You’re out of free articles.
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
Sign In or Create an Account.
By continuing, you agree to the Terms of Service and acknowledge our Privacy Policy
Welcome to Heatmap
Thank you for registering with Heatmap. Climate change is one of the greatest challenges of our lives, a force reshaping our economy, our politics, and our culture. We hope to be your trusted, friendly, and insightful guide to that transformation. Please enjoy your free articles. You can check your profile here .
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Subscribe to get unlimited Access
Hey, you are out of free articles but you are only a few clicks away from full access. Subscribe below and take advantage of our introductory offer.
subscribe to get Unlimited access
Offer for a Heatmap News Unlimited Access subscription; please note that your subscription will renew automatically unless you cancel prior to renewal. Cancellation takes effect at the end of your current billing period. We will let you know in advance of any price changes. Taxes may apply. Offer terms are subject to change.
Create Your Account
Please Enter Your Password
Forgot your password?
Please enter the email address you use for your account so we can send you a link to reset your password:
That won’t stop these investors from trying.

Sometimes it’s called the “missing middle,” sometimes, more ominously, the “valley of death.” Whatever the terminology, it’s undeniable that a chasm lies between a climate company’s early funding rounds and its eventual commercial scale-up, one that’s getting harder and harder to bridge. From the first half of last year to the first half of this one, total Series B funding declined by nearly a quarter; beyond Series C rounds — what the market intelligence platform CTVC calls “growth funding” — it declined by a third.
“The capital needs of these businesses have just outgrown their early stage backers,” Frank O’Sullivan, a managing director for S2G Ventures’ energy investments, told me. “But the infrastructure investors have absolutely no appetite whatsoever for taking on an unproven technology and scaling.” S2G makes both early stage and growth stage investments, and O’Sullivan co-authored a white paper last year on the problem of the “missing middle.” The paper found that of the $270 billion in private capital for clean energy raised between 2017 and 2022, just 20% was allocated to late-stage and growth-focused investments, while 43% went to earlier rounds and 37% toward deploying established tech.
Of course, some of climate tech’s funding gap can be attributed to broader trends in the venture market and economic landscape. Covid-related disruptions and low interest rates led investors to throw money at promising startups, only to see their valuations drop as inflation (with rising interest rates to match) and geopolitical uncertainty cooled down the overheated market. Other companies went directly onto the public market via special purpose acquisition companies, only to underperform expectations. “There is capital to be deployed,” O’Sullivan told me. “But a lot of the companies that need that capital are struggling, really, to swallow hard and take significant restructuring of their previous valuations.”
With clean tech in particular, there’s also frequently a mismatch between the abilities of venture firms, which often make their biggest returns on software startups, and the demands of climate tech. The latter tends to require huge investments in physical infrastructure and support for first-of-a-kind projects, and generally has a longer timeline to profitability than, say, an app. “Venture funding, in some sense, was built for scaling software companies,” Lara Pierpoint, managing director of the new catalytic capital program Trellis Climate, told me. “You’re talking about a capital light business that generally is creating something that enters a white space, and for which there’s huge amounts of market potential.”
It’s much more difficult to build expensive infrastructure that aims to displace fossil fuel facilities and the entire economy that relies on the cheap, reliable power they provide. So while VCs may be enthusiastic about taking a relatively small financial bet on a high-potential early-stage company, that may be all they’re able to do.
Trellis, on the other hand, is a part of the climate nonprofit Prime Coalition and funds first-of-a-kind climate projects with philanthropic capital. The nonprofit structure and philanthropy-focused funding model mean that Trellis can take a different tack on missing middle financing than traditional venture or equity investors. For example, Pierpoint told me it can choose whether to invest in a company or just a specific project. Trellis can also help de-risk projects by providing an “insurance backstop” — basically backup capital in case primary project funding falls short. “We’re looking at expanding the kinds of resources and dollars we can bring to the table in general for the ecosystem, because we think that venture can’t do this alone,” Pierpoint told me.
