Currence

Starting to think of this as the **Energy Hoarding** era. Geopolitical turmoil, the AI race, the industrialization race. The old way to meet those challenges was just to build more generation. But it's hard/slow building these days. Lower social license (esp on powering data centers), NIMBYism, bureaucracy and policy risk, all of that creates an aura of anxiety, a snatch and grab attitude towards securing capacity. I went on the Currents podcast from Norton Rose Fulbright with my co-founder Mark Taylor to get into the weeds on the race to power AI data centers. What are the strategies that companies are taking to get on line as quickly as possible? How realistic is BTM gas? Can flexibility and community benefits (being a good grid citizen!!) help move things faster? Is the ultimate option, for the few that can afford it, to just buy the platform, like Google did with Intersect Power earlier this year. Incidentally, our research on The Fastest MW shows that purchasing solar + BESS projects out of the ERCOT queue is the fastest way to get online. Check out this great conversation and reach out if you want to read our research: https://lnkd.in/gT4z9H-Q

Micro-reactors are having a moment. For the first time, they're getting more early-stage investment than SMRs. Earlier this month, Valar Atomics raised a $1bn Series B, the largest micro reactor round ever ($750m of new money). Days earlier, Antares raised $470m, ~$1.2bn into micro-reactors in a single week, pushing them to an all time high and ahead of SMR developers. Why now? ⚖️ Regulation may finally make the business model possible. Until now regulation effectively prohibited the micro-reactor model. The NRC's proposed Part 57/70/71 rules from May would change that, they’ll let a developer build, fuel and ship a reactor as a sealed product, reform arduous site security and operations requirements, and enable whole fleet licensing with a single application and 6–12 month reviews instead of years. 🛒 Customers are starting to show up. Defence is first: the DoD put up to ~$300m behind BWXT’s Project Pele, while the Air Force selected Oklo to supply power to an Alaska base, subject to licensing. In data centres, Radiant has a binding Equinix preorder, with deposits, for 20 units. 🏛️ Washington has flipped from safety, to safety and innovation. Four vendors: Antares, Valar, Deployable Energy and Aalo, met the DOE’s July 4 criticality challenge. The reactors are still a long way from commercial deployment, but, for the first time in decades, and with strong and bipartisan political backing, the policy posture has clearly shifted toward actively building the industry. How big is the micro-reactor opportunity?  I see two markets: 1. Replacing diesel. Military bases, mines, O&G fields and remote communities. These customers already pay a premium because they aren’t on the grid. The open question is whether micro-reactor economics can beat diesel once fuel, logistics and financing are included. 2. Powering distributed data centres. I’m sceptical about micro-reactors for today’s giant AI training campuses: at GW scale, economies of scale dominate. But inference could be different. If compute increasingly moves to smaller (~10MW) facilities distributed around towns and cities, micro-reactors’ modularity, siting flexibility and potential deployment speed become much more valuable. That would create a much bigger TAM, and potentially more price-insensitive buyers, giving the industry an order book to take it down the cost curve. Big caveat: no micro-reactor is operating commercially yet, and both economics and HALEU supply remain unproven. For the full analysis and data, clients can use the link below.

Come work with us! We're hiring for two big roles to build out the commercial team at Currence. We're in a massive infrastructure supercycle and the biggest players in it use our data and research to make decisions that will shape the coming decades in power, data centers, and climate tech innovation. >> Head of Revenue (New York or London) Own the entire commercial function across sales, marketing, CS, account management, and RevOps. You'll inherit a working engine and an experienced team. You'll partner directly with our founders on where we grow. Client list of blue chip logos plus more in the pipeline. https://lnkd.in/gkUwBXMm >> Senior Account Executive (New York) Win new business across energy, infrastructure, and financial services. Our content and research generate a steady flow of inbound, so a lot of the game is turning that into revenue while opening new logos of your own. https://lnkd.in/geiuMrW9 Let's build an energy system for the future: faster, cheaper, better, cleaner.

