Ten futures, one question

The Top 10 Emerging Technologies report is the fourteenth in an annual series through which the World Economic Forum names scientific advances it judges to be nearing a point of consequential decision. This year’s ten are everything-to-grid energy, direct lithium extraction, passive radiative cooling, PFAS destruction, precision fermentation, exosome drug delivery, personalised mRNA cancer vaccines, quantum simulation for drug discovery, world models, and lattice-based cryptography.

Each is presented twice: a descriptive overview written by the Forum and Frontiers, and a strategic outlook produced by the Dubai Future Foundation that works backwards from a “desirable future” and closes with an Imagining 2031 vignette. The report reads across the ten and names three recurring tendencies — the technologies are becoming more personal, more distributed, and more materially efficient. Its closing essay advances two structural theses: that the geography of production is decoupling from the physical conditions that once fixed it, and that value is migrating from what can be manufactured at scale to what can be produced at the point of use. It ends on two tensions it treats as decisive — trust and access — and on the observation that decisions about technology are being redistributed to a much wider set of actors than the industries that used to hold them.

One feature of the report belongs in any faithful summary, because it bears on everything that follows: the selection method changed this year. The expert survey of previous editions was replaced by an AI-based nomination workflow developed by Frontiers, which used three large language models to generate and cross-check candidates from Frontiers article data and industry news, scored them for “trendiness” against the frequency of related concepts in the literature, and drew business-traction signals from CB Insights. More than 1,200 candidates were narrowed to eighty, then to twenty, then voted to ten by an advisory council. The report states plainly that AI was used in its research and writing.

This is a credible and unusually well-written report whose persuasiveness runs ahead of its evidence — and the distance between the two is the thing an intelligent reader most needs to keep in view.

The science is sound and current, and the report largely refuses the utopian register that makes the genre easy to distrust. But the ten technologies sit at very different points on the evidence curve, and the report’s uniform confidence flattens that difference.

The report is at its most persuasive precisely where its evidence is least mature.

A personalised vaccine with a real signal in a randomised trial is narrated in the same settled present tense as a delivery technology with, on the report’s own admission, no approved product and an unsolved manufacturing problem. Read at a single confidence level, the report invites the reader to bank ten futures at once. Read as its authors intend — as signals approaching decision, not as arrivals — it is a serious and rewarding document, and should be priced as foresight rather than as evidence on which capital or clinical decisions can rest directly.

I. The evidence is real, and uneven Established / Emerging

Sorted by the maturity of the evidence rather than the order of presentation, the ten fall into three groups, and the grouping matters more than any single claim.

At the mature end sit the standards and the shelf products. Lattice-based cryptography is settled: the US National Institute of Standards and Technology finalised its post-quantum standards in 2024, and the migration is now a matter of timetables, not doubt. Precision fermentation has products in market and incumbent capital behind them — the clearest near-term case. Passive radiative cooling rests on established physics and is already written into building codes in California and China.

In the middle sit technologies whose science is real but whose performance at scale is unproven: direct lithium extraction, the coordination of millions of grid-connected assets, and the verification of PFAS destruction at scale — each promising, each early.

At the near end of the curve sit the technologies carried largely by promise. Exosome drug delivery has, on the report’s own account, no approved therapy and an unsolved manufacturing problem. Quantum simulation for drug discovery is evidenced principally by a doubling in market value — a signal about capital, not about cures. World models are newer still, and the report’s own warning that such a system “can be internally consistent and still wrong” is the most useful sentence in the document.

Personalised mRNA cancer vaccines are the report’s evidentiary anchor and its most instructive case. The melanoma result — a personalised vaccine added to immunotherapy, reducing recurrence in a randomised trial now advancing to phase 3 — is close to established as a genuine efficacy signal. The pancreatic result is more fragile than its placement suggests: it is small and early, and the striking survival figure describes patients whose immune systems responded, a comparison that selects for favourable biology. The finding is real and promising; the framing invites a within-trial correlation to be read as a population survival rate. That single choice is the report’s central weakness in miniature.

II. Why the persuasion outruns the proof Expert judgement

The mechanism is the house form itself. The Imagining 2031 vignettes render each technology as already woven into ordinary life, and the prose does not modulate with the evidence: the standards-grade case and the first-in-class demonstration are narrated with the same assurance. The effect is to lend the weak cases the confidence earned by the strong ones. This is a property of the form, not a failing of any author — and the report would be more trustworthy, not less, if its confidence visibly rose and fell with its proof.

III. The method changed, and it deserves scrutiny Expert judgement

The move from an expert survey to an AI-driven nomination workflow is the report’s most consequential and least examined decision. It is disclosed with real candour, and the intent — breadth and consistency at scale — is legitimate. But three features shape how much independent weight the selection carries. The “trendiness” signal is drawn from the frequency of concepts in the literature, mapped by a tool built by Frontiers, from Frontiers article data — and Frontiers is also the co-publisher whose journals supply the report’s own expert voices. The business-traction score imports the market’s enthusiasm into what is meant to be a scientific judgement. And any filter tuned to article frequency and recent funding will, by construction, favour the already-visible; the genuinely early or contrarian is precisely what it is least likely to surface. The advisory council’s human vote is a real check, and a good one. But disclosure is not neutralisation. A selection produced this way is best read as one well-resourced organisation’s defensible view of the frontier, not as a neutral reading of it.

