The Convergence Imperative
Climate Finance and AI Infrastructure are the Same Fight
I remember watching in disbelief on September 1, 2021, as Hurricane Ida turned the Vine Street Expressway into a river in the heart of Philadelphia. The Schuylkill River crested at nearly 17 feet. In just six hours, up to 10 inches of rain exposed what decades of infrastructure neglect had produced: a city unable to handle what climate scientists warned would be inevitable.
The federal government eventually allocated $163 million for recovery and flood prevention. Yet, Philadelphia city officials last estimated it has $1.2 billion in unmet costs from that single storm, where FEMA verified 11,000 homes were damaged. Residents couldn’t afford repairs before Ida, let alone after. That gap—between what communities need and what policy delivers—is where climate investments keep falling short.
Last November, Governor Josh Shapiro made that gap harder to close.
Shapiro signed a budget deal, withdrawing Pennsylvania from the Regional Greenhouse Gas Initiative. Participation in the multi-state initiative would generate revenue and place critical limits on carbon emissions and increased energy savings. When climate policy conflicts with political expediency, climate loses.
Under both Democratic and Republican leadership in one of the nation’s most flood-prone states, climate policies continue to fall short of what residents need.
Pennsylvania is courting artificial intelligence infrastructure that will dramatically increase electricity demand. Amazon announced $20 billion in data center investments across the state. According to PJM Interconnection, our regional grid operator, data centers could drive electricity demand up 60% by 2030.
That surge will drive costs up for Pennsylvanians. Natural gas powers 40% of data center electricity. When utilities expand capacity, households bear the cost. Without guardrails, Pennsylvania risks locking in decades of fossil fuel infrastructure at a time when we need to be cutting emissions.
Let’s be clear about the trade-offs: economic development fueled by energy infrastructure worsens the environmental impacts Philadelphia is already struggling to manage, while eliminating a potential funding mechanism for climate resilience.
Did Pennsylvania choose Silicon Valley over the Schuylkill River? Have we traded climate policy for server farms and called it progress? The next storm will reveal the answer.
THE PATTERN: COMPOUNDING INFRASTRUCTURE BURDENS
Much more than $1.3 trillion USD was mobilized amidst the COVID-19 pandemic. $1.3 trillion happens to be the same amount we need to fuel the next five years to ensure our most vulnerable communities globally are more resilient and adaptive to climate change. The financial challenge is not for lack of resources or the ability to move quickly and at scale. Political will holds us back.
Worldwide, data centers now consume approximately 1.5% of global electricity, which is projected to double by 2030. A single large facility can consume five million gallons of water daily, equivalent to the needs of a city of 50,000 people. In 2023, data centers caused approximately $6 billion in public health damages from air pollution in the United States alone. The burden falls disproportionately on Black and Indigenous communities, with growth in rural areas and tribal nations having limited regulatory oversight.
Whether you are an environmental activist, a tech entrepreneur, or a resident ratepayer, we are now facing the same problem. It is an age-old fight for land, water, and public health.
Carbon emitted anywhere affects climate everywhere. But local air pollution from fossil fuel combustion is different—it concentrates in specific communities. That’s why optimizing AI operations solely for carbon reduction doesn’t address the immediate public health costs borne by nearby residents. Pollution travels, but its damage isn’t evenly distributed.
The same AI model, trained with the same energy, yields health costs that vary by a factor of 10 depending on where it runs.
Source: IEEE Spectrum, 2025
Pennsylvania is already absorbing costs from infrastructure decisions made elsewhere. For example, backup generators at Northern Virginia data centers create pollution that drifts into Maryland, West Virginia, Pennsylvania, New York, New Jersey, Delaware, and DC, contributing $190 to 260 million annually in regional public health costs—and that’s at current usage levels. If these generators emit at their maximum permitted capacity, costs could reach $1.9 to $2.6 billion annually. That is why decisions on where to build data centers are also public health decisions.
Every age of innovation repeats this cycle. This one might be the most severe as the green transition and nationwide buildout of digital infrastructure converge. Communities face compounding infrastructure burdens. The lack of funding and political will to operationalize climate solutions will only add to trust deficits in AI deployment.
THREE GAPS KEEPING CAPITAL OUT
I have identified three systemic gaps that prevent capital from reaching the communities that need it most:
The Infrastructure Gap: Who Pays for Progress?
Each industrial revolution—railways, mass production, television, the internet—created massive disruptions. Those most invested in old ways rarely grasp how those ways will become obsolete. First movers benefit most.
Since the Rockefellers, Vanderbilts, and Carnegies built the railroads, new infrastructure creates concentrated wealth. That wealth creates political power and influence, leaving communities to bear the transition costs.
In economic terms, these costs are negative externalities. They are not accounted for on any company balance sheet. Inevitably, the public pays.
We need digital infrastructure—but we have a responsibility to build it in ways that mitigate environmental harms and meet community needs.
For instance, a recent study highlights Pennsylvania stormwater drainage capacity requires $7.8B, one of eight required investments. Instead of upgrading school air-conditioning systems, schools are closing due to rising heat and air-quality issues. The need for investment is as urgent as the need for long-term, community-informed strategies that improve climate resilience.