As with all nonprofits, generating big returns isn’t the focus for Trellis. But for traditional investors, that’s the primary goal. And while growth investments in more technically mature solutions are likely to generate consistent returns, O’Sullivan told me they don’t often provide the rarer but more alluring 10x returns that make early-stage venture capital particularly enticing. “So it’s a more balanced portfolio, typically, in that growth equity category. It’s just that you don’t see the high highs,” he said, explaining that a two to 3x return on investments is more realistic.
Brook Porter, a partner and co-founder at the growth-stage firm G2 Venture Partners, told me that focusing on the missing middle can be extremely profitable, though, and that the key to making real money is correctly identifying a company’s “inflection point” — that is, when it’s poised for significant growth and impact. That is, of course, every investor’s dream. But G2’s whole strategy revolves around identifying exactly when this critical juncture will be, tracking more than 2,000 companies per year to identify the ones best poised for breakout scale-up.
The firm spun off in 2016 from Kleiner Perkins’ Green Growth Fund, where Porter and his three co-founders previously worked as senior partners. This is where they honed their theory of inflection point investing, funding companies such as Uber, drone-maker DJI, and Enphase Energy. Porter told me that helping startups move from proof-of-concept to building “that machine of a business” requires a lot of hand-holding, and that “there aren’t as many investors with that skill set,” so it could take a while for this approach to scale.
On the other end of the funding spectrum, large institutional investors like banks, hedge funds, and asset management firms certainly have the money to help bridge the missing middle, but O’Sullivan and Pierpoint told me they’re generally more interested in fulfilling their internal climate mandates by building out more wind and solar, which generates near-guaranteed returns. These investment giants then look at their remaining cash and think, “Well, we should do something more avant garde. Let’s put money into early-stage venture,” O’Sullivan explained. That’s how many seed and Series A-focused funds raise money.
As O’Sullivan sees it, what’s happening now is “a flaw of the structure of capital allocation at the very highest level.” He thinks we could start by reorienting incentives such that large investors such as banks, asset managers, and pension funds get paid in part for helping bring new climate solutions to market, as opposed to just funding the same old, same old. That would allow them to write “right-sized checks” on the order of $50 million to $100 million to ready-to-scale companies — larger than what a VC firm would write, but smaller than what the big infrastructure investors are used to.
How would those alternate funding models actually work? Well, that’s the real question. Pierpoint said she’s often asked whether a new kind of investor or asset class will be necessary to fill the gap, and while she doesn’t have an answer, what she does know is that the group of climate tech companies that’s ready to commercialize “can’t wait 15 years until we have the exact right form of capital.”
“There needs to be urgency on the part of philanthropists, on the part of infrastructure equity investors, on the part of venture capitalists, to really start showing that we can do this,” Pierpoint told me, “and that we can bring together the right capital stacks to make this happen.”
Log in
To continue reading, log in to your account.
Create a Free Account
To unlock more free articles, please create a free account.
A new report from a coalition of energy and data analytics organizations offers recommendations for the country’s demand response leader.
By many measures, California is the most advanced U.S. demand response market. Its aggressive clean energy targets, widespread home electrification, and near-universal smart meter deployment make it a natural testbed for programs that call upon distributed energy resources — from home batteries and electric vehicle chargers to smart thermostats — to ease grid strain and pay customers for helping out.
The state has been running these initiatives in one form or another for decades, starting with agreements that paid commercial and industrial customers to cut their power during periods of grid stress. Over time, those programs expanded to households, allowing ratepayers to let utilities cycle their air conditioners on and off and, eventually, control their smart thermostats too. But the theoretical potential of California’s demand response strategy has far outpaced the realized grid benefits.
“Load flexibility has underdelivered for a long time,” Ric O’Connell, executive director at the grid policy nonprofit GridLab, told me.
