‍

Communities remain invisible within AI infrastructure decision-making, just as SMEs remain invisible within climate investment pipelines.

Who receives and benefits from the capital being mobilized for AI infrastructure and climate action? Who shapes the terms on which that capital is deployed? Between locally beneficial outcomes and the institutions allocating billions of dollars lies a missing intermediary layer: the capacity to translate technical and financial information, document needs and performance, support negotiation, coordinate implementation, and monitor commitments over time.

‍

AI AND CLIMATE FINANCE: TWO HISTORIC INVESTMENT WAVES

Two historic investment waves are reshaping the global economy. Between now and 2030, data-center infrastructure is projected to require nearly $7 trillion in capital expenditure worldwide, with approximately 70% of that investment supporting AI workloads. At the same time, annual global climate finance has crossed the $2 trillion threshold, reaching an estimated $2.1 trillion in 2025. Yet these extraordinary figures conceal a fundamental allocation problem: in both cases, capital is moving at historic scale while routinely overlooking the actors whose participation will determine whether this investment produces durable results.

While moderating a COP30 panel on centering SMEs in climate action, Global Climate Finance Forum founder Marilyn Waite captured this issue in one deceptively simple question: “Climate finance for what?” What, exactly, are we attempting to finance—and are current investment systems directing capital toward activities capable of producing the outcomes they promise?

But this question has another dimension: finance for whom? Who receives and benefits from the capital being mobilized for AI infrastructure and climate action? Who shapes the terms on which that capital is deployed? Whose knowledge counts when projects are selected, designed, and implemented?

The prevailing narratives of these investment booms center on the hyperscalers building data centers, the developers and utilities enabling their expansion, and the multinational companies, banks, and investment funds announcing major climate commitments. They pay far less attention to the people and enterprises positioned closest to the consequences—and often closest to the solutions.

FINANCE FOR WHOM?

In the AI infrastructure buildout, exclusion begins with information. Nondisclosure agreements and opaque development processes conceal basic project details until plans are too advanced for meaningful community input. Furthermore, residents and local officials often lack the technical and legal capacity needed to interpret what eventually becomes public. They are therefore asked to respond to decisions they had little opportunity to shape and to negotiate only after the most consequential terms have already been set.

This is first an injustice: the people who will live with a facility’s effects on electricity costs, water systems, land, noise, housing, and public infrastructure should have a meaningful role in determining how it is built. But excluding them is also a strategic failure. Residents, local governments, utilities, and community institutions understand existing pressures that may be invisible from a project model. Their knowledge can help developers and investors anticipate risks, negotiate appropriate protections, and design projects capable of maintaining community trust and political viability over decades—not merely through the next permitting or financing cycle.

Climate finance has developed a parallel blind spot. Small and medium-sized enterprises are delivering transformative climate solutions across the Global South—restoring forests, producing clean energy, advancing circular agriculture, and building resilient local economies. Yet despite delivering nearly 80% of locally-developed climate solutions in emerging markets, these high-impact “SMEs” receive less than 10% of climate finance flows.

Investors and financial institutions frequently pass over SMEs not because their innovations are ineffective, but because these businesses do not resemble the assets conventional finance is structured to fund. Smaller transaction sizes, limited collateral and track records, local-currency exposure, and less standardized reporting can place SMEs outside investor mandates or make them appear more costly and risky to finance than larger, more familiar projects. Yet their proximity gives them knowledge of local markets, ecosystems, supply chains, and customer needs that distant capital providers cannot easily reproduce—advantages that current investment practices fail to capture.

Communities remain invisible within AI infrastructure decision-making, just as SMEs remain invisible within climate investment pipelines. In both cases, capital selects for institutional familiarity and financial legibility rather than local knowledge and demonstrated effectiveness. The result is not only inequitable. It is an inefficient allocation of finance that deprives investors of vital knowledge, viable opportunities, and the local partnerships required to turn financial commitments into lasting outcomes.

THE MISSING INTERMEDIARY LAYER: CASES IN THE CARIBBEAN AND AFRICA

This is the structural condition connecting the AI infrastructure buildout to climate finance. Between locally beneficial outcomes and the institutions allocating billions of dollars lies a missing intermediary layer: the capacity to translate technical and financial information, document needs and performance, support negotiation, coordinate implementation, and monitor commitments over time. Without that connective tissue, communities struggle to influence infrastructure decisions, effective enterprises struggle to demonstrate their value in terms investors recognize, and capital fails to incorporate or reach the actors best positioned to improve its outcomes.

We frame this failure through three connected gaps in financial architecture. The intermediary gap is the absence of institutions and capacity capable of connecting local actors to capital and decision-making. The resulting infrastructure gap appears in what investment overlooks or miscalculates—from locally appropriate solutions to long-term grid, environmental, and community costs. The trust gap follows when communities bear those costs without meaningful visibility, influence, or recourse. These are not separate failures. They are cascading expressions of the same architectural weakness.