Greening our infrastructure also requires patient capital. The clawback on the EPA’s $3B Environmental Justice Block Grants is another example of communities being kept out of the decision making. Meanwhile we now know that optimizing data center operations for carbon reduction can actually increase public health costs in nearby communities. The same AI model, trained with the same energy, yields health costs that vary by a factor of 10 depending on where it runs.
Even though siting decisions are public health decisions, the public rarely gets a voice in the back-door agreements that determine where data centers will emerge.
Often, the political processes used to vet digital infrastructure investment strategies in the public domain favor corporate stakeholders over community insights. It promotes the idea of economic growth in a region to help address what communities need now so that they can invest in more sustainable practices later. History tells us a different story.
The Trust Gap: Unkept Promises
Data center announcements are often buttressed by elusive promises of jobs, economic development, and affordable energy benefits that may never materialize for certain communities. Data centers rarely create quality jobs for local residents in under-resourced communities—short-term construction work doesn’t substitute for permanent operational roles that often go to out-of-state talent.
There’s a massive trust gap in the AI boom in the U.S., both in terms of resistance to AI adoption and negative impact on low and moderate-income communities. 65% of Americans living in the bottom quartile feel they will be left behind and not realize gains from AI. They have every right to be skeptical. In Mansfield, Georgia, residents report that data centers are draining wells and leaving homes without potable drinking water. Aspirations for how the data center would transform this community differed from its actual impact.
How can disparately impacted communities advocate for themselves when they lack the information and agency to intervene before decisions are made? Public interest organizations fill in a gap when publicly elected officials ignore community experience.
The Intermediary Gap: We’re Missing Voice at Scale
Whether I’m speaking with community groups, policymakers, or tech leaders, there’s a common desire to focus on what’s possible and imagine a future with optimism. We need better narratives. And when we can embrace a new story and bring communities along for the ride, by including their voice, everyone benefits—or at least that’s my hypothesis.
Philanthropy is often seen as a catalytic player when gaps exist. Yet all 50 states are considering AI legislation with no coordinated consumer voice. Tech companies spent $85.6 million on federal lobbying in 2024—a 26% increase from 2023—while philanthropic funding for community-centered AI advocacy remains nascent.
When promises are made to bring jobs, resources, and new benefits to communities starving for investment, community leaders often operate from fear rather than curiosity—they lack information to evaluate what’s real. They hear that policy threatens innovation, that AI and data centers won’t create local jobs, and everything is getting more expensive. And while the economics have shifted, they benefit the few.
Without coordinated voice, communities can’t make basic demands: that grid upgrades be paid through broader cost-sharing rather than passed on to residents, that states like Pennsylvania adopt clean energy procurement strategies to stabilize prices, that this cycle of innovation won’t abandon climate commitments to chase AI investments. The intermediary gap means communities can’t negotiate for economic opportunity that doesn’t come at the cost of climate resilience—even though both are possible.
PHILADELPHIA, PUERTO RICO, JAMAICA: BUILDING THE COLLECTIVE
As we witnessed at COP30 in Belém, where Indigenous protestors disrupted the largest climate conference, each deficit requires us to stop bypassing frontline community voices. Our culture of scale inhibits impact, when solutions do not account for, or adapt to, the needs of local communities. There’s no one-size-fits-all—we need to right-size our approaches.
As the former CEO of B Lab U.S. & Canada, I’ve seen that purpose and profit can coexist. The tools exist, and so does the capital. The question remains whether we build the power to distribute the benefits equally.
Climate finance has a parallel problem. The unnecessarily complex process of due diligence and blended finance investments burdens the entrepreneur, disadvantages small businesses, and impedes climate solutions. Program officers often have a strong advantage. They see the problems and the marketplace of solutions but are limited in their ability to act and become a bridge because of institutional constraints and strategic priorities.
Compound Impact exists to do exactly this—share insights, build collective power, and create the bridges capital needs to reach frontline communities.
Let’s build a collective.
In the months ahead, I’ll ground this work in three places: Philadelphia, Puerto Rico, and Jamaica.
Philadelphia has the highest poverty rate of any major American city. Puerto Rico is the poorest territory in the nation. Jamaica—chosen by the Global Climate Finance Forum as its annual meeting site because it requires no visa from more than 80 countries—is now the latest example of the need for climate action after Hurricane Melissa. These are places where poverty can be alleviated if we invest the right way and engage marginalized communities. That’s where I am focused.
I am conducting this research out loud because the pace of change is too fast to figure out alone. We need to learn from each other—how to build power, form coalitions, unleash capital, and advocate in unison.
Floods don’t check voter registration, and rising utility bills hit working families hardest regardless of party affiliation.
If this resonates, take one step: Reply or DM me with your biggest question about mobilizing capital for climate resilience. I’ll tackle reader questions in the coming weeks—and if you know someone who should join this conversation, tag them too.
Let’s keep asking critical questions out loud, testing what works, and building the collective power we need to act.