A new joint report from GridLab, data analytics firm Kevala, and the energy consulting firm Energy and Environmental Economics released on Tuesday argues that California’s early-mover advantage has, in many ways, become a liability. While the technology to run more effective, streamlined demand response programs has finally arrived, decades of legacy initiatives have left the state and its confused consumers tangled among dozens of fragmented offerings, outdated compensation structures that don’t reward active participation, and rules that make it unnecessarily difficult for small, household devices to participate in wholesale electricity markets.
“The communications, the control, the metering — none of that stuff was really available 10 years ago, and you just sort of paid people to sign up,” O’Connell told me. “And then we didn’t really switch it as the technology became available for better measurement.”
But now that the technology is better, the report points out that the opportunity is bigger than ever: California has an unprecedented base of smart, connected devices — including millions of EVs, electrified buildings, and home batteries — that, if properly harnessed, could help smooth out the state's electricity demand and avoid the kind of costly new infrastructure buildouts that drives up everyone's rates.
One of the primary recommendations in the report, titled “Unlocking California’s Flexible Load,” is to pay customers for the actual value they provide to the grid — such as how often and for how long they reduce or shift their electricity use during demand response events. While that may seem obvious, historically, utility and state programs have paid customers simply for signing up and remaining "available" to cut power use — regardless of whether they actually deliver when called upon. That model made some sense before smart meters and other tools could verify performance, but today it often just wastes money while failing to deliver meaningful load reductions.
Changes like this could help California capture far more of the value demand response has long promised. A 2024 GridLab study with The Brattle Group found that virtual power plants — networks of distributed resources that collectively act like large, traditional power plants — could save California utilities and consumers $550 million per year while meeting more than 15% of the state’s peak electricity demand.
The potential is especially striking with EVs. Their charging patterns can already help shift overall electricity demand to less grid-constrained hours, while bidirectional charging may one day turn them into giant grid batteries capable of sending power back to the grid — an increasingly common capability known as vehicle-to-grid, or V2G. The report reveals that if just 10% of California’s projected 9.7 million EVs participated in V2G programs, they could supply nearly a third of the state’s 2036 long-duration battery storage target, according to a press release about the report.
As the report also makes clear, though, getting there will require more than simply changing how the program pays customers. Another major recommendation is consolidating the programs and streamlining how they’re administered. O’Connell said the utilities running their own programs — long held back by institutional inertia — are beginning to recognize the inefficiency problem, waking up to the fact that “the person doing the smart thermostat program is in a different department than the person who’s doing the behind the meter battery program,” he told me, explaining that he’s already working with Con Ed in New York to consolidate its offerings. Based on his conversations with California’s utilities, he said he expects them to announce consolidation plans soon, as well.
O’Connell said the utilities running the programs — long held back by institutional inertia — are beginning to recognize the inefficiency problem, waking up to the fact that “the person doing the smart thermostat program is in a different department than the person who’s doing the behind the meter battery program.”
It can be a hard sell to get the investor-owned utilities to put real muscle behind these programs, however, as they make money by building new infrastructure like large power plants, not by avoiding the need for it through demand flexibility.
“I think in many ways the IOUs have been indifferent to load flexibility. It’s not core to their business,” O’Connell told me. But with political tension over affordability mounting, customers increasingly worried about electricity rate hikes, and huge new large loads like data centers seeking to connect to the grid as quickly as possible, utilities are facing more pressure than ever to make better use of the infrastructure they already have.
Another core recommendation is designed to ensure that demand flexibility programs actually benefit all customers by capping customer compensation below the total cost that the utility avoided in new infrastructure buildout. For example, if a customer’s individual participation in such a program saves a utility $100 in spending, they should receive less than $100 for providing that flexibility. This is designed to ensure that all California customers end up saving on their utility bills, regardless of whether they’re able to flex their loads or not.
