The scientific advances ready to change the world: the Top 10 Emerging Technologies 2026

ELI5 / TLDR

Every year the World Economic Forum asks a jury of experts to pick ten technologies that have stopped being lab curiosities and are about to start changing real life in the next five years. This year’s ten run from a paint that cools buildings by beaming heat into space, to cancer vaccines built for one specific patient’s tumour, to encryption designed to survive computers that don’t exist yet. The pattern underneath them: things are getting more personal (treatments fitted to one body, not an average) and more local (making lithium, protein, or electricity near where it’s used instead of shipping it across the planet). Most are real and moving, but “in the next five years” is doing a lot of work in that sentence.

The Full Story

The format is simple. The WEF has published this list for fourteen years, long enough to know its own batting average: some picks scaled, some took far longer than expected, some stalled. The ranking here is not a ranking at all, just ten items in no order, chosen for two things together: likely to scale, and consequential if they do. Host Robin Pomeroy reads out each name; researcher Kimmie Bettinger explains it. What follows is the ten, grouped loosely by what they actually do.

Turning the grid into a two-way street — “everything to grid” energy

Today most things connected to the electricity grid only pull power out. The idea here is to let them push power back in too. Your electric car, a building, a warehouse battery, rooftop solar — each becomes a small store of energy that can absorb a surplus when the wind is blowing hard and release it when demand spikes.

Think of it like a city where every house has a small water tank. Instead of one giant reservoir straining to meet the 6pm rush, thousands of tanks quietly balance each other. The thing making this newly possible is a mix of better battery chemistries, smarter control software, and two-way flow of electricity.

He calls this the technology that’s really the foundation for a completely reimagined grid.

The timing matters because grids are already stressed by heat waves, and now data centres for AI are arriving as a giant new draw. The honest drawback, which the hosts flag: this rewards people who own a house and a car. If you rent, your landlord may pocket the benefit.

Getting lithium out of the ground faster — direct lithium extraction

Lithium sits in most batteries, so the energy transition runs on it. The current way to get it is almost medieval: pump lithium-rich brine into enormous turquoise ponds in a high-altitude desert (the Atacama in Chile is the famous one) and wait up to two years for the sun to evaporate the water.

Direct lithium extraction pulls the lithium straight out of the brine in hours, uses less water, and — crucially — works in a small, modular plant rather than needing one specific kind of desert.

They’re these huge turquoise ponds that are perfectly geometric in shape, and they stretch for kilometers… we wait, sometimes for up to 2 years, for the sun to evaporate the water.

Right now refining is concentrated — China around 62%, Chile 13%, Argentina 11% — and a single batch of lithium can cross several continents (dug up in Chile, refined in China, packed into a battery in Detroit). Modular extraction opens the door to doing the digging and refining in the same place. The catch: lithium’s price has fallen more than 80% since 2022, and nobody rushes to build new plants for something getting cheaper.

Paint that cools by beaming heat to space — passive radiative cooling

Start with the trap the hosts call the cooling paradox: air conditioning uses electricity, electricity makes heat, the heat warms the surroundings, the surroundings demand more cooling. A loop that feeds itself.

Passive radiative cooling breaks it. Picture a special paint or coating that throws heat off a surface at exactly the wavelength that slips through the atmosphere and escapes into deep space — no electricity needed. Painted on a roof, it can pull a building below the temperature of the air around it.

It comes as paint, roof tiles, window films, even heavy fabrics, and can be retrofitted onto existing buildings. The pitch is strongest in poorer, hotter countries where AC is unaffordable; reported energy savings reach up to 40%. Urban “heat islands” — cities that run 0.5 to 4°C hotter than the countryside around them — are the obvious target. One UK firm even coats power cables so they stay cool enough to carry more current.

Destroying “forever chemicals” — PFAS destruction

PFAS are the chemicals that make pans non-stick and firefighting foam work. They are built around the carbon-fluorine bond, one of the strongest in all of chemistry — which is exactly why they never break down. They’ve turned up in Arctic rainwater and in the blood of nearly everyone tested.

Until now the best we could do was contain them. The new thing is destruction: a handful of methods that actually snap that bond. One runs contaminated water across specialised electrodes, using electric current to tear the molecules apart.

We’ve found PFAS in places like the Arctic. It’s in rainwater on every continent in the world and even actually in the bloodstream of almost everyone we’ve tested for it.

