler's efficiency of 0.9, applied to the local grid's emission factor and the customer's load profile. The estimate is 320 tons of CO2 avoided per unit per year, with a confidence interval of ±15 percent. The team rates the impact as high and the confidence as medium, and notes that the additionality depends on whether the customer would have replaced the boiler anyway.

Stage 4: Financial evaluation. The platform's financial team reviews the firm's cash flow, the order pipeline, and the warranty obligations. It notes that the firm's receivables are concentrated in three customers and that the warranty provision is thin relative to the engineering risk. It produces a financial score and a recommended loan structure: a RMB 20 million term loan with a five-year tenor, a two-year grace period, and a guarantee from the local government's risk compensation fund covering 50 percent of the principal.

Stage 5: Matching and placement. The platform packages the evaluation and the loan recommendation and presents it to the banks in its network. Two banks express interest. The loan is placed with one of them at a rate 80 basis points below the bank's standard rate for an unrated SME, reflecting the platform's evaluation and the government guarantee.

The platform did not supply the capital; the bank did. The platform did not supply the guarantee; the government did. What the platform supplied was the information that made the bank willing to lend and the pricing that reflected the firm's actual risk.

The example also illustrates the limits. The platform's evaluation took six weeks and cost roughly RMB 150,000, which is viable only because the platform is subsidized and because the loan size is large enough to amortize the cost. A smaller loan — RMB 5 million, say — would not cover the evaluation cost, and the platform would not process it. The intermediary layer, like the rest of the financial system, has a minimum efficient scale, and firms below that scale remain outside it.

10.3.6 The Standardization Problem

The intermediary layer is fragmented. Each platform has its own evaluation methodology, its own rating scale, and its own data format. A firm evaluated by one platform cannot easily use that evaluation with a financier that works with a different platform. The fragmentation raises the cost of the evaluation, because it must be repeated for each financier, and it reduces the value of the evaluation, because it is not comparable across firms.

Standardization is the obvious solution, and it is harder than it looks. The evaluation methods are proprietary to the platforms, and the platforms have commercial reasons to resist a common standard. The financiers have their own internal models and are reluctant to rely on an external standard they did not design. Regulators have been slow to mandate a standard because the methods are still evolving.

The experience with credit rating agencies is instructive. The rating agencies converged on a common scale — the AAA-to-D scale — not because a regulator mandated it, but because the market needed comparability and the agencies that adopted the scale attracted more business. The convergence took decades, and it was accompanied by regulatory recognition (the Nationally Recognized Statistical Rating Organization designation in the United States, the external credit assessment institution recognition in the EU) that gave the scale legal force.

The environmental performance rating is at an earlier stage. The scales in use are not comparable, and there is no regulatory recognition that would give one scale precedence. Standardization will likely follow the same path: market convergence on a scale that the largest financiers accept, followed by regulatory recognition. The process will take years, and in the interim the fragmentation will persist.

Public policy can accelerate the convergence. A government that uses a particular rating scale in its own guiding funds and subsidy programs creates demand for that scale, which encourages the platforms to adopt it. The Chinese government's green bond endorsement mechanism, which requires third-party verification against a common standard, is an example of this approach. The mechanism has not been extended to technology firm ratings, but the logic is the same.

10.3.7 The Data Infrastructure

The intermediary layer depends on data. The evaluation methods described above require firm-level data on technology performance, environmental impact, and financial condition, and the data must be timely, comparable, and verifiable. The current data infrastructure is weak on all three dimensions.

The timeliness problem is the most acute for early-stage firms. The firms are small, their reporting systems are rudimentary, and the data they provide is often annual and stale. A platform that evaluates a firm on the basis of last year's financial statements is evaluating a firm that may have changed substantially.

The comparability problem is the lack of common definitions. Two firms may report their carbon avoided using different boundary assumptions, different emission factors, and different allocation methods. The numbers are not comparable, and the platform must normalize them, which adds cost and introduces error.