IV. What it gets right, said as plainly as its faults Expert judgement

A fair appraisal acquits as readily as it convicts, and this report earns its acquittals. It refuses hype: the strategic outlooks are full of the conditions and risks that lesser foresight omits. Its two structural theses — the decoupling of production from place, and the migration of value to the point of use — are a genuine synthesis that makes the ten legible as a pattern rather than a list, and will be worth thinking with long after several of these technologies have stalled. And its closing move — that decisions devolved to municipalities, hospitals and sovereign funds are, in substance, a redistribution of responsibility for whether these technologies are trusted and shared — reframes emerging technology as a governance problem rather than a horse race. It is the most durable idea in the document.

V. How good, relative to its field Expert judgement

A review should say not only whether a report is good, but how good, against the work it sits beside.

Against its own back-catalogue, this is among the more disciplined editions the series has produced: less breathless than the Forum’s foresight has sometimes been, and anchored by a governance essay that earlier editions lacked. The move to AI-assisted selection is the sharpest break from its own tradition.

Against the other general-audience surveys, its character is distinctive. MIT Technology Review’s annual breakthroughs are more journalistic and consumer-facing, willing to name a single winning product but rarely engaging the systems around it; the Forum is more systemic and more reluctant to name losers. McKinsey’s technology outlooks are heavier on adoption curves and investment figures, and sit closer to the market they describe; RAND’s foresight is more sceptical and security-minded. This report lands between the two — nearer McKinsey in optimism, nearer RAND in its willingness to name unresolved risk.

Against the evidence authorities, the comparison is the most clarifying, because it is the one the report cannot win and does not try to. Nature Outlooks and Lancet Commissions are peer-reviewed, narrow and evidence-graded; the OECD’s science and technology outlooks are drier, denser and built around policy instruments. Each is a better guide to whether a specific claim is true. None offers what this report offers: synthesis across ten unlike domains and a frame for the decisions they force. Its place in the field follows from that. As an evidence authority it is mid-tier and does not pretend otherwise; as a cross-domain orientation for decision-makers it is at or near the top of its class. The reader who wants adjudication should go to Nature; the reader who wants a map should start here.

VI. What it means for Oman and the Gulf Strategic interpretation

This reading is Cordelia’s, not the report’s, which mentions neither Oman nor — beyond a single reference to Abu Dhabi’s Estidama — the Gulf, and asks none of the questions a small, energy-rich, land-poor state must ask.

The report’s central thesis fits Gulf conditions almost too well. If production is decoupling from natural endowment, then precision fermentation reframes food security as a matter of clean power and capital rather than arable land; direct lithium extraction opens battery-grade supply from the geothermal and oilfield brines a hydrocarbon economy holds in abundance; passive radiative cooling addresses the region’s defining cost; and the health technologies sit squarely within a sovereign health-intelligence ambition and Vision 2040.

But the report’s access tension cuts against small states more sharply than it notices, and this is the reading a sovereign investor most needs. The same decoupling that frees production from place concentrates its new inputs — strain libraries, refining capacity, sequencing and manufacturing, model libraries — in a few hubs. A state can capture that concentration or be captured by it, and which of the two is not a technology question but a question of which capabilities are built at home and which are rented. The report frames that choice with force. It cannot, and does not try to, tell a sovereign fund which of the ten are investable today. That judgement is a different instrument from this Review — and the distance between the two is the distance between a public appraisal and a paid one.

Five questions run beneath every Cordelia appraisal: what does the report assume, what does it leave unsaid, how good is its evidence, what would change our view, and how far do its conclusions travel? The eight lenses put those questions to work.

Applied honestly, the six fixed absences acquit the report on some counts and convict it on others. Regulation and financing are, in fact, addressed — each strategic outlook engages both directly. Workforce is partially addressed. But ethics is thin: the specific questions — consent and equity in personalised medicine, biosecurity in engineered strains, the labour displacement the report names but does not sit with — are gestured at rather than examined. Middle-income and developing-country implementation is largely absent: artemisinin supply in Sub-Saharan Africa is invoked as motivation, not analysed as a delivery problem. And GCC-specific context is absent — despite Dubai Future Foundation authorship, there is no Gulf analysis beyond one reference to Estidama. That last gap is the one Cordelia exists to fill.

The framework is not a score. It is a consistent, comparable summary of judgement that places this report — and any report, from any institution — on the same map. It sits beside the issue’s overall confidence: High.

Appraisal, not instruction.

What is at stake in a report like this is not whether ten particular technologies arrive on schedule. Half of them will not, and the report is honest enough to say that none is a finished story. What is at stake is the quality of attention that institutions bring to the frontier.

The report’s gift is a habit of mind: it insists, rightly, that whether a technology arrives well is settled less in the laboratory than in the municipalities, hospitals, regulators and sovereign funds onto whom the decisions have been quietly devolved. Its risk is a second, unintended habit — to hold ten unlike bets at one level of confidence, because the prose holds them there.

The defence is the question this issue began with. Of every item on the page, ask how much is proven, how much is funded, and how much is merely well-written — and notice that the three have different answers for each. Read as the beginning of a question rather than the end of one, this is among the more thoughtful documents its genre produces, and worth recommending to any strategist, policymaker or investor who needs an intelligent orientation to where the frontiers sit. It is not the ground on which to rest a clinical or capital decision, and it does not ask to be. The judgement built on top of it is the reader’s own — which is, in the end, the report’s own most durable claim.

The source report. World Economic Forum, in collaboration with Frontiers, Top 10 Emerging Technologies of 2026 (June 2026) — read it on the World Economic Forum’s site. Every quantitative claim discussed above is the report’s own, reproduced for appraisal and treated as directional; the characterisations of evidence maturity are the house’s reading of the field.

Would a reader who disagrees with our conclusion still trust our reading of the evidence?