In partnership with
Hey everyone, it’s Thursday, FutureProof is back.
Apologies for the radio silence last week. Life briefly mistook my calendar for a demolition site, so there was no edition. Normal service, or at least my version of it, resumes today.
This week has a strong “fossil shocks are clean-tech marketing now” vibe.
Oil volatility is making electrification look cheaper, safer, and frankly more adult. Sodium batteries are edging into real EVs. African e-mobility is accelerating. AI is finding rare diseases, designing cleaner infrastructure, and, somehow, wandering into generative biology.
Because apparently the future took one week off and then returned with a spreadsheet, a lab coat, and a geopolitical headache.
Highlights this week:
Climate data: Climate.gov rises again, Carbon Brief maps 1.8m climate papers, and Jamaican women turn adaptation into policy evidence
AI: rare diseases get fresh answers, Midjourney takes a medical pivot, and biology starts looking worryingly programmable
EVs: small electric cars return, sodium batteries arrive, and Africa’s EV transition gets supercharged by oil shocks
Clean energy: the Iran crisis pushes renewables, electrification, and energy security into the same sentence where they always belonged
Podcasts: Schneider Electric’s Philippe Rambach explains AI’s energy paradox, while Johan Oosthuizen unpacks the governance risk behind critical minerals
Some weeks whisper.
This one came back from the break carrying receipts, salt batteries, resurrected climate data, and yet another reminder that fossil fuel dependence is a terrible retirement plan.
Let’s get into it.
Climate
Climate.gov Was Deleted. Climate.us Just Brought It Back From The Dead
I’m delighted by this one, because it turns out climate data is harder to bury than inconvenient political talking points. After Climate.gov was gutted and redirected, former team members and volunteers have rebuilt it as Climate.us, restoring the archive, the tools, the teaching materials, and access to the Fifth National Climate Assessment. Quiet heroics. Loud implications.
Key Highlights
Climate.us has restored the material lost when Climate.gov was taken down, including a 15-year archive of climate explainers, expert blogs, maps, datasets, classroom resources, and visual climate indicators.
The new nonprofit is led by former Climate.gov staff, meaning the people who built one of the US government’s most accessible climate science resources are now protecting it outside the reach of political vandalism. Kudos to them.
Climate.gov had more than 15 million visits a year and over half a million social followers, making this less a niche science archive and more public infrastructure with graphs, citations, and a pulse.
Why This Matters: Accurate climate information is basic civic infrastructure: without trusted data, communities, businesses, educators, and policymakers are left making decisions with blindfolds on, which is apparently still some people’s preferred operating model.
Kismet: One of the Climate.us contributors, Robert Simmon, created the first global “Earth at Night” view and the Blue Marble image featured on the original iPhone, so yes, the people rescuing climate data also helped shape how millions first saw the planet in their pocket. 👉 Full story here
Carbon Brief Just Built A Climate Science Galaxy. It Has 1.8 Million Stars
And in a neat companion to the Climate.us revival story above, this is another reminder that good climate data has a nasty habit of surviving the people who would rather it didn’t. Project Cosmos is less “database” and more intellectual cartography for the climate age. Carbon Brief has mapped more than 1.8 million climate-related publications and over 40 million citation links, creating a vast, searchable picture of how climate science has grown, connected, argued, matured, and occasionally had to drag public policy behind it like a sulking toddler.
Key Highlights
Project Cosmos is Carbon Brief’s new database of climate change research, built over 18 months, covering more than 1.8 million unique publications and just over 40 million citation relationships.
The project starts from the full body of IPCC working group and special reports, pulling out more than 100,000 references, then expanding outward through second-order citations, IPCC-citing studies, and 22 climate-focused peer-reviewed journals.
Carbon Brief has used the database to create the “Cosmos 500”, ranking the most influential climate publications, authors, and institutions, while also opening the door to future analysis of research trends, blind spots, and shifting scientific focus.
Why This Matters: Good climate policy, corporate strategy, adaptation planning, risk modelling, and public literacy all depend on knowing what the science actually says, where it came from, and how strong the evidence base is, because vibes, unfortunately, remain a catastrophic substitute for knowledge.
