◇AI2028–2030
AI-Generated Video Indistinguishable from Reality
Generative video models such as OpenAI's Sora, Runway Gen-3, Google's Veo and Kuaishou's Kling have made rapid progress in producing coherent, high-resolution clips from text prompts. While current outputs still contain artifacts on close inspection, the rate of improvement suggests that within a few years AI-generated video will be indistinguishable from real footage for most viewers and many professional use cases. Adobe, Disney, Netflix and major news organisations are already integrating generative video into production workflows for pre-visualisation, storyboards and background shots. The implications are broad: lower costs for film and advertising production, new tools for creators, but also serious risks around misinformation, non-consensual imagery and evidentiary integrity. Several governments and the EU's AI Act are mandating provenance labelling, watermarking (C2PA) and disclosure of synthetic media. Detection tooling is improving in parallel but is widely expected to lag generation. By the late 2020s the consensus view is that high-fidelity synthetic video will be a routine creative medium, with the social and regulatory infrastructure for distinguishing authentic footage still maturing. Deepfake-related fraud, including voice-cloning scams and synthetic CEO videos, is already materialising and is expected to grow before defensive technologies stabilise.
#generative video#deepfake#Sora
HighEarly Adoption
◈Space2028–2030
Reusable Heavy-Lift Rockets in Routine Operation
Reusable heavy-lift rockets — vehicles capable of placing 50+ tonnes into low Earth orbit (LEO) and recovering both stages for repeated flight — are expected to enter routine commercial operation in the late 2020s. SpaceX's Falcon 9 has demonstrated routine reuse of the first stage since 2015, with over 300 booster landings and more than 250 reflights as of 2024. The Falcon Heavy configuration flies periodically. The next step is fully reusable two-stage vehicles. SpaceX's Starship-Super Heavy system, which has been conducting integrated flight tests since 2023 and successfully caught the Super Heavy booster with 'Mechazilla' arms at the launch pad in October 2024, has the stated goal of carrying 100–150 tonnes to LEO in fully reusable mode, with marginal launch costs projected in the low tens of millions of dollars. Blue Origin's New Glenn, which launched its first test flight in 2024, is designed for 45 tonnes to LEO with a reusable first stage. China's Landspace (Zhuque-3) and CAS Space are pursuing similar vehicles, and the European Space Agency is investigating reusable first-stage technologies through its Prometheus and Themis programmes. The U.S. Federal Aviation Administration is updating its launch-licensing framework to accommodate higher flight rates. By the late 2020s the consensus view is that several fully reusable heavy-lift vehicles will be flying weekly or more frequently, transforming orbital launch economics and enabling the construction of large orbital infrastructure such as space stations, solar-power demonstrators and human-rated Mars vehicles.
#reusable rockets#launch#Starship
HighPrototype
◉Society2028–2030
Global Internet via Satellite Constellations
Low-Earth-orbit (LEO) satellite constellations are delivering broadband internet to virtually any location on Earth, including regions poorly served by terrestrial infrastructure. SpaceX's Starlink constellation has launched more than 6,000 satellites and provides service in more than 100 countries; Eutelsat OneWeb operates more than 600 satellites in a constellation focused on enterprise and government customers; Amazon's Project Kuiper has begun launching satellites and plans to begin service in 2025–2026; and China has announced its 'Guowang' and 'G60' constellations totalling more than 25,000 planned satellites. The International Telecommunication Union (ITU) estimates that roughly 2.6 billion people remained offline in 2024, mostly in low- and middle-income countries, with cost and lack of infrastructure being the principal barriers. LEO broadband is gradually closing this gap in remote and underserved areas. The economic and social effects are substantial: improved access to education, telemedicine, financial services and markets for rural communities; new options for remote work and distributed teams; and improved disaster response and resilience. The principal challenges include the cost of user terminals (currently USD 200–600), spectrum coordination, orbital-debris management (the U.S. Federal Communications Commission has set new rules for de-orbiting post-mission), and competition with terrestrial fibre in dense urban areas. By the late 2020s the consensus view is that LEO satellite broadband will be a mainstream global connectivity option, complementing terrestrial networks and providing genuine competition to incumbent telecom operators in many markets. Astronomical light pollution and the long-term sustainability of LEO are ongoing concerns that ITU and UNOOSA are addressing.
