In partnership with
Hey everyone, it’s Thursday, and FutureProof is back.
This week, FedEx discovers that spending $300 million on electric trucks could save $800 million in diesel, India overtakes the US on EV adoption, Google puts AI to work on disease, and solar starts eating into the diesel-generator economy.
Highlights this week:
The Stakes: Faster warming, worsening planetary indicators, and insurers warning of growing financial risk.
Climate: AMOC risk becomes political drama, while carbon accounting turns into a fight over responsibility.
AI: Google applies AI to disease, while Anthropic points Claude at cyber defence.
EVs: FedEx orders 2,000 electric trucks, India passes the US, and Ember maps the global EV shift.
Clean energy: Solar challenges diesel in Iraq, while wind and solar slash the emissions we would otherwise be producing.
From me: Why AI is exposing enterprise data problems, plus Sagar Paul on making data fit for machine-speed decisions.
Plenty happening. Let’s get into it.
Interestingly, the common thread this week is that technologies once sold mainly on environmental grounds are increasingly winning on harder measures: cost, resilience, energy security, productivity and risk.
Kismet: FedEx electrifies because diesel is expensive; Iraqi villages adopt solar because diesel-backed electricity is unreliable. Different markets, same direction.
The Stakes
Climate Change Is Accelerating, and Insurers Are Starting to Back Away
This week’s climate signals are grimly consistent: observational data suggest global warming has accelerated sharply, planetary vital signs are hitting record extremes, and insurers are beginning to retreat from risks they increasingly struggle to price. I think the most important shift here is that climate change is moving from an environmental warning into something far more immediate - a threat to property values, mortgages, public finances and the basic machinery of risk itself.
Researchers analysing five major temperature datasets found warming averaged about 0.40°C per decade from 2011–2025, roughly double the longer-term rate; if that pace persists, the 20-year Paris definition of 1.5°C warming could be crossed around 2028.
The 2026 State of the Climate report finds 7 of 16 planetary vital signs at record extremes, with ocean heat, sea level, greenhouse gases and ice loss among the indicators moving in the wrong direction — several of them increasingly quickly.
European insurance insiders warn that withdrawal from climate-exposed risks has already begun and could accelerate. Only about a quarter of Europe’s natural-catastrophe losses are currently insured, raising an awkward question: if insurers stop carrying the risk, who does? Increasingly, homeowners and taxpayers.
Why This Matters: Climate risk stops being an abstract future problem very quickly once insurers, banks and mortgage markets start repricing the physical world beneath them.
Kismet: Some insurance insiders say catastrophe modellers and actuaries are being pressured to “manage down” future risk estimates to keep deals moving, which is an extraordinary echo of what happens when financial systems discover inconvenient numbers before everybody else does. 👉 Links Inline
Climate
Climate Policy Gets the Political Thriller Treatment
A former Climate Confident guest, Amos Wittenberg, sent me this rather brilliant experiment: When The Sea Slows, a near-future political drama that asks what happens when climate tipping points stop being abstract science and start landing on ministers’ desks, budgets, grids, farms and elections. It is fiction, but deliberately grounded in current climate science - think The Thick of It meets AMOC collapse, with fewer jokes and rather more existential dread.
The story follows a new Department of Climate Resilience from 2029 onwards as an accelerating AMOC slowdown collides with energy security, flooding, insurance retreat, migration, food production and public finances.
Instead of treating climate change primarily as a mitigation argument, it explores the much messier politics of resilience: who pays, what gets protected, whether gas plants stay open, how much risk governments should plan for, and what happens when every option annoys somebody.
You can read it chapter-by-chapter or listen to the narrated version, with technical explainers threaded throughout so the fictional politics remain anchored to real science rather than Hollywood climatology.
Why This Matters: Climate communication often struggles because graphs explain risk but rarely make us feel how that risk cascades through institutions; this project turns those system interactions into politics, trade-offs and human decisions.
Kismet: The story eventually lets readers choose between two political outcomes — meaning climate resilience becomes, quite literally, an election you get to vote in.
👉 Full story here
Carbon Accounting Is Having a Quiet Civil War
Carbon accounting comes up constantly on both my Climate Confident and Resilient Supply Chain podcasts, and with good reason: the rules decide who “owns” each tonne of emissions, which in turn shapes corporate targets, procurement, regulation and investment. Right now, that rulebook is being fought over - politely, technically, and with enormous consequences.