Kismet: The oldest work in the Cosmos database dates back to 1483: a Latin print translation of Theophrastus’s Historia Plantarum, meaning a project about modern climate science somehow reaches all the way back to one of humanity’s earliest major attempts to systematically understand plants. 👉 Full story here
Jamaican Women Are Turning Climate Adaptation Into Something Policymakers Can’t Ignore
Climate adaptation can sound abstract, bureaucratic, and embalmed in committee language, but in Jamaica, women are making it visible, personal, and impossible to file away under “stakeholder engagement”. Through the Envisioning Resilience initiative, women are using photography to document water insecurity, extreme heat, hurricane damage, drought, and community recovery, then bringing those stories directly into adaptation policy conversations.
Key Highlights
Six women joined the first Jamaica phase of Envisioning Resilience in 2023, with seven more added in 2025, receiving photography and policy training to tell climate stories from their own communities.
The photo essays capture lived realities often missing from official planning, from families in Rose Town collecting water after decades without working pipes, to Kingston street vendors coping with rising midday heat, to Treasure Beach families rebuilding after Hurricane Melissa.
Jamaica’s government says the project is helping strengthen its National Adaptation Plan process by bridging technical planning and lived community experience, which is exactly the kind of sentence that sounds dull until you realise it means policymakers are finally being made to listen.
Why This Matters: Climate adaptation works best when it is shaped by the people already adapting every day, especially women, youth, and frontline communities who understand risk in ways that spreadsheets, consultants, and air-conditioned conference rooms routinely miss.
Kismet: After Hurricane Melissa blocked a main road in Treasure Beach, one resident opened their own backyard as a makeshift trail so neighbours could move through the community, a tiny act of improvised resilience that says more about adaptation than a thousand glossy strategy documents. 👉 Full story here
AI News
AI Is Coming For Healthcare’s Blind Spots, And Midjourney Just Took A Wild Left Turn
AI in healthcare is moving from shiny demo theatre into some genuinely useful territory: spotting rare disease clues doctors may never have time to revisit, accelerating the design of broad-spectrum defences against respiratory infections, and, in the strangest plot twist of the week, Midjourney announcing a medical scanner that sounds like an MRI wandered into a luxury spa and found religion. Caution required, obviously. But the direction of travel is fascinating: more data, earlier detection, faster research, and fewer humans left waiting years for answers because the system is too fragmented, underfunded, or allergic to common sense.
Key Highlights
OpenAI, Boston Children’s Hospital, and Harvard researchers used an AI-assisted workflow to reanalyse 376 previously unsolved rare disease cases, helping clinicians surface leads that resulted in 18 confirmed diagnoses after expert review and lab validation.
Midjourney has announced Midjourney Medical, including plans for a water-based ultrasound scanner designed to create a 3D body map in around 60 seconds, starting with body composition mapping while working through the long, joyless, and very necessary regulatory path.
Intercept, a new $500m philanthropic initiative backed by funders including Stripe, Anthropic, the OpenAI Foundation, and others, wants to radically reduce respiratory infections through broad-spectrum preventatives and cleaner indoor air technologies, because apparently “everyone getting sick every winter” was not, in fact, carved into stone tablets.
Why This Matters: Healthcare’s next leap may come less from one miracle drug and more from systems that notice patterns earlier, connect evidence faster, and make prevention normal rather than heroic, especially in areas where today’s incentives leave huge human and economic costs quietly piling up in the background.
Kismet: In the rare disease study, the AI surfaced a possible 22q11.2 deletion linked to DiGeorge syndrome even though that structural variant was not listed in the input data, a reminder that the most useful AI in medicine may not be the one that “knows the answer”, but the one that spots the thread humans should pull. 👉 Links Inline
AI Is Getting Cleaner Per Query. Now Comes The Awkward Bit
AI’s infrastructure giants are clearly working to shrink the energy and water intensity of compute, with NVIDIA pushing 45°C liquid cooling that can slash water use in the right climates, while OpenAI and Broadcom’s new Jalapeño inference chip promises substantially better performance per watt. The catch, because there is always a catch and physics likes to invoice eventually, is Jevons’ Paradox: if efficiency makes AI cheaper and easier to use, total demand may surge faster than per-unit impacts fall.