#broadband#satellite#connectivity
HighEarly Adoption
◇AI2028–2032
AI-Designed Drugs Reaching Clinical Approvals
AI is rapidly moving from a research curiosity to a core tool in pharmaceutical pipelines. Following DeepMind's release of AlphaFold, structure-prediction models have become standard infrastructure at most large biotech firms, and generative models are now producing novel candidate molecules for targets that were previously considered undruggable. Several AI-discovered molecules have entered clinical trials, including Insilico Medicine's INS018_055 for idiopathic pulmonary fibrosis and Exscientia's opioid-receptor modulators. The FDA has signalled openness to AI-generated submissions as long as traditional safety and efficacy standards are met. Industry analysts at McKinsey and the Boston Consulting Group expect the first AI-originated drugs to reach the market in the late 2020s, with broader adoption in the 2030s. The technologies involved include protein-structure prediction, generative chemistry, active learning for assay prioritisation and large language models that summarise biomedical literature. AI is also shortening the time between target identification and candidate selection, although it does not yet compress the lengthy clinical-trial phase. Remaining bottlenecks include the cost and duration of Phase II and III trials, regulatory alignment on model documentation, and the difficulty of validating targets in complex diseases such as Alzheimer's. Overall impact is expected to be a steady increase in pipeline productivity rather than a sudden disruption.
#drug discovery#pharmaceuticals#AlphaFold
HighEarly Adoption
✦Biotech2028–2032
CRISPR Therapies for Common Diseases
CRISPR-based gene editing has moved from laboratory tool to approved medicine. In December 2023 the U.K. Medicines and Healthcare products Regulatory Agency and the U.S. Food and Drug Administration approved Casgevy (exa-cel), the first CRISPR-Cas9 therapy, for sickle-cell disease and transfusion-dependent beta-thalassaemia. The therapy edits a patient's own haematopoietic stem cells ex vivo to reactivate fetal haemoglobin production. Several other CRISPR programmes are in clinical trials, including Intellia's NTLA-2001 for transthyretin amyloidosis (in-vivo liver editing), Verve Therapeutics' base-editing programme for familial hypercholesterolaemia, and Editas Medicine's work on inherited blindness. Industry analysts at Nature Biotechnology and the MIT Technology Review expect that the late 2020s and early 2030s will see the first CRISPR therapies for non-monogenic conditions such as high cholesterol, HIV and certain forms of cardiovascular disease. Remaining challenges include the cost of personalised manufacturing (the first Casgevy treatments are listed at roughly USD 2 million per patient), the need for safer in-vivo delivery vehicles, and the long-term safety monitoring required for permanent genetic edits. Lipid-nanoparticle and adeno-associated virus delivery systems are improving steadily, and base- and prime-editing are reducing off-target effects. By the early 2030s a small but growing portfolio of approved gene-editing therapies for common conditions is the consensus expectation, contingent on continued regulatory and manufacturing progress.
#CRISPR#gene editing#therapeutics
HighPrototype
✦Biotech2028–2032
Universal Flu Vaccine
A universal influenza vaccine would protect against a broad range of flu strains, including those that have not yet emerged, by targeting conserved regions of the virus rather than the rapidly mutating head of the haemagglutinin protein. The U.S. National Institute of Allergy and Infectious Diseases (NIAID) published a strategic plan for a universal flu vaccine in 2018, and several candidates have since advanced through clinical trials. Moderna, Pfizer, BiondVax, Scicom and the NIH's Vaccine Research Center are pursuing approaches including mRNA-encoded haemagglutinin stems, chimeric proteins, and nanoparticle arrays that present conserved epitopes. Early Phase I and II trials have demonstrated broadened antibody responses, and a small number of candidates are entering Phase III efficacy studies. Success would reduce the need for annual reformulation and annual vaccination, improve global preparedness against pandemic strains such as H5N1, and lower the disease burden of seasonal influenza, which the WHO estimates causes 290,000–650,000 respiratory deaths each year. The consensus expectation is that the first broadly protective flu vaccine will reach the market in the late 2020s or early 2030s, with broader coverage against multiple subtypes developing through the 2030s. Manufacturing readiness, regulatory alignment and procurement strategies for low- and middle-income countries remain significant implementation challenges.