Everyone's talking this week about the $730bn of hyperscaler capex, and whether it's a sign of the top or whether it's a bubble. I don't know. I don't think you do either. 🤓 But I keep thinking about a different number -- 30,000,000. That's the number of paid Copilot seats as of June, up from 20m in April. That's just about 7% of the roughly 450m paid M365 commercial seats globally. So, a tiny fraction, but it's accelerating, and has plenty of room to run inside all those thousands of corporates that are figuring out their AI strategies right now. That's what we're talking about in Powerstack this week. 👇 Why? It's because nobody signs a 15-year PPA against a demand curve. You sign it against a counterparty. Copilot sits inside software that corporates fund out of operating budgets and renew every year, which is about as good a signal as you get that Microsoft is still going to be good for it in 2040. And here's the thing. Very little of that $730bn actually buys generation. Microsoft spent $41bn in the second quarter and about two thirds went to CPUs and GPUs. The power capex is increasingly sitting on the balance sheets of the infra funds, and they're the ones looking at this credit call, right down to how many users are hooked on next-gen Clippy. Check it out. And if you're curious about our take on the Texas moratorium, Currence clients can see our analysis in your inbox or at the link in the comments. #datacenter #power

At Currence we've been tracking and making sense of the regulations around the US data center buildout: large load tariffs, moratoriums, bans, ordinances, and more. Our data shows that new moratoriums for data centers have increased dramatically over the first half of 2026, with the largest increase from 45 in May 2026 to 79 in June 2026 On Monday, Texas governor announced a 'moratorium' that placed a pause on all grid connected data center approvals. Here is what you need to know: -Texas is one of the largest data center markets in the world, this is a big deal. -Data centers now require approval and audit by the Texas PUC and ERCOT, requiring projects to disclose: public funding, annual and peak electricity consumption, annual and peak water consumption, cooling system type, owners and controlling interests, and a range of community impact reduction measures covering light, noise, and emergency responses. -No clear timeline, or guidelines on what is required to pass an audit, have been given. Timing could not be more political, with elections coming up in November and growing public concern around data centers. This looks more like a simple request for information rather than a moratorium. Without a clear timeline and specifications around power, cooling and design details, this will result in worse outcomes for ratepayers and communities. #datacenters #texas #moratoriums #largeloadtariffs #AI #power #ERCOT

Our second LDES Leaderboard is out now! We rank the top LDES technology providers using Currence’s proprietary deployment, cost, financing, and round trip efficiency data, and in this second edition, you can track the progress of the top names. Check it out:

We hereby officially begin the “mark your calendars and start overbooking yourself” phase for our favorite time of year — New York Climate Week! 🥳 We're back with our fourth annual Currence (CTVC) x Planeteer Capital's guide to the week: 📅 https://lnkd.in/gnWQpGc8 This year's guide is housed in Luma for maximum filtering plus sleek calendar and map views. Think of it as your friendly filter for the events most relevant to climate tech founders and funders specifically. Hosting an event you think should be on to the guide? Submit it directly through the calendar! Expect much more over the next few weeks, this is a living artifact that will keep getting richer as new events roll in. Subscribe to get those RSVPs in early. Stay tuned, we'll be posting what Planeteer's hosting soon. (Spoiler, there'll be lots! 😉) Sophie Purdom Kim Zou Mark Taylor Maggie Mullooly Holly Secon Allie Mullen Omar Smith #NYCW2026 #newyorkclimateweek #NYCW #climateweek

We're hosting a breakfast September 22nd during Geothermal Rising's 2026 conference to dive into the state of commercialization of advanced geothermal. Despite Fervo's new drilling records, hyperscalers scrambling for offtake, and capital pouring in, it's still early days for the sector. Guy Cohen will share Currence's proprietary research on the land grab, the top ranked developers, and how much of the announced pipeline is real. Register at the Luma link in the comments below ↓

Our new H1 2026 Climate Tech Investment Trends report is out! Climate tech VC hit $26.1bn in H1, up 55% year on year. H1 was a big one for climate tech investment and exits. Driven in large part by data centers. Low-carbon developers made up 34% of all funding, with DayOne ($4.5bn) and Nscale ($2bn) alone accounting for a quarter of the total. Clean power IPOs had a record half, led by Fervo Energy and X-energy. Bloomberg just wrote up the report. Huge shout out to Coco Liu. New to the report this year: we've expanded scope with a new Innovation Trends section covering robotics and space-based technologies alongside the core climate tech data. Read the full H1 2026 report here: https://lnkd.in/g7Rpc73c