Key Highlights
NVIDIA’s next-generation Rubin AI infrastructure is designed for 100% liquid cooling, using coolant up to 45°C in closed loops that can enable dry-cooler operation and near-zero cooling water consumption in favourable climates.
OpenAI and Broadcom have unveiled Jalapeño, an LLM-optimised inference chip designed from scratch for modern AI workloads, with early testing showing substantially better performance per watt than current state-of-the-art accelerators.
The sustainability question is shifting from “Can AI be made more efficient?” to “Will efficiency gains reduce total footprint, or simply make AI so cheap and abundant that demand eats the savings for breakfast?”
Why This Matters: Efficiency is necessary, but not sufficient: if AI is going to scale responsibly, the industry needs cleaner chips, smarter cooling, renewable power, better siting, waste heat reuse, transparent reporting, and governance that recognises rebound effects before they become another infrastructure hangover.
Kismet: NVIDIA says its servers can run with coolant warmer than a hot tub, entering racks at 45°C and leaving at roughly 55°C, which means the future of AI cooling may be less “walk-in freezer” and more “industrial-grade underfloor heating with attitude”.
👉 Links Inline
Biology Just Got A Programming Language. That’s Either Brilliant Or Deeply Weird
A new Stanford and Arc Institute preprint introduces Proto, a high-level programming language for generative biology, letting researchers combine AI models, biological constraints, and optimisation tools into structured design programmes across DNA, RNA, proteins, ligands, and their interactions. The big idea is simple enough to be alarming: make biological design more like software design, except the code runs in cells and the debugging process has a suspicious amount of pipetting.
Key Highlights
Proto turns biological design into composable programmes built from four main primitives: sequences, constraints, generators, and optimisers, allowing researchers to specify what they want biology to do rather than hand-craft every molecular detail from scratch.
The team experimentally validated Proto-designed alternatives to traditional genetic engineering workflows, including cell-line-specific introns in human cells and synthetic promoter-repressor pairs in bacteria.
In one test, 32% of successfully assayed Proto-designed introns showed significant differential splicing in the intended direction, while 46% of tested ProtoRepressors significantly reduced gene expression against their target promoters.
Why This Matters: If this approach holds up beyond preprint-land, generative biology could become faster, more modular, and easier to apply in medicine, crop resilience, bio-manufacturing, enzyme design, and climate biotech.
Kismet: The paper includes a design for a 20-kilobase DNA sequence predicted to encode the Morse-code message “PROTO” through chromatin accessibility in the mouse genome, because apparently even DNA now wants to do branding.
👉 Full story here
Electromobility
Small EVs Are Back, And The SUV Blob Should Be Nervous
Europe’s carmakers are rediscovering the radical idea that city cars should fit in actual cities, with models like the Renault Twingo E-Tech, Renault 5, Cupra Raval, Citroën ë-C3, Smart #2, and BYD Dolphin Surf pushing back against the absurd arms race of electric SUVs. Smaller batteries, lower costs, tighter design, and more personality could make EVs cheaper, cleaner, and far more practical for the streets most people actually use.
Key Highlights
Renault’s new Twingo E-Tech is expected to start at €19,490 in France, using a 27.5kWh battery and offering 163 miles of range, which is plenty for daily urban use and refreshingly unlike hauling a cathedral to buy milk.
Europe’s smaller EV segment is heating up fast, with Renault, Citroën, Peugeot, Mini, Fiat, Cupra, Smart, BYD, and Leapmotor all chasing buyers who want electric mobility without the bulk, cost, and manufacturing footprint of oversized SUVs.
Battery cost reductions and smarter engineering are making small EVs commercially viable again, with Renault even cutting Twingo development to two years and reducing parts from roughly 1,500-2,000 to just 750.