#vaccine#influenza#immunology
HighPrototype
✦Biotech2028–2032
Personalized Cancer mRNA Vaccines
Personalised cancer vaccines use messenger RNA to instruct a patient's immune system to recognise and attack the unique mutation profile (neoantigens) of their own tumour. The approach builds on the success of mRNA COVID-19 vaccines but tailors the encoded antigens to each individual. The most advanced programme is the collaboration between Moderna and Merck on mRNA-4157, which is being combined with the checkpoint inhibitor Keytruda in adjuvant melanoma treatment. Phase IIb results reported in 2023 showed a roughly 44% reduction in the risk of recurrence or death compared with Keytruda alone, and a Phase III trial is enrolling. BioNTech is running parallel trials in pancreatic cancer, colorectal cancer and head-and-neck cancer, and is advancing its Individualised Neoantigen Therapy (INT) platform. Genentech is pursuing a similar approach with its Synthetic Long Peptide vaccines. Industry consensus is that the first approved personalised cancer mRNA vaccines will reach the market in the late 2020s, initially for high-recurrence-risk melanoma and possibly pancreatic cancer. Manufacturing turnaround time — typically six to eight weeks to sequence a tumour, design the construct, manufacture it and ship it — is the main practical bottleneck, along with the cost of a fully individualised therapy. Broader uptake across multiple tumour types is expected through the 2030s as manufacturing infrastructure matures and reimbursement pathways are defined.
#mRNA#oncology#immunotherapy
HighPrototype
△Energy2028–2032
Perovskite-Silicon Tandem Solar Cells at Commercial Scale
Perovskite-on-silicon tandem solar cells stack a perovskite top cell on a conventional silicon bottom cell, capturing a broader portion of the solar spectrum and pushing cell efficiency beyond the practical limit of single-junction silicon (around 26%). The technology has progressed rapidly: Oxford PV holds the world record for a commercial-format tandem cell at 28.6% efficiency, and LONGi, Trina, JinkoSolar and Hanwha Q Cells have all reported >33% lab-cell efficiencies. The International Technology Roadmap for Photovoltaics (ITRPV) expects tandem cells to enter mass production from 2025 onwards, with rapid scaling through the late 2020s. Manufacturing challenges include the stability of perovskites under heat, moisture and ultraviolet light; lead containment; and the cost of additional deposition equipment on existing silicon lines. Once solved, tandem modules should deliver a meaningful efficiency uplift at modest incremental cost, reducing the levelised cost of solar electricity further. The IEA's World Energy Outlook 2024 projects solar PV to become the largest source of global electricity by the mid-2030s, and tandem technology is expected to be a meaningful contributor to that growth from the late 2020s onwards. By 2030 the consensus view is that commercial tandem modules will be available from multiple major manufacturers at scale, with a steadily rising market share through the 2030s as manufacturing processes mature and bankability improves.
#solar#photovoltaics#perovskite
HighPrototype
▣Robotics2028–2032
Robotic Surgery in Rural Hospitals
Robotic surgery systems — initially exemplified by Intuitive Surgical's da Vinci platform — have become standard in many urban and academic hospitals for urological, gynaecological, colorectal and cardiac procedures. The next frontier is broad deployment in rural and underserved hospitals, where surgical expertise is scarce and patients often have to travel long distances for complex procedures. Several trends are converging to make this possible. New lower-cost systems such as Medtronic's Hugo, CMR Surgical's Versius, and Asensus Surgical's Senhance are expanding the competitive landscape and reducing capital costs. The integration of 5G networking is enabling telesurgery, in which a surgeon at a remote site controls a robot at the patient location, with sub-second latency. The first major demonstration was a transatlantic laparoscopic cholecystectomy in 2001; subsequent trials in China have demonstrated broader use of 5G telesurgery. Augmented reality overlays, AI-assisted planning and intraoperative imaging are improving surgical precision. The WHO reports a global shortage of approximately 15 million healthcare workers, and surgical access is highly uneven. Consensus projections place robotic surgery in a meaningful fraction of rural hospitals in advanced economies by the late 2020s, with broader deployment in middle-income countries in the 2030s. Reimbursement, training and connectivity infrastructure are the principal implementation challenges. By the early 2030s robotic-assisted procedures are expected to be a routine option in mid-sized community hospitals, including those in underserved regions.