The Greenhouse Gas Protocol remains the de facto global standard, splitting emissions into Scopes 1, 2 and 3 - but critics argue the system can double-count emissions and lean too heavily on estimates.
One proposed alternative, e-ledgers, would assign every tonne of emissions once and transfer that carbon record through the value chain with the product. Cleaner data, stronger procurement signals, but only if everybody participates.
The biggest philosophical fight is Scope 3: should an oil company remain responsible for the emissions from burning its products, or should those emissions move down the ledger to the end user? That is not bookkeeping trivia; it changes corporate accountability completely.
Why This Matters: Change the accounting rules and you change what companies optimise for - so carbon accounting can quietly redirect billions in procurement, energy investment and decarbonisation spending.
Kismet: Some of the strongest corporate backing for e-ledgers comes from ExxonMobil and BlackRock, which is a useful reminder that whenever someone proposes a “cleaner” accounting system, it is worth checking whose emissions become someone else’s problem. 👉 Full story here
AI News
Google Is Turning AI Into a Disease Detective
Google is pushing AI well beyond chatbots and into something rather more consequential: helping scientists understand disease, design treatments, diagnose patients faster and even forecast outbreaks. The breadth is striking - from proteins and genomes to mammograms and Ebola - and this is the kind of AI progress I find genuinely exciting.
AlphaFold helped researchers in Barcelona virtually screen 30,000 compounds against Chagas disease, identifying two existing FDA-approved drugs that could potentially be repurposed.
Google tools used with Stanford enabled whole-genome diagnoses in under eight hours in time-critical genetic cases, while DeepSomatic found previously missed mutations in childhood leukaemia samples.
In mammography studies covering 125,000 women, Google says its system caught 25% of historically missed interval cancers while potentially cutting specialist reading workloads by up to 40%; its flu forecasting model also ranked first in the CDC’s 2025–26 challenge.
Why This Matters: The most important AI story may turn out not to be who builds the cleverest chatbot, but who can compress years of biological discovery and clinical analysis into hours.
Kismet: Google is now using AI to design entirely new enzymes from scratch, including proteins that can manufacture drug components and others capable of breaking down plastics under extreme conditions. 👉 Full story here
Anthropic Wants Claude Defending Power Grids and Open-Source Software
Anthropic has launched its Cyber Mission, and this is a much more interesting AI story than another model leaderboard: Claude is being pushed into defending the operational technology behind power grids, water systems, factories and transport networks, while also scanning the open-source code that modern software quietly depends on. Cybersecurity is becoming one of the clearest arenas where AI could genuinely tilt the balance between attackers and defenders.
The new Critical Infrastructure Defense Program brings frontier Claude models, Anthropic engineers and threat research together with firms including Accenture, CrowdStrike, Dragos, Hitachi, Nozomi Networks, Palo Alto Networks and Rockwell Automation.
Anthropic is also launching OSS Scanner, giving open-source projects free recurring scans for vulnerabilities, complete with exploit proofs-of-concept and suggested fixes; Anthropic expects a true-positive rate above 90%.
The company’s bet is that within two years AI could favour defence, helping catch bugs before software ships, generate safer code and actively protect systems — although right now the economics still often favour attackers because finding vulnerabilities has become much cheaper than fixing them.
Why This Matters: Critical infrastructure and open-source software are both foundational and chronically under-resourced, so putting frontier AI directly into their defence could be one of the highest-value applications of the technology.
Kismet: Anthropic notes that some operational-technology vulnerabilities can remain unfixed for years - occasionally decades, because patching a live industrial system can itself be more dangerous than leaving the flaw in place. 👉 Full story here
Electromobility
FedEx Orders 2,000 Electric Trucks, and Expects $800m in Fuel Savings
FedEx has placed a $300m-plus order for 2,000 electric box trucks from US manufacturer Harbinger, one of the largest binding orders for medium- or heavy-duty electric trucks yet. What really jumps out at me is the economics: this isn’t corporate virtue signalling - FedEx expects the vehicles to save vastly more in diesel than they cost to buy.
Harbinger estimates each Class 5 and 6 truck will cut FedEx’s fuel bill by about $20,000 a year, with the vehicles built for a 20-year operating life.
Across 2,000 trucks, that adds up to roughly $800m in lifetime fuel savings against a purchase price north of $300m - before counting potential maintenance savings.
The switch is also expected to avoid more than 1.7 million tonnes of CO₂ over the vehicles’ operating lives, while helping FedEx work towards fully electrifying its operating fleet by 2040.