This particular recommendation comes in response to a problem the state encountered with its legacy rooftop solar compensation system, Net Energy metering, which ran from 1996 to 2022. The program pays existing solar customers, who have been grandfathered into the program, well above the actual value of the power they export to the grid, thereby shifting billions of dollars in costs onto customers without solar.
Lastly, the report recommends creating a simpler path into wholesale electricity markets. While sophisticated players —- think large businesses or major demand response aggregators such as Voltus or Sunrun — can sell load reductions directly into those markets, the process remains too complicated and paperwork-heavy for smaller aggregators bundling together resources such as household EVs and batteries. For now, the report argues, those smaller players should keep enrolling customers through simpler, utility-run programs while regulators work to make wholesale market participation more accessible.
Ultimately, O’Connell hopes the report can help California move past the institutional battles that have historically held demand flexibility back. “One of the problems with California is there’s no kind of neutral,” he told me. “We were trying to be that neutral party that’s like, here’s the roadmap to get everyone to actually unlock this potential.”
The goal, he said, was to “name all the problems of the past” — and, in doing so, give California’s utilities, regulators, aggregators, and customers a clearer path forward.
Current conditions: Lake Powell just dropped to its lowest level since the reservoir straddling the border between northern Arizona and Utah began filling 60 years ago • A dangerous new heat dome has formed over the American Southeast, driving midday highs north of 110 degrees Fahrenheit in cities such as Jacksonville, Florida • Temperatures in Bandar-e Mahshahr are rising past 124 degrees, making the Iranian port city at the northern end of the Persian Gulf, near the border with Iraq, the current hottest place on Earth.
Less than two weeks ago, Amazon confirmed its plans to build a data center complex powered by a 7.65-gigawatt, off-grid natural gas plant. As my colleague Emily Pontecorvo wrote, the facility would handily surpass the output of the nation’s biggest power station, the 7-gigawatt Grand Coulee hydroelectric plant in Washington State, and Georgia’s Plant Vogtle, which recently vaulted to No. 2 after the completion of the country’s only two wholly new nuclear reactors in decades increased its output to nearly 5 gigawatts. An even bigger gas plant is now eyeing the top spot on the list. On Monday, ChatGPT-maker OpenAI inked a deal for a sweeping new data center campus in Ohio, backed by $105 billion from chipmaker Nvidia. As part of the agreement, SoftBank’s SB Energy will construct a 9.2-gigawatt gas plant that will be owned by the U.S. government and financed by Japan, according to The Wall Street Journal. “Today, we are helping secure the critical infrastructure required to build these factories,” Jensen Huang, Nvidia’s chief executive, wrote in a blog post on the company’s website. “We are investing in the long-lived foundations of AI factories so our customers can deploy the most productive compute platform in the world, generation after generation.”
The biggest impediment, at least according to North America’s quasi-governmental grid watchdog, is power. “The only thing China is ahead of us in the AI race is power,” Jim Robb, the chief executive of the North American Electric Reliability Corporation, told reporter Arianna Skibell on the Politico Energy podcast episode that went live Monday. “We have better models, we have better engineers, we have better scientists — but we’re challenged in our society to build the infrastructure that’s going to be required to support the growth.”
Europe’s hellish summer continues to shatter records. Just weeks after wildfires scorched Spain and France in what the French president called the country’s “hardest” challenge “since World War II,” Belgium is now battling its biggest blaze in recorded history. Hundreds fled as the flames approached the German border, though rainfall on Monday helped slow the spread. But the High Fens fire has already exposed political fissures in the country. On Monday, Belgian Defense Minister Theo Francken blamed anti-American sentiment for preventing the government from purchasing Chinook helicopters that would have strengthened the country’s firefighting capacities, according to The Brussels Times, an English-language news website. In the Flemish-language Het Laatste Nieuws, the country’s most widely circulated newspaper, columnist Isolde Van den Eynde complained that the episode highlighted the gap between how much government infrastructure exists for climate policy and how little there is for actually dealing with warming-fueled disasters. “While quite a few citizens are wondering where our little army of climate ministers is,” she wrote, “soldiers are on the ground.”