#medical robotics#surgery#telemedicine
HighEarly Adoption
▣Robotics2028–2032
Humanoid Robots in Warehouses at Scale
Warehouse automation has evolved rapidly from fixed conveyor systems to fleets of mobile robots following Amazon's 2012 acquisition of Kiva Systems. The next step is general-purpose humanoid robots capable of performing tasks designed for human workers — picking items from shelves, packing boxes, loading pallets, operating existing material-handling equipment. Several companies have begun pilots. Agility Robotics' Digit robot has been deployed by Amazon and GXO Logistics; Figure's Figure 02 is being tested by BMW at its Spartanburg plant; Apptronik's Apollo has been piloted with Mercedes-Benz; Boston Dynamics' new electric Atlas is being developed for industrial applications; and 1X Technologies' NEO is targeting consumer and commercial use. The integration of large vision-language-action models such as Google's RT-2 and Figure's Helix is dramatically improving the ability of these robots to perform novel tasks with minimal reprogramming. Industry analysts at RBC Capital Markets and Goldman Sachs project the humanoid robot market will reach USD 38 billion by 2035, with warehousing and logistics being one of the earliest and largest segments. Key challenges include battery life, manipulation in cluttered environments, reliability over long operating hours, and integration with existing warehouse management systems. The consensus expectation is that humanoid robots will be deployed at meaningful scale in warehouses operated by major retailers and logistics firms from the late 2020s, with broader deployment through the 2030s as costs fall and reliability improves.
#humanoid#logistics#warehouse
HighPrototype
▣Robotics2028–2032
Fully Autonomous Mining Operations
Autonomous mining systems operate haul trucks, drills, loaders and dozers without on-board human operators, supervised from remote operation centres. Rio Tinto's Pilbara iron-ore operations pioneered the approach with its Mine of the Future programme, and the company now operates more than 200 autonomous trucks across multiple sites and moves roughly a third of its Pilbara material autonomously. BHP, Fortescue, Suncor, Codelco and several North American potash and oil-sands operators have followed. Caterpillar and Komatsu have delivered more than 700 autonomous haul trucks collectively. The next wave of automation includes autonomous drilling, autonomous explosive charging, and surface and underground loaders. The International Council on Mining and Metals (ICMM) and the World Economic Forum expect most large surface mines in advanced economies to operate with significant autonomous fleets by the early 2030s. The principal benefits are safety, with haul-truck-related fatalities reduced to near zero at fully autonomous sites, and lower operating costs through improved fuel efficiency, tyre life and equipment utilisation. Underground mining automation is harder because GPS does not work underground, but companies such as Sandvik and Epiroc have made significant progress using lidar, SLAM and 5G-connected teleremote operation. The consensus expectation is that fully autonomous mines — encompassing haulage, drilling, loading and ancillary operations — will become the standard for new greenfield surface mines in the late 2020s, with retrofits of existing mines continuing through the 2030s. Workforce transition is a significant social and economic challenge that governments and companies are still learning to manage.
#mining#autonomous vehicles#automation
HighEarly Adoption
➤Transportation2028–2032
Autonomous Trucking on Highways
Autonomous trucking refers to class 8 heavy-duty trucks operating without a driver on selected highway corridors, typically between logistics hubs on interstate or motorway networks where a human driver handles local street driving at either end. Several companies are actively running pilot and early commercial operations: Aurora Innovation launched commercial driverless freight service between Dallas and Houston in 2024; Kodiak Robotics operates driverless routes in Texas; Plus.ai has commercial operations in the U.S. and China; Waymo Via is testing in Texas and Arizona; and Embark Trucks (which ceased operations in 2023) illustrated the financial risks of the sector. In China, TuSimple and Inceptio have demonstrated drivered and driverless operations. The American Transportation Research Institute (ATRI) estimates that driver pay and benefits represent roughly 40% of marginal operating costs per mile, providing a strong economic incentive for automation. The principal technical challenges — long-tail edge cases, weather, and emergency vehicle interactions — are being addressed through high-definition mapping, redundancy in sensors and compute, and operational design domains limited to fair-weather interstate driving. Regulatory frameworks are evolving at the U.S. federal and state levels, and the EU is preparing updated rules for automated driving systems. The consensus expectation of industry analysts at S&P Global Mobility and BloombergNEF is that meaningful deployment of driverless trucking on U.S. interstate corridors will begin in the late 2020s, scaling through the 2030s. The implications for the roughly 3.5 million truck drivers in the U.S. alone are significant and will require workforce-transition policy.