Why This Matters: Small EVs solve more than tailpipe emissions: they reduce material use, battery size, energy demand, street congestion, manufacturing emissions, and cost, making electrification more democratic rather than another luxury lifestyle accessory with wheels.
Kismet: The new Smart Fortwo replacement is called the Smart #2, pronounced “hashtag two”, which may be the most 2026 thing imaginable: a tiny city car burdened with a name that sounds like a social media intern escaped into product strategy.
👉 Full story here
Salt Batteries Are Rolling Into EVs, And Lithium Just Got Company
Sodium-ion batteries are moving from “interesting chemistry” to “actual vehicles”, with CATL expecting 10,000-20,000 EVs to use its sodium batteries this year and Morgan Stanley arguing the market could grow sharply through the next decade. The big promise is not glamorous: cheaper materials, better cold-weather performance, fewer critical mineral headaches, and big implications for grid storage, commercial fleets, and affordable EVs.
Key Highlights
CATL says its sodium-ion batteries can operate in temperatures as low as -20°C or -30°C, making them especially interesting for cold-weather EVs, fleet vehicles, and storage systems where lithium iron phosphate batteries can struggle.
Morgan Stanley sees sodium-ion battery deployment rising from around 2% market share in 2027 to 20% by 2030 and 37% by 2035, while estimating a potential $800bn capital formation wave by 2035. Financial analysts have discovered salt. Alert the philosophers.
Sodium-ion batteries reduce reliance on lithium, copper, cobalt, and graphite, but they are not magic: supply chains for hard carbon anodes, Prussian blue-type cathodes, electrolytes, and manufacturing capacity still need to scale.
Why This Matters: Sodium-ion will not replace lithium everywhere, but it could reshape the economics of low-cost EVs, cold-climate fleets, renewable energy storage, and energy security by moving part of the battery supply chain towards cheaper, more abundant materials.
Kismet: The wonderfully ridiculous phrase “salt is the new oil” may actually contain a serious point: sodium is so widely available that battery competition could shift from mineral access to manufacturing excellence, supply-chain scale, and system design.
👉 Full story here
Hormuz Shock Turns Africa’s EV Shift From Sensible To Urgent
Oil shocks have a funny way of clarifying the spreadsheet, and across Africa the case for electric motorbikes, buses, and trucks has suddenly become a lot harder to dismiss. With petrol and diesel costs rising after the Middle East energy crisis, e-mobility start-ups are seeing demand surge, financing queues stretch, and Chinese-backed supply chains move fast into a market where two wheels may matter more than four.
Key Highlights
Petrol costs for African moto-taxis have jumped from about $4.20 to $5.10 a day since the war, while electric motorbikes can cover the same distance for roughly $2.30, which is the kind of maths even a fossil lobbyist would struggle to smother with a consultancy report.
EV infrastructure start-ups have raised around $300m since last month, including Spiro, which secured $215m recently and is adding another $55m, while selling 10,000 electric motorbikes on the continent in a single month.
Policy and fleet economics are lining up fast: Ethiopia has banned combustion vehicle imports and now has more than 115,000 EVs, Rwanda has banned new combustion moto-taxi registrations in Kigali, and electric truck operators in East Africa see five-year costs 20-40% below diesel once fuel savings are counted.
Why This Matters: Africa’s EV transition may not look like Europe’s car-led pathway at all; it could be driven by motorbikes, buses, battery swapping, trade corridors, fuel import pressure, and governments tired of watching oil shocks mug their economies in broad daylight.
Kismet: Spiro is targeting 1m electric motorbike sales in 2027 and says it wants 90% of materials sourced locally across African manufacturing hubs, which hints at a far bigger prize than cleaner transport: African resources transformed in Africa, instead of exported cheaply and bought back expensively, because apparently that lesson only took a few centuries. 👉 Full story here
Clean Energy
Trump’s Iran War May Have Accidentally Become A Clean Energy Accelerator
Energy shocks have a habit of doing what climate diplomacy struggles to: making fossil fuel dependence look expensive, fragile, and politically stupid in real time. China’s climate envoy, the IEA’s Fatih Birol, and the Private Infrastructure Development Group are all pointing in the same direction: the Iran crisis is pushing governments, investors, and emerging markets to treat renewables, electrification, storage, and energy security as one joined-up agenda, because oil dependency has once again wandered onto the stage wearing clown shoes and carrying a live grenade.