It’s early, mostly government-funded, and driven by two motives: clean drinking water, and land that was written off as contaminated suddenly becoming usable again.

Brewing protein in a tank — precision fermentation

Give a simple microbe, like yeast, a new set of genetic instructions, and it will brew a specific molecule you want — a protein, a fat — that is chemically identical to the version from an animal. You can make whey protein with no cow involved.

The argument is scale: feeding a projected 9 billion people by 2050 bumps against hard limits on land, water and emissions, and the current way of making food doesn’t stretch that far. The same trick also produces cosmetic ingredients and chemicals we used to pull from fossil fuels. It’s already trickling to market — one firm sells a fermentation-made egg substitute through Walmart.

Bettinger’s own caveat, and it’s a thoughtful one: her husband is a cattle rancher, and a technology that makes protein in a city tank quietly pulls income away from farming families who’ve done it for generations — a bigger deal still in countries where most people work the land.

A pause on scaling

Halfway through, the conversation steps back to ask why good technology sometimes never arrives. The answer, from fourteen years of hindsight, is rarely the science. It’s three other things. First, whether the supporting pieces exist — mRNA was spotted in 2014 but went nowhere until a way to deliver it into cells arrived around 2018. Second, whether someone brave enough will use the rough, expensive early version — often a desperate doctor or a government with a crisis. Third, money and nerve. Useful framing for reading any “emerging tech” list with a cold eye.

A courier already inside your body — exosome drug delivery

We now have drugs that can switch off a mutation or edit a gene. The problem is delivery: a drug is useless if it can’t reach the right spot, and the body attacks anything foreign.

The trick is to hijack a courier the body already trusts. Cells naturally send tiny membrane-wrapped packets — exosomes — to message each other. Load one with a drug, address it, and the body waves it through. Because it’s recognised as “self,” it can even cross the blood-brain barrier, the wall that has long kept drugs out of the brain.

Because the exosome is a courier that the body recognizes, it can actually cross the blood-brain barrier… diseases like Alzheimer’s or Parkinson’s that have been really hard to treat for decades are now on the table.

Early trials span Parkinson’s, Alzheimer’s, cancer (including some success against pancreatic cancer), and long COVID.

A vaccine built for one tumour — personalized mRNA cancer vaccines

Chemotherapy, in the hosts’ phrase, is cracking a nut with a sledgehammer — controlled damage, not a precise strike. The alternative here: biopsy a patient’s tumour, read the mutations unique to it, and brew a vaccine aimed at that exact cancer in that exact person, in weeks.

Cracking a nut with a sledgehammer… You might open the nut, but it does create a lot of damage.

This inverts the entire economics of pharma, which is built on one blockbuster drug sold to millions. The cost is the obvious wall — early treatments above $100,000 per patient — survivable in a rich country, out of reach for billions. Possible fixes float by: hybrid off-the-shelf-plus-personalised vaccines, AI to speed discovery, and treatment labs sited down the hall from the patient.

Modelling molecules atom by atom — quantum simulation for drug discovery

Ordinary computers can’t capture how a molecule really behaves, so they approximate by stripping out complexity. A quantum computer can model the molecule directly, watching a candidate drug fold and lock onto its target atom by atom — before anyone makes it in a lab.

Why this matters: nine of ten drug candidates that reach clinical trials fail, partly because the cruder modelling oversold them. Better simulation upstream could shift that ratio, which changes the economics — and could make rare diseases (individually small markets, collectively hundreds of millions of patients) worth pursuing.

AI that learns by bumping into things — world models

Bettinger’s nine-month-old can’t define gravity, but he drops his spoon off the high chair, feels the pull, hears the thud, and builds a working model of the world in his head. Today’s most powerful AI mostly learns the opposite way — from text describing the world, not from the world itself.

He could never tell you what gravity is, but he already has a little model in his head around how the world works.

World models learn like the baby: from experience and physical interaction, fed by video, depth sensors, pressure readings, motion capture. The payoff is robots that cope with messy, ambiguous situations, climate models that grasp how a storm actually moves, factories that learn as they run. Self-driving cars on the streets of San Francisco are an early hint of this sensory learning at scale.

Encryption that survives quantum computers — lattice-based cryptography

Today’s encryption could one day be cracked by a quantum computer powerful enough — which doesn’t exist yet, but criminals already act on a “harvest now, decrypt later” basis: steal the scrambled data now, wait years, unlock it once the hardware arrives.