#autonomous trucks#logistics#freight
HighPrototype
◇AI2030–2035
Autonomous Vehicles at Scale
Autonomous vehicles (AVs) at scale refers to the deployment of driverless taxis and freight vehicles operating without human supervision across most urban and highway environments. Waymo, Cruise and several Chinese operators already run commercial driverless services in geofenced cities such as Phoenix, San Francisco, Austin, Wuhan and Beijing. Industry roadmaps from BloombergNEF and S&P Global Mobility project that the number of cities with commercial robotaxi service will expand materially through the late 2020s, with highway-based autonomous freight following shortly after. The remaining engineering challenges include long-tail edge cases involving construction, emergency vehicles, weather and unusual pedestrian behaviour, as well as regulatory frameworks that vary by jurisdiction. Achieving scale depends on continued reductions in sensor and compute costs, validated safety cases, and public acceptance. Insurance and liability models are being reworked in parallel. Most analysts expect that by the early 2030s a meaningful fraction of urban ride-hailing will be driverless in selected markets, and that autonomous trucking will become common on interstate corridors. Wider consumer ownership of fully self-driving personal vehicles is expected to lag fleet deployment by several years because of cost, trust and infrastructure considerations. The cumulative effect on cities, including reduced parking demand, redesigned streets and changes in public transit, is expected to unfold gradually over the 2030s and 2040s.
#self-driving cars#mobility#robotics
HighEarly Adoption
◇AI2030–2035
AI Personal Assistants with Long-Term Memory
Personal AI assistants that maintain a long-term memory of an individual user's preferences, contacts, schedule, files and history are widely expected to become a mainstream product category in the 2030s. Today's large language models handle individual conversations competently but reset between sessions unless explicitly primed. Research from Anthropic, Google DeepMind and OpenAI on memory architectures, retrieval-augmented generation and agentic planning is gradually closing this gap. Apple Intelligence, Google Gemini and Microsoft Copilot have all begun integrating limited personal context, but the depth and continuity required for a true digital assistant remain a research goal. Key technical challenges include selective memory (deciding what to remember and what to forget), privacy preservation across cloud and edge devices, multi-modal memory (text, images, audio) and the ability to plan and execute multi-step tasks across applications. Equally significant are the social and trust issues: users will need confidence that their assistant acts in their interest, that memory is private, and that the assistant's recommendations are not biased by commercial incentives. Regulatory frameworks such as the EU AI Act and emerging US state laws will shape deployment. By the mid-2030s, a personal assistant that has followed an individual for years — knowing their tastes, contacts and goals — is widely considered plausible, assuming continued progress in models, context windows and on-device inference.
#LLM agents#personal AI#memory
MediumPrototype
◇AI2030–2035
Real-Time Cross-Lingual Translation Earpieces
Wearable earpieces that translate spoken conversation in near real time are rapidly becoming practical. Google's Pixel Buds, the WT2 Plus translator earbuds and several competitors already deliver two-way translation across dozens of language pairs, with latency measured in seconds rather than minutes. The combination of edge AI chips, smaller transformer-based speech models and the latest large-vocabulary speech recognition has pushed accuracy and latency close to the threshold needed for fluid, conversational use. Widespread consumer adoption is expected through the early 2030s as accuracy improves for low-resource languages and as latency drops below one second for common pairs. Key remaining challenges include handling code-switching, overlapping speakers, strong accents and noisy environments, as well as supporting the more than 3,000 of the world's 7,000+ spoken languages that currently lack adequate machine translation. UNESCO and the OECD have both flagged language preservation as a concern if translation technology concentrates on a small set of dominant languages. By the mid-2030s the consensus view is that always-on translation earpieces will be a mainstream consumer-electronics category, comparable in scale to wireless headphones today, with significant implications for travel, cross-border commerce, education and the structure of multilingual societies.
#speech translation#wearables#natural language
HighPrototype
✦Biotech2030–2035
Lab-Grown Meat at Cost Parity
Cultivated meat — animal tissue grown in bioreactors from starter cells, without raising or slaughtering animals — has moved from a single USD 330,000 burger in 2013 to limited commercial sales in Singapore and the United States. Upside Foods and Good Meat received the first U.S. approvals in 2023, and the Singapore Food Agency had cleared cultivated chicken as early as 2020. The Good Food Institute reports that more than 150 companies are now active in the sector. Cost remains the principal barrier: estimates from McKinsey & Company and CE Delft suggest production costs fell from roughly USD 17,000 per kilogramme in 2019 to a few hundred dollars per kilogramme by 2024, but parity with conventional chicken (around USD 3–5 per kilogramme) requires further reductions in growth-medium costs, bioreactor scale-up and downstream processing. Most analysts project that cultivated meat will reach cost parity with premium animal proteins (such as certain seafood and premium beef) in the early 2030s, with broader parity later in the decade. Regulatory pathways are being defined in the EU, the U.K., China and several other markets. Environmental impact is contested, with some life-cycle analyses suggesting significant land and water savings but potential emissions from energy-intensive facilities. By the mid-2030s the consensus expectation is that cultivated meat will be a niche but growing segment of the global protein market, complementing plant-based alternatives.