Puerto Rico, sitting within reach of an emerging Caribbean AI and energy hub, demonstrates what happens when capital moves faster than the institutional capacity needed to govern it. In the neighboring Dominican Republic, a public-private investment exceeding $500 million is positioning the country as a regional hub for artificial intelligence and digital connectivity. Google’s planned Digital Port will connect the country directly to two Google Cloud regions in the United States and provide capacity for additional international submarine cables. A separate proposal, Project Hostos, would carry electricity generated in the Dominican Republic to Puerto Rico through an undersea transmission cable. The project has received a U.S. presidential permit but still requires additional local approvals and commercial agreements.

These projects are not components of a single coordinated investment plan. Read together, however, they show how public and private capital is beginning to reshape the economic and energy relationships among neighboring islands. The Dominican Republic is being positioned as a producer and regional infrastructure hub. Puerto Rico may become a purchaser of electricity generated elsewhere, even as its own residents continue constructing resilience household by household.

At the end of 2025, nearly 192,000 rooftop solar systems were operating across Puerto Rican homes and businesses, representing 1,456 megawatts of capacity and helping to stabilize the island’s wider grid. This growth has been enabled partly by net-metering policy and public support, but it has also accelerated because residents cannot consistently depend on the centralized electricity system. It is therefore less a story of coordinated grid transformation than of community improvisation filling the space where stronger intermediary infrastructure should have been.

The proposed connection to the Dominican Republic did not cause Puerto Rico’s rooftop-solar boom. Rather, the boom reveals the unresolved institutional environment into which this next wave of infrastructure capital may arrive. The intermediary gap in Puerto Rico is not the complete absence of institutions. It is the failure to build sufficient coordinating capacity among federal funders, Puerto Rican agencies and regulators, utilities, private developers, municipalities, technical experts, and affected communities.

The same structural condition appears differently in climate finance: SMEs often lack the technical knowledge and tools needed to compile consistent, credible, and investable data, creating a major bottleneck to reducing perceived risk and unlocking capital.

Kenya’s Vermi-Farm Initiative demonstrates both the scale of what investors risk overlooking and the role intermediary support can play. The enterprise currently supports more than 22,600 farmers, 87% of whom are women and youth. More than 14,200 farmers are using regenerative agricultural practices through its programs, which have processed over 265,000 kilograms of food waste and distributed more than 180 tonnes of organic fertilizer.

Yet co-founder Royford Mutegi explained that one of the startup’s earliest challenges was obtaining “access to the right resources in terms of technical expertise” to shape the data Vermi-Farm was already generating and clearly articulate its impact to investors. Nonfinancial support from organizations including the International Trade Centre helped the enterprise construct standardized models for traceable data and access investor-readiness programs. This enabled Vermi-Farm to produce, in Mutegi’s words, “data that is investable, data that can be traced, data that can be backed up” and attract attention from funders such as Livelihood Impact Fund, the Roddenberry Foundation, and the Z Zurich Foundation over the past two to three years.

Where Puerto Rico shows what can happen when capital enters without the intermediary capacity needed to coordinate its deployment, Vermi-Farm demonstrates what becomes possible when that layer is deliberately built.

CLOSING THE INTERMEDIARY GAP

Closing the intermediary gap requires more than inviting overlooked actors into the room. It requires mechanisms that embed their participation and value into the architecture of the transaction itself. In the AI infrastructure buildout, Community Benefit Agreements paired with legal and technical assistance can perform this function. In climate finance, the closest counterpart is blended finance combined with portfolio aggregation and capacity-building.

A Community Benefit Agreement (CBA) can convert community input into enforceable project terms before construction proceeds. Rather than relying on voluntary promises, communities can negotiate commitments concerning environmental impacts, infrastructure costs, local benefits, transparency, monitoring, and accountability. But an agreement is only as meaningful as the process that produces it. If residents cannot interpret energy forecasts, water requirements, utility agreements, or environmental assessments, a CBA may simply formalize terms negotiated under the same imbalance of information and capacity described earlier. Technical assistance is therefore not an optional supplement to the agreement. Communities need access to legal counsel, technical expertise, and comparable project information before consequential decisions are fixed.

Blended finance structures address a parallel imbalance in climate investment by changing how risk is allocated. Guarantees, first-loss capital, and other concessional layers can absorb a defined portion of potential losses, improving the risk-return profile for commercial or institutional investors. This does not make an enterprise risk-free, nor does it substitute for a viable business model. It creates room for investors to evaluate commercially promising SMEs without requiring each one to possess the collateral, transaction size, reporting systems, and long track record expected of a large corporate borrower.