Hawaii, meanwhile, was still reeling from the first hurricane to damage the Big Island in more than a century. Tropical Storm Lala, which strengthened into a Category 1 storm at its peak, knocked out power for nearly 200,000 homes and businesses across the state. As I told you yesterday, the utility that covers 95% of Hawaii has warned it could be months before power is restored. Today we got a clearer sense of the other damage. More than 100 homes were washed away in the storm, and the damage to roads and bridges, according to The New York Times, cut off access to a town with the only hospital in its region.

Exxon Mobil’s oil fields off the coast of Guyana are booming, generating nearly $5 billion in profit last year and only expanding. Chevron last summer spent $53 billion to buy Hess and gain a foothold in the once-poor nation on South America’s Caribbean shores. It’s no wonder The Economist declared South America “the world’s hottest oil patch” last summer.
Now America’s oil goliaths are looking across the Atlantic for their next windfall. On Monday, the Financial Times reported that Exxon had revived its plans to build a liquified natural gas plant in Mozambique’s restive Cabo Delgado, despite the threat of terrorism from an Islamist insurgency in the region. At the same time, Chevron confirmed to Reuters the discovery of new oil and gas deposits in one of its blocks off the coast of Angola, the second-largest producer in sub-Saharan Africa.
Sign up to receive Heatmap AM in your inbox every morning:
Sunrun built America’s biggest business selling and leasing residential batteries and solar panels on the promise of going off-grid and helping homeowners produce enough power to pare down their utility bills. Now the company is doing the same for data centers. On Monday, the San Francisco-based giant announced a deal with the virtual power plant provider Voltus to provide access to its thousands of residential solar-plus-storage systems in the PJM Interconnection and Midcontinent Independent System Operator electrical grids, covering much of the eastern half of the lower 48 states. “We are providing critical capacity from home batteries supported by funding from hyperscalers,” Sunrun CEO Mary Powell said in a statement. “This is just the beginning of what distributed energy assets can achieve.”
Good news for some of my friends over at the farmer’s market in my neck of Brooklyn: New Jersey is preparing to allow farmers to harvest sunlight for crops and electricity. On Monday, the New Jersey Board of Public Utilities voted to award 16 projects totaling more than 52 megawatts for the state’s first agrivoltaics program. Over the next three years, the program will scale up to more than 200 megawatts of projects. “This pilot can help agriculture and the solar energy industry learn if active agriculture use can be a renewable energy partner in shaping New Jersey’s future,” New Jersey Secretary of Agriculture Ed Wengryn said in a statement. “Getting these projects operating is the best real-life laboratory to learn the challenges the two industries face.”
Manila is a striking metropolis with ancient-looking Chinese and Spanish colonial architecture, gleaming new towers, and vast new neighborhoods forming out of landfilled parts of its eponymous bay. When I visited for a reporting trip in 2024, I learned that the name of the Philippines’ capital comes from the Tagalog phrase meaning “where there is nilad,” a type of flowering mangrove shrub that historically blossomed along the city’s riverbanks. Today those channels that line that city’s streets and wind through the world’s oldest Chinatown are filled with trash. Plastic bottles and garbage are common sights in a fast-growing economy held back by its limited supply of mostly dirty electricity. President Ferdinand Marco Jr. now says there’s “only” one solution to the pollution crisis: Burn it. Last week, his administration told The Philippine Star that new waste-to-energy plants could come online in as little as a year. Environmentalists who say incinerators will only add to air pollution are already pushing back. The government has put out a tender for up to 400 megawatts of capacity, Renewables Now reported. Meanwhile, in a sign of just how much the energy market is heating up in the country, the Philippines’ biggest renewables installer, First Gen Corporation, just turned down a bid from the American investment giant KKR, saying the offer didn’t match the installer’s surging value.