Lattice-based cryptography hides data in what Bettinger calls a mathematical fog — a vast multi-dimensional grid laced with random noise, where the right answer is buried among thousands of wrong ones. Even a quantum computer gets lost in it.

Even a quantum computer gets lost in this mathematical fog.

There’s a bonus: a related idea, homomorphic encryption, lets you compute on data without ever unlocking it. One example — a model trained on 300,000 patient records from three hospitals without exposing a single patient’s file. The US has already made lattice-based methods core to its quantum-safe standards.

The threads tying it together

Two patterns run through the ten. More personal: medicine aimed at one patient rather than a population average. More local: lithium, protein and electricity produced near where they’re used, loosening long supply chains. And a quieter third — doing more with less, cooling without electricity, results without extra strain on the planet. The closing note is about method itself: AI is starting to flip the order of science, running broad experiments before forming a hypothesis instead of after, which may shift where discovery happens and who funds it.

Key Takeaways

Everything to grid energy — EVs, buildings and batteries become two-way nodes that both draw and return power, smoothing demand spikes; enabled by new battery chemistries and control software. Benefits owners more than renters.
Direct lithium extraction — pulls lithium from brine in hours, not the up-to-two-years of evaporation ponds; modular, less water, frees the process from needing a high-altitude desert. Lithium’s price has fallen 80%+ since 2022, denting investment.
Passive radiative cooling — paints/films/tiles that emit heat into deep space, cooling surfaces below ambient air with zero electricity; up to 40% energy savings; cheap and retrofittable, strongest case in hot poor countries.
PFAS destruction — new methods (e.g. electrodes) finally break the carbon-fluorine bond in “forever chemicals,” moving from containing pollution to eliminating it; early, government-funded.
Precision fermentation — reprogrammed microbes brew animal-identical proteins, fats, cosmetics and chemicals in tanks; aimed at feeding 9bn within planetary limits; risks displacing farming livelihoods.
Exosome drug delivery — uses the body’s own messenger packets as trusted couriers to carry drugs, including across the blood-brain barrier; trials in Parkinson’s, Alzheimer’s, cancer, long COVID.
Personalized mRNA cancer vaccines — a vaccine built in weeks from a patient’s own tumour mutations; inverts pharma’s blockbuster model; currently $100,000+ per patient.
Quantum simulation for drug discovery — models molecules directly via physics instead of approximating, attacking the fact that 9 in 10 drug candidates fail trials; could make rare diseases economically viable.
World models — AI that learns from physical experience (video, sensors, motion) rather than only text; powers adaptable robots, better climate models, learning factories.
Lattice-based cryptography — hides data in a noisy mathematical grid even quantum computers can’t crack; counters “harvest now, decrypt later”; already core to US quantum-safe standards.
Selection criteria: likely to scale and consequential if it does — picked by a WEF expert advisory council, list is 14 years old.
Why tech stalls (from hindsight): missing supporting pieces, no first adopter willing to use a rough early version, or lack of money/nerve — almost never the science itself.

Claude’s Take

This is a competent, genuinely accessible tour, and the analogies (the fog, the spoon-dropping baby, the sledgehammer on a nut) earn their place. As a “what’s on the technology horizon” briefing for a curious non-specialist, it does the job cleanly.

The thing to keep in mind is what it is: a WEF marketing format, essentially an annual hype index with the serious numbers living in a report the podcast keeps pointing you toward. The two-person chat is warm but thin — Pomeroy mostly reads sentences from the report and Bettinger expands them, so you get the optimistic case in full and the hard parts in passing. The “next five years” claim deserves a raised eyebrow on at least half the list: world models and useful quantum simulation are real research directions but “scaling by 2031” is a brave bet, and personalised cancer vaccines at $100k+ are a long way from the billions of people who’d need them. To the credit of both the report and the hosts, they don’t hide the drawbacks — the lithium price problem, the renter-vs-owner unfairness of grid energy, the threat precision fermentation poses to farming families. That honesty is the best thing here and the reason it clears a 7 rather than landing lower.

The most interesting idea is buried at the end and isn’t even one of the ten: AI flipping the scientific method, letting you experiment broadly before forming a hypothesis. That reframing — and the meta-lesson on why technologies stall (it’s almost never the science) — is more durable than any single item on the list. Treat the ten as a well-curated reading list of things to watch, not as predictions. Score: 7/10 — clear, useful, and honest about trade-offs, but a survey skim by design, with the real depth one click away in the underlying report.