Yet risk-sharing alone is insufficient. Blended finance structures are most effective when they combine appropriate risk allocation with technical assistance, clear impact metrics, strong implementation partnerships, and portfolio aggregation. For example, SME loans can be pooled into a diversified portfolio and supported by a first-loss tranche or guarantee, helping spread risk, reduce transaction costs, improve underwriting efficiency, and achieve the ticket size institutional investors require. But financial structuring must be matched by practical support that helps SMEs measure and communicate their performance in terms legible to investors. Rather than imposing reporting systems designed for large corporations, this support should allow enterprises to begin with foundational practices and strengthen their data and reporting as their capacity matures.

CBAs and blended finance structures are not identical instruments, but they are mirror images in function: both build actors routinely overlooked by conventional investment processes into transactions from the beginning. Capital becomes more effective when communities have the capacity to shape the infrastructure they will host and when SMEs have the financial and institutional structures required to demonstrate, finance, and scale the solutions they are already delivering.

FIRST PRINCIPLES FOR INVESTORS

The mirror also applies to how investors think and act in their efforts to deploy AI without pushing its costs onto communities, and accelerate climate action by backing the enterprises already delivering it.

1. Intermediary capacity is the binding constraint—and the investment opportunity. Investors cannot assume that capital will distribute itself effectively once it has been committed. They must also fund the architecture that carries it to local actors: community organizations, legal and technical advisers, SME pipeline builders, investor-readiness programs, local financial institutions, data and reporting systems, and environmental impact assessors. In climate finance, this also means participating in blended finance structures that give intermediaries the capacity to originate, aggregate, and support transactions conventional investors would otherwise overlook.

2. Information asymmetry is the first barrier to effective capital direction. In data center siting, nondisclosure agreements can prevent communities from learning deal terms before permits are approved, while utility rate structures are often presented in forms difficult for non-specialists to interpret. In climate finance, the information gap runs in the opposite direction: investors often cannot find or compare promising SMEs because opportunities are fragmented across markets and reporting systems are designed around the capacity of larger companies. In both cases, capital is allocated without the knowledge needed to understand its consequences or recognize its most effective uses. Investors must therefore support early disclosure, accessible technical translation, SME pipeline development, and proportionate reporting systems before decisions are locked in.

3. The timing is structural. Private equity investment in digital infrastructure typically operates on a five-to-seven-year fund cycle. By the time community impacts become measurable—in electricity bills, in water stress, in grid load—the original investment may already be repricing or exiting.

The AI chips driving today’s buildout depreciate faster than accounting recognizes, often becoming obsolete within two to three years despite being carried on company books for five or six. Communities signing agreements today are therefore committing to infrastructure built around technology assumptions that may shift before the promised benefits fully materialize.

The physical infrastructure operates on the opposite timeline. In regulatory accounting, transmission lines, substations, and capacity additions can remain in the rate base for thirty to forty years. Capital may exit in year five and hardware may turn over in year two, but the grid built to support the facility remains, and its costs continue to be distributed across ratepayers.

Communities are stranded on the longest timeline, with the least leverage, at the moment it matters most. A CBA is a mechanism that could align these clocks by establishing obligations, protections, monitoring, and accountability that survive changes in technology, ownership, and investor participation. That is why the absence of CBAs at scale—and of the technical assistance communities need to negotiate and enforce them—remains a structural condition the industry must address, even as state regulators move to protect ratepayers through large-load tariffs.

Climate SMEs face a different version of the same mismatch. Investors often expect them to demonstrate the collateral, reporting systems, operating history, and profitability of mature corporations before patient capital and capacity-building have given them time to develop those capabilities. Enterprises are effectively denied financing because they do not yet possess the evidence that appropriate financing would help them generate.

Aligning these clocks requires capital structured around how viable enterprises actually grow: patient and, where possible, local-currency finance; guarantees or first-loss layers that absorb early-stage risk; technical assistance that strengthens operations and data over time; and reporting requirements that progress as enterprise capacity matures.

Across both investment waves, the investor clock cannot be the only clock that matters. The actors closest to the consequences must be built into the transaction before its terms are fixed—not consulted after the risks, benefits, and responsibilities have already been allocated. Otherwise, investors may deploy capital at historic speed while financing fragility, opposition, and missed opportunity.

The window is still open. Not for long. But open.

About This Piece

This OpEd is co-authored by two researchers working from different vantage points on the same structural problem. The piece argues that the AI infrastructure buildout is surfacing — at unusual speed — the same structural failures that have long prevented capital from reaching SMEs providing local solutions to communities in the Global South, and low-and middle-income (LMI) communities in the U.S. context. It is addressed to impact investors, foundation program officers, and family office advisors as a call to engage, not a critique of inaction.

Jorge Luis Fontanez is the founder of Compound Impact Collective and author of Compound Impact on Substack, tracking capital flows at the intersection of AI infrastructure, climate finance, and community development. He is the former CEO of B Lab U.S. & Canada and an inaugural member of the Global Climate Finance Forum.

Meghna Parameswaran is the Project Coordinator for the Secretariat of the Global Climate Finance Forum, which aims to rewire global climate finance to empower SMEs driving climate innovation across the Global South.

‍

‍

‍