Counterintuitive chart of the month: Bank tellers survived the ATM. They did not survive the smartphone.

Measuring poverty used to mean flying data collectors to remote villages every decade. A satellite AI model just mapped every 6-km area of Africa for 36 GPU-hours total, which is roughly the cost of a tank of gas. Cool paper by @marshallburke.bsky.social, et al.: arxiv.org/abs/2604.23166

The current U.S. grid: -44% cleaner than 1990 -Texas nearly matches the national average -Illinois almost cleaner than California (nuclear+renewables OG) -Coal states still high (though ND dropping thanks to wind)

Does new transmission actually move the needle on clean energy? Here's a month of California data before and after SunZia came online. Wind is up by a third, and gas down by a third. Judge for yourself.

The electricity bill for this image is higher than Scottie's salary in 1997

Over 85% of California's industrial heat demand is below 600°C. That's the threshold where electrification starts to look commercially viable. New research puts a facility-level number on just how much of the "hard-to-abate" sector might not be that hard.

Bad news: Climate change could eliminate more than 90% of the range of 1 in 10 plant species by 2100. Good news: Well... technically there isn't any. But this paper does have terrific figures!

Roughly 40% of the $6.8 billion in annual lifesaving benefits from Uber are pure externalities, per Anderson/Davis (2026): $2.7B in benefits to other drivers, pedestrians, and cyclists who never opened the app. That alone exceeds Uber's annual US ridesharing shareholder surplus.

Updated through today. The ceasefire announced April 5 knocked $35 off oil prices. Then Trump called it “on massive life support,” prices climbed back >$107, and the 2026 Hormuz line became something historical crises never were: a spike, a partial recovery, and a second climb.

This chart from our paper is the one I keep coming back to. Without IRA: 13-61 GW/yr of clean electricity IRA: 23-131 GW/yr OBBBA: 29-59 GW/yr Even OBBBA is near 2024 record levels, because demand growth changed the baseline. Policy matters, but so does the starting point.

Today is Florence Nightingale’s birthday. Beyond nursing, she used data visualization to prove soldiers were dying not from battle wounds but preventable disease. One of the earliest examples of a chart used to change policy. Deaths fell by 2/3 within a year.

The global EV transition in one chart: Leader: Norway Almost there: Netherlands, China Accelerating: U.K., France, Germany Starting: Thailand, South Korea Barely begun (yet still selling millions): U.S., India, Indonesia Doing its own thing: Brazil (ethanol)

11 models. Same scenario. 30 percentage points of spread in power sector CO2. Why? Assumed electricity demand, discount rates, temporal resolution, coal economics. Ranges >> point estimates.

Wild figure from a new paper on the large premium for regulated returns for natural gas and electricity utilities relative to capital cost benchmarks (dots are rate cases since 1980). Utility returns barely budged as rates plummeted after 2008. The gap is now 1-5 pp. Cost to consumers: ~$7B/yr.

New paper in Nature Reviews Clean Technology with @rynacui.bsky.social and co-authors: What did the IRA actually accomplish, and what does the One Big Beautiful Bill Act undo? We’re on Shift Key with @robinsonmeyer.bsky.social today breaking it down: heatmap.news/podcast/shif...

New on SHIFT KEY: The OBBBA swept away many of the climate incentives in the IRA after less than three years. What did that do to our climate trajectory? @bistline.bsky.social and @rynacui.bsky.social join @robinsonmeyer.bsky.social to discuss: podcasts.apple.com/us/podcast/s...

Projected increase in household energy costs from curtailing IRA's power sector credits: ~$55-640/year in 2035. States matter more than the national number.

Paper @science.org today led by Junna Wang and my @ucdavis.bsky.social colleague Xiaoli Dong. 7-16% of plants at risk of extinction under climate change, largely from vanishing climates, not slow dispersal. But plants will move, reshaping local species composition www.science.org/doi/10.1126/...

California has a lot of solar and energy storage coming. Texas has even more.

Interesting place for a battery manufacturing boom to happen, given the politics.

New paper: What did the IRA do to clean energy? And what does OBBBA change? Here's what 3 years of data and 11 models say: - $729B in clean energy investment (+92%) - Battery manufacturing up 8x - Solar investment up 87% - Wind investment DOWN 42% - ZEV purchases: $252B (+173%)

Three years of IRA. One Big Beautiful Bill. 11 models. 14 authors. Tomorrow we publish a review in @natureportfolio.nature.com synthesizing what happened, and what models say comes next.

Cool NBER paper: Data centers caused $25B in environmental damages in 2025 (VA is 1/4 of all current national damages). Planned facilities could add another $21B. But if AI boosts GDP by just 0.5%, the environmental damages from AI energy use amount to only 2.4% of those gains.