Why This Matters: Commercial EV adoption gets much harder to dismiss as environmental window-dressing once fleet operators start buying thousands of trucks because the spreadsheet says diesel is the expensive option.
Kismet: Medium-duty delivery trucks are almost comically well suited to electrification - predictable routes, comparatively modest daily mileage, lots of room for batteries, and long overnight stops back at depots where charging can happen while everybody sleeps. 👉 Full story here
While the US Hits Reverse, Emerging Markets Are Flooring the EV Accelerator
The contrast is getting hard to ignore: as US policy retreats and its BEV share falls to 6%, countries including India and the Philippines are accelerating electrification for reasons that are as much about economics and energy security as climate. The centre of gravity in the EV transition is shifting, and it is not waiting for Washington.
India has now beaten the US on fully electric car share for six consecutive months, while 49 of the 56 countries tracked by Ember now have a higher BEV share than the US.
In the Philippines, EV registrations jumped 130% in the first eight months of the year and electrified vehicles now account for nearly a quarter of passenger-car sales, up from 12% in 2025 and 4% in 2024.
The economics are doing a lot of the persuading: Philippine petrol prices surged 68.3% after the Iran conflict, while EV fleets across oil-importing nations displaced about 200 million barrels of crude between March and August, saving roughly $20bn.
Why This Matters: EV adoption is increasingly being driven by something even more politically durable than climate policy, the desire to stop sending money abroad for volatile imported oil.
Kismet: India, Brazil, Malaysia and New Zealand are now among the countries beating the US on BEV market share, which is quite a reversal for a technology transition once framed almost entirely around wealthy Western markets. 👉 Links Inline
Ember Just Built the EV Tracker I Didn’t Know I Needed
Ember has launched an open, monthly EV Sales Data Tracker covering 56 countries and more than 90% of global EV sales - and for a data nerd like me, this is catnip. You can compare markets, split BEVs from plug-in hybrids, track registrations over time, and even see how much oil the global EV fleet is already displacing.
So far in 2026, 25.2% of new cars registered globally have been electric when BEVs and plug-in hybrids are combined.
Ember estimates there are now 88.5 million EVs on the road, displacing around 1.4 million barrels of oil every single day, roughly $140 million worth at $100/barrel.
The country rankings are wonderfully revealing: Norway leads at 97.8% BEV share, but Nepal is already at 74.6% and Vietnam at 46.9%, another reminder that this transition is emphatically not confined to rich Western markets.
Why This Matters: Good data kills stale narratives, and a transparent monthly tracker makes it much harder to claim EV adoption is stalling when the global numbers are saying the opposite.
Kismet: Ember says the global electric car fleet is now doubling every 2-3 years - fast enough that last year’s EV talking points can become obsolete before you’ve finished repeating them. 👉 Full story here
Clean Energy
Solar Is Coming for the Diesel Generator
Iraq offers another glimpse of a much bigger shift I’m watching across places from Pakistan to Africa: where electricity grids are weak, expensive or unreliable, cheap solar plus batteries can leapfrog the old diesel-generator economy entirely. In Iraqi Kurdistan, that is already happening village by village - less glamorous than a giant solar farm, perhaps, but arguably more transformative.
The village of Kulak powers 32 homes with nearly 200 solar panels, spread across houses, schools, mosques and even a hospital, eliminating the need for diesel generators.
Iraq is targeting 12 GW of solar by 2030, including a 1 GW TotalEnergies project in Basra and a planned 300 MW solar park in Karbala.
The economics get especially compelling in remote areas: extending transmission lines across mountains to a handful of homes can be ruinously expensive, while modular solar-and-battery microgrids can grow with demand.
Why This Matters: In many developing markets, solar is not merely replacing fossil electricity - it is bypassing expensive, unreliable infrastructure and giving communities a cheaper route to dependable power.
Kismet: In places dominated by generator backup, households can effectively pay twice for electricity - once for the grid and again for diesel when it fails, so solar’s advantage is not just lower emissions, but escaping an entire parallel energy economy.
👉 Full story here
Without Wind and Solar, Power-Sector Emissions Would Be 28% Higher
One argument I keep hearing is that renewables have merely been piled on top of fossil fuels rather than replacing them. Zeke Hausfather ran the counterfactual, and the answer is pretty stark: freeze wind and solar at 2005 levels and today’s power sector would be emitting vastly more CO₂ - clean energy has been meeting new demand and displacing fossil generation at the same time.
Without wind and solar growth since 2005, global power-sector emissions in 2025 would have been about 16.9 billion tonnes of CO₂ instead of 13.2 billion, roughly 28% higher.