#cultivated meat#cellular agriculture#food
MediumPrototype
△Energy2030–2035
Grid-Scale Solid-State Batteries
Solid-state batteries replace the flammable liquid electrolyte of conventional lithium-ion cells with a solid ceramic, polymer or sulphide electrolyte. The benefits include higher energy density, lower risk of thermal runaway, longer cycle life and the ability to use lithium-metal anodes. Toyota, QuantumScape, Solid Power, Samsung SDI, CATL and several Chinese manufacturers are pursuing automotive-grade cells; QuantumScape has reported cells exceeding 800 cycles with high capacity retention, and Toyota has targeted limited vehicle deployment by 2027–2028. The same technology can be scaled to grid storage, where longer life, higher efficiency and improved safety materially reduce the levelised cost of storage. BloombergNEF projects that solid-state battery cells will enter automotive mass production in the late 2020s, with cost falling below USD 80/kWh at the pack level by the mid-2030s. For grid-scale applications, where energy density is less critical, the principal benefit of solid-state technology is cycle life and safety, potentially enabling 10,000+ cycle batteries suitable for 20-year project lifetimes. The consensus expectation is that solid-state batteries will be a substantial grid-storage segment in the 2030s, complementing iron-air, sodium-ion and redox-flow technologies for long-duration storage. Manufacturing scale-up is the principal bottleneck: solid electrolyte synthesis and the 'dry room' requirements are more demanding than for conventional cells, and existing gigafactory capacity is sized for liquid-electrolyte designs.
#batteries#storage#solid-state
HighPrototype
△Energy2030–2035
Small Modular Reactors at Scale
Small Modular Reactors (SMRs) are advanced fission reactors with an electrical output typically below 300 MWe per unit, designed to be factory-built and shipped to site rather than constructed bespoke. The objective is to reduce the cost overruns and schedule slippage that have plagued large-scale nuclear builds in Europe and the United States. Several SMR designs are now in advanced stages: NuScale's VOYGR (recently cancelled by the UAMPS Carbon-Free Power Project but with continued international interest), GE Hitachi's BWRX-300 (under construction in Canada at Darlington), Rolls-Royce SMR's 470 MWe PWR (in UK generic design assessment), and the U.S. Terrapower Natrium sodium-cooled reactor (under construction in Wyoming). Russia's floating Akademik Lomonosov has been operational since 2020, and China's HTR-PM pebble-bed reactor has been grid-connected since 2021. China National Nuclear Corporation's ACP100 Linglong One is also nearing completion on Hainan Island. The International Energy Agency and the OECD Nuclear Energy Agency both project that SMRs will become a meaningful part of the global nuclear fleet from the early 2030s onwards, particularly for industrial heat, district heating, hydrogen production and replacing retiring coal plants. Key challenges include first-of-a-kind costs, supply-chain readiness, regulatory alignment across jurisdictions, and public acceptance. By the mid-2030s the consensus view is that several SMR designs will be in commercial operation, with broad deployment through the late 2030s and 2040s contingent on cost reduction.
#SMR#nuclear#fission
HighPrototype
△Energy2030–2035
Long-Duration Energy Storage (100+ hours)
Long-duration energy storage (LDES) refers to technologies that store electricity for periods of 10 hours or more, and in some cases several days or weeks. Such storage is increasingly seen as essential for high-renewables grids where wind and solar output can be low for extended periods, sometimes called Dunkelflaute events. The U.S. Department of Energy's Long Duration Energy Shot targets a 90% cost reduction for 10-hour-plus storage by 2030. Multiple technologies are competing: iron-air batteries (Form Energy, which deployed its first commercial system in 2023), sodium-ion batteries (CATL, BYD), zinc-based chemistries (Eos Energy), redox flow batteries (Invinity, VRB,ESS), liquid-metal batteries (Ambri), gravity storage (Energy Vault), thermal storage (Malta, Antora) and compressed-air energy storage (Hydrostor). BloombergNEF's 2024 Energy Storage Outlook estimates that global stationary storage deployments will rise from roughly 70 GWh in 2024 to more than 1,500 GWh by 2030, with long-duration technologies representing an increasing share. The consensus expectation is that the first commercial 100-hour-plus storage projects will be operating by the late 2020s, scaling through the 2030s as renewables penetration rises and the value of long-duration discharge increases. Costs are expected to fall below USD 100/kWh of delivered energy for the most promising chemistries, making them competitive with peaking gas plants in many markets.