Everyone talks about the tension between growth and climate, but few measure it rigorously. This new World Bank chart is one attempt. It shows whether economic growth (accounting for the emissions it generates) is a net poverty reducer or net poverty creator, country by country.

PJM installed fleet: 197 GW PJM new queue applications, day one: 220 GW (106 GW of natural gas alone) The waiting room is bigger than the building.

Excited to see Watershed make TIME’s 100 most influential companies of 2026, one of only 10 in sustainability. I’m proud to work alongside some of the most rigorous, mission-driven scientists and engineers I’ve ever met. Still so much work to do. Let’s go.

California’s grid has real problems. California is also building more solar and energy storage than almost anywhere on earth. These facts coexist. The second one doesn’t get enough attention.

Impressive piece and graphics. The lighting figure brought to mind my favorite 700-year energy figure. The cost of lighting in the UK has fallen from £35,000 per million lumen-hours in 1300 to essentially £0 today. One of the most underrated progress charts ever made.

Join us tomorrow as we chat about energy systems and policy uncertainty! Register here: princeton.zoom.us/webinar/regi...

New peer-reviewed finding: Have you tried plugging it in? (Eight models confirm electrification of transport and buildings is the primary driver of emission reductions in every scenario consistent with the Paris Agreement… and can lower energy wallet spending)

OPEC told UAE they could pump 3.15M bpd, but UAE has 4.85M bpd of capacity. Today they quit. This chart explains everything.

Here’s a meme summary of our paper.

🧵 #ModelMonday We published a new paper on carbon-free electricity procurement. The most important findings are about how much your assumptions change the answer. A thread on what models reveal (and hide) about CFE. 👇

I’ll be at Stanford on Monday talking about AI, energy, and climate. Come by or tune in and yell at me online, a notification is a notification events.stanford.edu/event/stanfo...

The 2-hour marathon was this generation's 4-minute mile. People debated whether human physiology could do it. The once-inconceivable achievement was broken today not by one runner, but two! Sabastian Sawe just ran 1:59:30 in London.

People still talk about energy storage as the missing piece that will eventually make renewables work. In Arizona, it's already >20% of total capacity and still climbing. At some point "the future" becomes "this year's data."

Excited to join the MES Expert Panel next Thursday on Energy Systems Modeling under Policy Uncertainty How do we build robust decarbonization strategies when the policy ground keeps shifting? Come ask us. Register here: princeton.zoom.us/webinar/regi...

Interesting forecast. Here’s what the last 25 years looked like.

This is the real “disappearing scientists“ story.

Two things are true on Earth Day 2026: 1. We have a long way to go, especially on carbon emissions. 2. The progress since the first Earth Day in 1970 is staggering: U.S. air pollution down 60-95% through policy, innovation, and action. Pessimism isn’t more honest than optimism. Both are choices.

A jawdropping chart from the IEA. Text query: 0.05 Wh Large model: 0.31 Wh Agentic AI: 1.14 Wh Agentic + reasoning: 7.6 Wh Agentic + reasoning (full): 50 Wh This is why nobody can agree on AI’s energy future.

Leaked memos show the Supreme Court blocked the Clean Power Plan in 2016 when justices feared companies couldn't afford to comply. Those companies cut CO2 enough to hit the plan's 2030 goal, a decade early. The emergency that changed how America is governed turned out not to be one.

With the Boston Marathon today, a good time to re-up one of the greatest figures in sports data: the distribution of marathon finish times (n=9,789,093). The spike at 4:00 is not a coincidence.

One chokepoint: 14 million barrels a day and ~20% of global oil consumption. This is what was flowing through the Strait of Hormuz before the conflict, and why its closure is being called the largest energy supply disruption since the 1970s oil embargo.

Every major oil shock, put on the same timeline. The 1973 embargo was the longest lasting. The 1990 Gulf War was the biggest reversal. The 2026 Hormuz crisis is the steepest climb and drop. But the line isn't finished yet.

After nearly 20 years building energy system models at EPRI, I'm joining Watershed as Head of Science. The science of decarbonization is only as good as the decisions it drives. Excited to close that gap. watershed.com/blog/john-bi...

A 1200+ year climate record is at risk. Do you know know anyone in Japan who could help? See below.

In the special issue described in this editorial, working with several developing country teams, we articulated practical bottom-up guidance for reaching net-zero across the economy in key sectors: power, transport, industry and AFOLU, and in the context of countries' other goals @iddri.bsky.social

Bought the Artemis II Lego set ”for my kid.” This thing is stunning.

Artemis II is circling the Moon this week using 8.8 million pounds of SLS thrust. The Raptor 3 engine SpaceX is building outputs ~100 GW of power. The entire U.S. runs on about 400-700 GW depending on the season. We are living in a remarkable era of rocket engineering.