Key Highlights
China’s climate envoy Liu Zhenmin said oil and gas shortages caused by the Iran crisis should push countries to speed up their energy transitions, warning that any short-term shift back to coal is just that: short-term panic with soot.
IEA chief Fatih Birol said governments are reviewing energy strategies, policies, partnerships, and technology choices, and expects the crisis to give an extra boost to global electrification.
Emerging market green bond sales in dollars and euros have jumped 28% to $31bn so far this year, as countries and investors look to fund renewables and energy-security projects in response to oil price volatility.
Why This Matters: Clean energy is no longer just a climate play; it is becoming a national security, inflation control, industrial strategy, and resilience play, which is what happens when fossil fuels keep mistaking geopolitical chaos for a business model.
Kismet: China remains the world’s largest oil and gas importer, yet its climate envoy argued it weathered the Hormuz shock better than many Asian peers because its energy system is already more diversified with wind, solar, EVs, storage, and a vast grid buildout, proving that the best resilience is built long before the crisis arrives.
👉 Links Inline
Electrification Could Put £1,200 A Year Back In UK Households’ Pockets
Following the previous story on how the Iran crisis is accelerating clean energy interest globally, the UK’s Climate Change Committee has landed with the home-front version: faster electrification is the clearest route to lower bills, stronger energy security, and less exposure to fossil fuel price shocks. Its latest progress report says households combining an EV, heat pump, flexible electricity tariff, and solar could save around £1,200 a year compared with sticking to petrol cars and gas boilers, which is what policy wonks call “consumer benefit” and normal people call “money”.
Key Highlights
The CCC says electrification is now central to cutting UK emissions, especially across transport, buildings, and industry, where progress has lagged behind the huge gains already made in cleaning up electricity.
By 2030, faster electrification could save the UK up to 80m barrels of oil and 1.5bn therms of gas each year, worth almost £8bn at current prices. That is not ideology. That is avoided fuel dependency with a receipt.
The UK still has serious policy gaps: 17% of the emissions cuts needed for its 2030 climate target are not covered by current government plans, while heat pumps, electric vans, heavy industry, aviation fuels, and engineered removals all remain worryingly undercooked.
Why This Matters: Electrification is where climate policy, household economics, energy security, and industrial strategy finally meet, and weakening EV or clean heating policies now would leave people paying more for imported fossil fuels while pretending delay is prudence.
Kismet: Carbon Brief’s analysis shows the word “electrification” appears far more often in this year’s CCC progress report than in previous editions, which is a tiny linguistic tell with big implications: the centre of gravity in climate policy is shifting from abstract emissions targets to plugging actual machines into cleaner power. 👉 Full story here
Podcast Episodes
Climate Confident:
AI’s Energy Paradox: The Same Tech Straining The Grid Could Help Save It
AI’s electricity appetite is real, but this week’s Climate Confident conversation with Philippe Rambach, Chief AI Officer at Schneider Electric, adds a badly needed dose of nuance: the right AI, applied to buildings, grids, factories, and demand management, can cut waste, shift energy use away from dirty peaks, and help electrification scale without everything catching fire metaphorically, financially, or actually. Philippe’s central point is blunt and useful: don’t start with the model, start with the business or climate problem, then choose the smallest, most effective AI tool for the job.
Key Highlights
Schneider says AI-powered room controllers can learn how individual rooms behave and cut heating and cooling energy use by around 20%, which matters because buildings are responsible for a huge 40% chunk of global emissions.
AI can help shift demand away from peak grid periods by forecasting building consumption, solar output, battery use, and electricity prices, turning energy flexibility into both a climate tool and a cost-saving machine.
Philippe argues most AI pilots fail because companies treat AI as shiny innovation theatre rather than transformation work: start with value, build cross-functional teams, and take responsibility all the way to production. Radical stuff, apparently.