Between 2006 and 2025, wind and solar avoided an estimated 23 billion tonnes of CO₂, with around 60% of those avoided emissions occurring in just the past five years.
The displacement is accelerating: since 2015, wind and solar supplied 57% of global electricity-demand growth, and in 2025 alone their additional generation met 99% of the increase in demand.
Why This Matters: Looking only at whether fossil generation has fallen misses the counterfactual completely, the relevant question is how much more coal and gas we would be burning had renewables never scaled.
Kismet: The climate benefit looks smaller in today’s temperature record partly because extra coal pollution would have temporarily cooled the planet through sulphate aerosols; that pollution disappears quickly, while the CO₂ warming it masks hangs around for centuries. 👉 Full story here
Latest blog post
AI Isn’t Breaking Enterprise Data - It’s Revealing How Broken It Already Was
In my latest post, I argue that the biggest barrier to autonomous enterprise AI may not be model capability at all, but semantic debt: years of undocumented assumptions, conflicting definitions and human workarounds buried inside corporate data. People have quietly been acting as the middleware; AI removes that safety net.
A field like “delivery date” can be perfectly clean, complete and formatted - yet still mean different things to logistics, finance and customer service.
As companies move from AI assistance to AI autonomy, those ambiguities become dangerous because machines can act on them at scale, repeatedly and very quickly.
The competitive advantage may therefore shift away from access to the latest model and towards something less glamorous but far more durable: better context, lineage, governance and shared definitions.
Why This Matters: Before leaders ask whether AI is smart enough to make operational decisions, they need to ask whether their organisation can explain, precisely and consistently, what its own data actually means.
Kismet: For decades, experienced employees have been quietly correcting enterprise systems in their heads; AI is forcing companies to document institutional knowledge they never realised was infrastructure. 👉 Full story here
Resilient Supply Chain:
AI Won’t Fix Your Supply Chain Data, It’ll Scale the Mess
In this week’s Resilient Supply Chain, I spoke with Sagar Paul of The Modern Data Company about a problem I suspect many companies are about to discover the hard way: enterprise systems often work because humans quietly correct, reinterpret and constrain the data as they go. Give an AI agent the same information without that hidden judgement layer, and ambiguity can suddenly become automation at machine speed.
Sagar identifies four recurring failure points: format inconsistencies, data quality, temporal misalignment and semantic ambiguity, with that last one especially dangerous once machines start making decisions.
“Real-time” data is not automatically useful; without context, relevance and decision rules, it can become digital exhaust - more alerts, more infrastructure cost, and very little additional value.
The strategic shift is from collecting more data to productising existing data: adding provenance, lineage, freshness, permissions and business meaning so both people and machines know what they are actually looking at.
Why This Matters: AI autonomy raises the cost of ambiguity, because weak data that humans once corrected informally can now drive thousands of wrong decisions before anyone notices.
Kismet: Sagar’s best analogy was milk: raw data is the bucket straight from the cow; useful data is the labelled carton telling you where it came from, how fresh it is, what is in it, and whether you should actually drink it. 🎧 Listen to the full episode
Don’t forget to follow the podcasts in your podcast app of choice to ensure you don’t miss any episodes.
Featured Chart(s)
South Australia is really showing the rest of the world how to successfully deploy renewables
Sales of EVs are bringing the global car sales numbers up, even as IC vehicle sales tumble
China’s EV sales hit 67.1% of new vehicle sales in September - onwards and upwards!
Mind-blowing: 40% of global shipping is just moving fossil fuels around!!!
Misc
Great to see gun ownership remaining low in Ireland (where I’m from), and Andalusia (where I live)
Mandatory Trump humour
Cartoonists didn’t think much of Trump’s betrayal of Ukraine, with his deal to buy diesel from war criminal Putin’s Russia
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
Enough of this woke nonsense!!!
The AI notetaker that gets the hard words right
Your AI is only as good as what you feed it. Feed it a transcript with your product names, acronyms, and numbers spelled wrong, and you spend the afternoon hand-fixing the output.
Wispr Flow Notetaker uses your dictionary and calendar, so product names, acronyms, numbers, and uncommon names come out spelled right. It pops up when your meeting starts and captures Zoom, Google Meet, Teams, and Slack huddles with one click, no bot joining the call.
Then your meetings show up inside Claude or ChatGPT through the built-in connector. No copy paste. Try Notetaker free on Mac and Windows.