The verifiability problem is the lack of audit. The firm's environmental claims are self-reported, and the platform's verification is limited to a desk review of the methodology and the supporting documents. A third-party audit of the environmental claim is rare, because the audit is expensive and the firm is small.

The digital infrastructure that could address these problems is developing. The industrial internet platforms that connect manufacturing equipment to data systems can, in principle, provide real-time performance data from installed units, which is exactly the data the environmental rating needs. The carbon accounting standards being harmonized under the GHG Protocol and the ISO 14064 series provide the definitions that the comparability problem requires. The satellite-based emissions monitoring that has been developed for methane and CO2 provides an independent check on self-reported data for some categories of impact.

The integration of these data sources into the intermediary layer is the next stage of development. A platform that can pull real-time performance data from a firm's installed units, apply a standardized emission factor, and verify the result against satellite data has a much stronger evaluation than a platform that relies on self-reported annual figures. The technology exists; the integration is a matter of investment and standards.

10.3.8 The Intermediary as a Rating Provider

The sci-tech finance intermediary is the natural producer of the environmental performance rating for technology firms. The intermediary has the technical capacity to assess the technology, the data access to measure the environmental impact, and the relationship with the financier to make the rating useful. The rating agencies that dominate the corporate bond market do not have these capacities for early-stage technology firms; their methods are designed for issuers with operating histories and audited financials. The gap is filled by the intermediary, and the policy priority is to strengthen the intermediary's capacity and standardize its output.

This is the point where the three mechanisms of the chapter converge. Venture capital supplies the equity risk capital and the screening signal. Government guiding funds supply the patient capital and the return tolerance. The intermediary supplies the information that makes both legible to the rest of the financial system. The three mechanisms are not alternatives; they are components of a single system, and the system functions only when all three are present.

10.4 The Catalysis Function in the Integrated Mechanism

10.4.1 Sequencing and Complementarity

The rating comes first, because it is the input to everything else. A firm without an environmental performance rating cannot be identified as a green technology firm, and a firm that cannot be identified cannot be catalyzed. The rating is produced by the intermediary, using the technology evaluation and the environmental data. The cost of the rating is a fixed cost, and it is the first cost the catalysisfunction must bear.

The insurance comes second, because it addresses the risks the rating reveals. A firm whose environmental performance is high but whose technology risk is high needs insurance or a guarantee to make the rating bankable. The insurance is priced against the rating: a higher-rated firm pays a lower premium, because its performance is more certain. The insurance and the rating are jointly produced, and joint production is more efficient than separate production.

Catalysis comes third, because it applies capital to the rated and insured firm. The catalysis is performed by the VC, the guiding fund, or the intermediary platform, and its form depends on the firm's stage. A seed-stage firm is catalyzed by a grant and a VC investment. A demonstration-stage firm is catalyzed by a blended finance package. A commercial-stage firm is catalyzed by a bank loan priced against the rating and the insurance.

The sequence is not rigid. A firm may be catalyzed before it is rated, if the catalyst is a VC that performs its own evaluation. But the sequence is the efficient order, because each step reduces the cost of the next. The rating reduces the cost of the insurance, the insurance reduces the cost of the capital, and the capital reduces the cost of the next rating.

10.4.2 The Cost of Catalysis

The cost of the catalysis function is the sum of the costs of the three components: the rating cost, the insurance premium, and the transaction cost of the capital. For a small firm, these costs can be large relative to the amount raised, and the scale problem is the binding constraint.

Consider the arithmetic for a firm raising RMB 20 million. The rating costs RMB 150,000, or 0.75 percent of the amount raised. The insurance premium, for a guarantee covering 50 percent of the principal at a 2 percent annual rate over five years, is roughly RMB 1 million in present value, or 5 percent of the amount raised. The transaction cost of the loan, including the platform fee and the legal costs, is roughly RMB 200,000, or 1 percent. The total cost of catalysis is roughly 6.75 percent of the amount raised.

