For the better part of a decade, the question that defined the technology industry was a question about software. How large could a model become, how fluently could it write, how many tasks could a chatbot fold into a single conversation? The 2026 edition of the Top 10 Emerging Technologies report, co-published on June 23 by the World Economic Forum and the open-access publisher Frontiers, suggests that the center of gravity has quietly moved. The technologies the report identifies as most likely to reshape industry, policy, and society over the next five years are, for the most part, no longer confined to screens. They act on power grids, drug pipelines, fermentation tanks, cooling systems, mines, and machines. After years of software-first artificial intelligence, the frontier is becoming physical.
The report names ten technologies, and the texture of the list is itself the argument. Everything-to-grid energy turns electric vehicles and buildings into two-way batteries that can return power to the grid on demand. Direct lithium extraction replaces slow evaporation ponds with engineered systems that pull battery-grade lithium from salt flats in hours rather than months. Passive radiative cooling materials keep buildings cool without drawing any power, reflecting sunlight back through the atmosphere into space. PFAS destruction breaks down so-called forever chemicals into harmless substances. Precision fermentation brews food ingredients and medicines using programmed microbes, electricity, and sugar. Exosome drug delivery borrows the body's own cellular packaging to carry medicine to diseased cells, and personalized mRNA cancer vaccines train a patient's immune system against their specific tumor. Quantum simulation narrows the field of drug candidates by modeling how molecules actually behave. World models give AI systems an internal sense of how the physical world unfolds, so they can predict the outcome of a scenario like a superstorm. And lattice-based cryptography hardens digital data against the quantum computers that may one day try to break it.
Eight of those ten technologies act directly on physical systems. Only world models and lattice-based cryptography live primarily in the realm of pure computation, and even those are pointed outward, toward storms and infrastructure rather than inward toward another benchmark. That ratio is the quiet headline of the 2026 report.
Why It Matters
The significance of this shift is easy to understate, because none of these technologies arrives with the theatrical flourish of a new chatbot. There is no demo to marvel at, no viral screenshot. Yet the report's framing is deliberate: competitive advantage, it argues, is migrating away from software and toward the ability to control infrastructure, materials, biological processes, and industrial data. The companies and countries that win the next phase may not be the ones with the most eloquent model, but the ones that can act on the physical world with precision—pulling lithium from brine, programming a microbe, returning electrons to a grid at exactly the right moment.
There is a second, subtler pattern threaded through the list. Several of the technologies break the long-standing link between geography and production. For most of industrial history, where something could be made was dictated by where the raw materials happened to sit, or where the climate cooperated. Direct lithium extraction loosens the grip of geology. Precision fermentation loosens the grip of farmland and weather, letting a protein or a drug ingredient be brewed in a tank in a city that could never grow the crop it replaces. Passive radiative cooling cools a building in a desert without a power plant behind it. Read together, these are not ten isolated gadgets but a set of levers that relocate where critical goods can be made—and, by extension, who gets to make them.
For the AI story specifically, the lesson is almost philosophical. The large language model did not vanish from the picture; it became infrastructure, a quiet layer underneath the harder problems of energy, medicine, and materials. World models are the clearest expression of this. Rather than predicting the next word, a world model learns from multi-modal data how reality itself behaves, so that an AI can reason about a flood or a factory floor the way a physicist reasons about a falling object. The intelligence is no longer the product. It is the lens through which a more stubborn, physical problem is finally brought into focus.
How the Field Is Reading It
What gives the 2026 report an unusual reflexive quality is how it was assembled. This year's list was produced through an AI-based nomination workflow developed by Frontiers, which screened more than 1,200 candidate technologies drawn from academic publications and industry sources, with cross-model validation. Human experts then assessed each candidate's novelty and real-world impact, and an Advisory Council reviewed the final selection. In other words, artificial intelligence was used to map the moment when artificial intelligence stops being the headline—an irony the report's authors seem entirely comfortable with.
The people behind the report frame the moment as a turning point rather than a retreat. Stephan Mergenthaler, Managing Director of the World Economic Forum, noted that while each technology could make a meaningful impact on its own, together they reveal new patterns across energy, medicine, and manufacturing that could challenge long-held assumptions about how technology addresses problems like food insecurity, climate change, and untreatable diseases. Frederick Fenter, Chief Executive Editor of Frontiers, was more direct about the inflection, describing it as a decisive shift in which the technologies with the greatest impact are moving from software toward the physical realm, even as AI continues to support progress on many fronts. Notably, Frontiers describes itself as an AI-first research publisher and used an AI-based discovery tool to build the very list that documents AI's changing role.