Why This Matters: As AI drives new electricity demand, the climate question becomes less “Is AI good or bad?” and more “Are we using the right AI, in the right places, to reduce real-world energy waste faster than compute demand grows?”
Kismet: Philippe said Schneider’s estimate for some energy-saving AI use cases is roughly a 1:500 ratio between the carbon emitted to run the model and the energy saved, which is the kind of number that makes the “AI is only a climate villain” narrative wobble awkwardly in public. 🎧 Listen to the full episode
Resilient Supply Chain:
Critical Minerals Won’t Save The Climate If Their Supply Chains Are A Mess
This week on the Resilient Supply Chain podcast, responsible sourcing specialist Johan Oosthuizen takes us beneath the shiny battery-and-EV narrative and into the messy, human, operational reality of critical mineral supply chains. The core argument is sharp: reporting is not enough, self-declaration is structurally weak, and the real resilience test lies three layers down in suppliers, contractors, communities, audit data, governance systems, and the uncomfortable gap between boardroom ambition and mine-site reality.
Key Highlights
Johan explains why critical mineral supply chains are ecosystems, not simple supplier lists: a large mine may employ 1,000-2,000 people directly, but can require around 10,000 people in its first-tier support network alone.
Self-reporting has obvious limits because companies tend to present “the best version” of themselves, while the real risk often sits buried in weaker contractors, hidden community tensions, poor document control, and unaudited management systems.
Audit data can become strategic, not bureaucratic, if procurement teams use it to shape sourcing decisions, build supplier capability, strengthen local economies, reduce operational disruption, and spot governance risk before it becomes a very expensive surprise wearing a hard hat.
Why This Matters: The energy transition depends on minerals, but minerals depend on trust, traceability, community consent, and governance quality; without those, clean tech supply chains risk swapping fossil fuel dependency for a new flavour of fragility.
Kismet: Johan’s line that “governance risk is bigger than geology risk” is the quietly explosive idea here: the minerals may be in the ground, but whether they can be responsibly, reliably, and profitably extracted often depends more on people, systems, and social licence than on the rocks themselves. 🎧 Listen to the full episode
Featured Chart(s)
Kudos to the UK for it’s success in reducing the carbon intensity of its electricity generation. Still a long way to go, but great job, so far.
Utility scale battery storage rollout in Australia is flattening the curve on electricity prices
Not all clean tech comes from China - a surprisingly large amount comes from the EU.
The % of new car sales that are electric. A few short years ago, these would all have been single digit percentages (with the possible exception of Norway!)
And it is not just cars that are going electric. Increasingly it is HGVs
Misc stuff
In 2014 French TV broadcast this weather forecast of what a typical august forecast would look like in 2050 (see top panel).
We’re only in 2026 and already we’ve blown through that prediction (bottom panel).
Beautiful, beautiful reply. Chef’s kiss.
And news reports have it the police have finally abandoned their years long search for the Benefits of Brexit with hopes fading of ever finding any.
Engage
If you made it this far, very well done! If you liked this newsletter, or learned something new, feel free to share this newsletter with family and friends. Encourage folks to sign up for it.
Finally, since being impacted by the tech layoffs, I'm currently in the market for a new role. If you know someone who could benefit from my tech savvy, sustainability, and strong social media expertise, I'd be really grateful for a referral.
If you have any comments or suggestions for how I can improve this newsletter, don’t hesitate to let me know. Thanks.
*** Be aware that any typos you find in this newsletter are tests to see who is paying attention! ***
And Finally
The world’s first man to lose so much money, he became a multi-billionaire - how sick is that?
Talk to your AI tools the way you'd talk to a colleague.
You don't send a colleague a three-word brief. You explain the context, the constraints, what you've already tried. But typing all that into ChatGPT takes forever — so you don't.
Wispr Flow lets you speak your prompts instead. Talk through your thinking naturally and get clean, paste-ready text. No filler words. No cleanup. Just detailed prompts that actually get you useful answers on the first try.
Millions of users worldwide. Works system-wide on Mac, Windows, and iPhone.