For a firm raising RMB 5 million, the same costs are proportionally larger. The rating cost is 3 percent, the insurance premium is 5 percent, and the transaction cost is 4 percent, for a total of 12 percent. At this scale, the catalysis cost is a significant fraction of the capital, and the firm may be better off with a simpler, more expensive loan that does not use the full mechanism.

The scale problem is the reason the intermediary layer has concentrated in regions with a high density of firms. The fixed costs of the evaluation and the platform are amortized over a large deal flow, which lowers the per-transaction cost. A region with a thin deal flow cannot support the intermediary, and its firms are catalyzed less efficiently. This is a market failure that public policy can address, by subsidizing the fixed costs of the intermediary in underserved regions. The subsidy is justified by the same logic as the guiding fund: the social return to the catalysis exceeds the private return, and the gap is the case for public support.

10.4.3 The Measurement of Catalysis

The final question is how to measure whether the catalysis function is working. The metric should capture the three components and their interaction.

The first component is the rating coverage: the share of green technology firms in a region that have an environmental performance rating. A high coverage means the identification function is working.

The second component is the insurance penetration: the share of rated firms that have insurance or a guarantee attached to their financing. A high penetration means the buffering function is working.

The third component is the capital mobilized: the amount of private capital invested in rated and insured firms per unit of public capital. A high ratio means the catalysis function is working.

The three components can be combined into a single index, but the combination should be interpreted with care. A region with high coverage but low penetration has an identification system that is not connected to the buffering system. A region with high penetration but low capital mobilized has a buffering system that is not connected to the capital market. The index should be reported as a triple, not a single number, so that the bottleneck is visible.

The data for the three components is available, in principle, from the platforms and the funds. The practical obstacle is the fragmentation of the data systems, which is the same obstacle that limits the intermediary layer. A region that builds an integrated data system — a single registry of rated firms, insured firms, and catalyzed capital — can measure the catalysis function and manage it. A region that does not remains in the position of the loan officer in the opening scene: she has a protocol, but she does not have the data to apply it at scale.

10.5 Conclusion: The Third Component

The chapter opened with a solar-cell manufacturer that could not express its environmental performance in the language its lender's model reads. The analysis has identified three mechanisms that, together, can translate. Venture capital supplies the equity risk capital and the screening signal for the early-stage firm. Government guiding funds supply the patient capital and the return tolerance for the segments where the private return is below the social return. The intermediary layer supplies the evaluation and the data that make the firm legible to the capital market.

Each mechanism is necessary, and none is sufficient. VC alone under-allocates to the long-duration segments because the fund arithmetic does not work. Guiding funds alone concentrate in the later stages and distort allocation through geographic mandates. Intermediaries alone cannot supply capital, and their evaluations are fragmented and expensive. The catalysis function is the joint product of the three.

The connection to the earlier chapters is direct. The rating that the intermediary produces is the same environmental performance rating that the credit model uses. The insurance that the firm attaches is the same risk-buffering layer that the loan file records. The catalysis is the application of capital to the rated and insured firm, and its success is measured by the reduction in the firm's default probability and the increase in its environmental performance. The three functions are sequential in the loan file and complementary in the system.

The solar-cell manufacturer's furnace upgrade will be financed when the rating, the insurance, and the catalysis are in place. The rating will express the 11,000 tons of carbon avoided in a form the credit model reads. The insurance will cover the performance risk of the new equipment. The catalysis will lower the cost of the loan that funds the upgrade. The loan officer will record the estimated PD reduction, and the second loan will be priced against the first.

The next chapter turns to the government-guided funds in more detail, and to the market failure in early-stage green capital that the funds are designed to correct. The analysis there will build on the design principles developed in this chapter and will examine the evidence on the funds' environmental additionality. The question is whether public capital can be structured to catalyze private capital without displacing it, and the answer will determine whether the catalysis function can be scaled beyond the regions and segments where it currently operates.