For researchers and policymakers, the more useful signal may be the categories themselves. A list dominated by energy storage, materials, and biology is a list that asks different questions than one dominated by apps. It points budgets, talent, and regulation toward laboratories and manufacturing lines rather than data centers alone—and it implies that the hard, slow, capital-intensive work of changing the physical world is back at the center of the innovation conversation.
What Comes Next
The report is designed to be used rather than merely admired. To help organizations plan around the ten technologies, Frontiers' chief editors co-developed a series of interactive Transformation Maps hosted on the World Economic Forum's Strategic Intelligence platform, offering a live view of how each technology is developing across sectors and connecting it to global priorities. The intent is to give policymakers, industry leaders, and research institutions a shared evidence base for decisions that will play out over years, not quarters.
None of these technologies is finished. Several sit at the awkward stage between a striking laboratory result and a dependable industrial process—the stage where most promising ideas either mature quietly or stall. Personalized cancer vaccines still have to clear the long gauntlet of trials and manufacturing logistics. Direct lithium extraction has to prove it can scale without new environmental costs of its own. Everything-to-grid energy depends as much on tariffs, standards, and utility cooperation as on the hardware itself. The report's five-year horizon is less a prediction than an invitation to watch which of these crosses from possibility into infrastructure.
What is worth watching, more than any single entry, is whether the pattern holds. If the 2027 list looks like this one—weighted toward the physical, with AI as the connective tissue rather than the spectacle—then 2026 will be remembered as the year the conversation turned. If software roars back, it will look like a pause. Either way, the direction of attention has shifted, and attention, in technology, tends to pull capital and talent behind it.
Closing Thoughts
There is something clarifying about a list that refuses to be dazzled by its own tools. The 2026 report was built with AI, published by an organization that calls itself AI-first, and yet its central message is that the most consequential work ahead lies in places AI cannot reach on its own: the chemistry of a salt flat, the immune system of a single patient, the physics of a building shedding heat into the night sky. Intelligence, in this telling, is necessary but not sufficient. It has to be married to matter.
That is a humbling note to end on, and probably a healthy one. The past decade taught us to be impressed by machines that talk. The decade the report gestures toward will ask us to be impressed by machines, materials, and biological systems that quietly do—that cool a city without power, that brew a medicine without a farm, that hand electrons back to a grid at midnight. If the screen was the stage for the first act of the AI era, the second act seems to be moving offstage, into the unglamorous and enormously consequential machinery of the physical world. The most interesting technologies of 2026 are the ones you will never see on a screen at all.
한글 요약
세계경제포럼(WEF)과 오픈액세스 학술 출판사 프런티어스(Frontiers)가 6월 23일 공동 발표한 '2026년 10대 신흥 기술' 보고서는, 지난 10년간 기술 논의의 중심이었던 '소프트웨어 중심 AI'에서 무게중심이 옮겨가고 있음을 보여줍니다. 선정된 10개 기술 가운데 8개가 전력망·소재·바이오·제조 같은 물리적 시스템에 직접 작용하는 것으로, 양방향 전력 거래(Everything-to-grid), 직접 리튬 추출, 복사 냉각 소재, PFAS 분해, 정밀 발효, 엑소좀 약물 전달, 맞춤형 mRNA 암 백신, 양자 신약 시뮬레이션, 월드 모델, 격자 기반 암호 등이 포함됐습니다.
보고서의 핵심 메시지는 경쟁 우위가 '가장 유창한 모델'에서 '물리적 세계를 정밀하게 제어하는 능력'으로 이동하고 있다는 것입니다. 여러 기술이 지리와 생산의 오랜 연결고리를 끊어, 기후나 지질 때문에 불가능했던 곳에서도 핵심 물자를 만들 수 있게 합니다. 이 흐름에서 AI는 더 이상 주인공이 아니라, 에너지·의료·소재 같은 까다로운 문제를 풀어내는 바탕 인프라로 자리합니다. 특히 '월드 모델'은 다음 단어를 예측하는 대신 현실 세계가 어떻게 작동하는지를 학습해, 초대형 폭풍 같은 시나리오의 결과를 예측합니다.
흥미롭게도 이번 목록은 프런티어스가 개발한 AI 기반 후보 발굴 워크플로가 1,200개 이상의 기술을 선별하고 전문가 평가와 자문위원회 검토를 거쳐 완성됐습니다. WEF의 슈테판 메르겐탈러와 프런티어스의 프레더릭 펜터는 이를 '결정적 전환점'으로 규정하며, 가장 큰 영향을 줄 기술이 소프트웨어에서 물리적 영역으로 옮겨가고 있다고 강조했습니다. 보고서는 WEF 전략 인텔리전스 플랫폼의 인터랙티브 '트랜스포메이션 맵'과 함께 정책·산업·연구 의사결정의 근거로 활용되도록 설계됐습니다. 참고: Frontiers, World Economic Forum.