Europe, on the other hand, is seeing its hydrogen ambitions stall out. New analysis by Hydrogen Insight found that project timelines across the continent are now being pushed past two years, “with the number of projects expected to begin commissioning by the end of 2029 falling by almost two thirds.”
Something you don’t see every day: The Trump administration is defending a climate policy imposed by the Biden administration that environmental groups like against Republican states. Last week, E&E News reported that the Department of Justice had asked a federal judge in Louisiana to dismiss a lawsuit brought by 10 GOP state attorneys general in a challenge to a Biden-era policy that stopped subsidizing flood insurance for properties in places increasingly at risk due to new climate extremes.
OpenAI’s new Ohio data center will rely on the country’s largest fossil-fueled power plant — which will be built on federal land.
This is an edition of Heatmap Daily, an evening review of the day’s news written by our executive editor. Sign up for it here.
This morning, OpenAI announced that it is leasing an enormous data center facility that will be built in Pike County, Ohio. The facility’s ownership structure will be arcane, to say the least: It will be built on federal land, operated by a subsidiary of the Japanese firm SoftBank, and partially backstopped by the chip designer Nvidia. The project is the most significant example so far of the increasingly creative off-book financing that’s now driving the artificial intelligence boom.
For our purposes, though, what sticks out about the facility is not its financing per se but the scale of its energy demand. The supercomputer will consume 10 gigawatts of electricity, or roughly as much power as New York City demands on a summer day.
To supply this energy, the Energy Department will build and own … a 9.2-gigawatt natural-gas-burning power plant on-site. It will be financed by the Japanese government and operated by SB Energy, the SoftBank subsidiary. Although this power plant was announced back in March as part of President Donald Trump’s trade deal with Japan, it wasn’t as clear then whether it would actually get built. Nvidia’s involvement raises the odds that it will reach completion. (In any case, it will get built in stages.)
There are several notable things about this extraordinary — and enormous — power plant, assuming that it does get built. Upon completion, it would rank as the largest power plant in the United States, nearly 40% larger than the Grand Coulee Dam. It would also become one of the largest natural gas power plants in the world, rivaling the Jebel Ali Power and Desalination Plant in Dubai. The scale of natural gas throughput required to feed the plant will resemble that required for a large liquified natural gas export facility; simply feeding the plant everyday could eat up a sizable chunk of, say, Ohio’s overall natural gas production.
There’s much we still don’t know about this power plant as well, including what kind of turbine it will use. That question will play a big role in its overall greenhouse emissions and air pollution footprint — although no matter what it will become a major polluter.
It will inaugurate, as well, a new era of national mega-gas plants. We learned earlier this month, for instance, that Amazon is behind a 7.65-gigawatt gas-burning facility being built in Texas dubbed Gigawatt Ranch. That enormous plant, if built, will also outrank the Grand Coulee Dam. (The market research company Cleanview first reported Amazon’s involvement in the facility.) The data center developer Nexus has proposed a 6-gigawatt gas-burning facility near Hubbard, Texas, as well — another enormous power plant. Since the beginning of the fracking boom, natural gas has been distinguished in part by its highly modular nature: For both regulatory and technical reasons, it’s been possible to erect a gas-burning power plant in a variety of sizes in a variety of places on the grid. The rise of these newly behemoth gas-burning facilities suggests that we might be in a new era of truly behemoth gas development.
And what makes the Ohio facility different from the Texas examples, too, is that it's going to be owned by the U.S. government. It's essentially going to be a public natural gas-burning power plant. That has interesting implications for climate and energy policy, because the government’s involvement could bring it under the auspices of future federal regulation — or even executive authority. While its continued operation will likely be protected by two-way federal contracts with Nvidia, SB Energy, and other counterparties, the Trump administration has already stretched the bounds of contract law to allow for, let’s say, entrepreneurial federal policy making on its chosen issues. AI is not exactly popular as is. In a different political moment, with a different mandate, how might a future Democratic president look at this site?