#LDES#storage#grid
MediumPrototype
◈Space2030–2035
Commercial Low-Earth Orbit Space Station
Commercial low-Earth-orbit (LEO) space stations are intended to replace some of the capabilities of the International Space Station (ISS), which NASA plans to de-orbit by 2030 through a controlled re-entry over the South Pacific Ocean Uninhabited Area. NASA's Commercial LEO Development programme has funded several private consortia to design and build successor stations: Axiom Space (attaching modules to the ISS before detaching as a free-flying station), Blue Origin's Orbital Reef partnership with Sierra Space, Starlab (a partnership of Nanoracks, Lockheed Martin and Voyager Space), and Northrop Grumman's independent design. The China Tiangong space station has been permanently crewed since 2022 and is expected to remain operational through the 2030s, with plans for a commercial extension. Russia has announced plans for the ROSS space station, initially in a high-inclination orbit, but its schedule is uncertain. Axiom has already flown three private astronaut missions to the ISS. The consensus expectation is that the first free-flying commercial LEO station will be operational by the early 2030s, with several competing facilities available by the mid-2030s. These stations will host commercial research, pharmaceutical production, materials science, tourism and government astronaut missions. Business models are still maturing, and the principal near-term customer for human-rating and operations is NASA itself, which plans to be one of many tenants rather than the sole operator.
#LEO#commercial space#space station
HighPrototype
◈Space2030–2035
Mars Sample Return Mission
A Mars sample return mission would collect samples of Martian rock, soil and atmosphere and return them to Earth for analysis in terrestrial laboratories. The NASA Perseverance rover, which landed in Jezero Crater in February 2021, has been collecting and caching samples since shortly after landing. As of 2024, Perseverance has deposited its first sample cache at Three Forks and is continuing to collect additional samples. The original Mars Sample Return campaign, jointly planned by NASA and ESA, envisioned a Sample Retrieval Lander, a Mars Ascent Vehicle and an Earth Return Orbiter returning samples in the early 2030s. In 2024 NASA commissioned a revised, faster and lower-cost architecture in response to cost growth and schedule slippage, with two competing designs being studied by NASA centres and private industry. The European Space Agency is contributing the Earth Return Orbiter and the Sample Transfer Arm. Independent review boards have suggested that samples are unlikely to be returned before 2030, with 2033 being a more realistic target. Scientifically, returned samples could resolve long-standing questions about past habitability of Mars, the history of water, and the possibility of past or present life. The mission also serves as a critical precursor to crewed Mars exploration by validating precision landing, ascent from Mars and large-scale interplanetary mission operations. Consensus schedules now place return in the 2030–2035 window, contingent on the revised architecture.
#Mars#sample return#planetary science
MediumPrototype
▣Robotics2030–2035
Autonomous Construction Sites
Autonomous construction sites use robotic and AI systems to perform a substantial fraction of building work — excavation, bricklaying, concrete pouring, drilling, drywall installation and site survey — with reduced human labour. Several precursors are already in commercial use. Built Robotics offers autonomous excavation and grading systems that retrofit existing construction equipment; Komatsu and Caterpillar sell autonomous mining and quarry machines; Canvas has deployed robotic drywall finishing; Hadrian X (FBR) has demonstrated robotic bricklaying at significantly higher productivity than manual work; and Construction Robotics' SAM100 semi-automated bricklayer is operating on U.S. job sites. Drone-based site survey and progress monitoring from companies such as DroneDeploy, Pix4D and Propeller is now standard. The World Economic Forum's Future of Jobs Report 2025 expects broad deployment of construction robotics through the 2030s, driven by labour shortages in many advanced economies and the need for affordable housing. Key challenges include unstructured site conditions, weather, integration with traditional building trades, safety regulation and the conservative nature of the construction industry. Industry analysts at McKinsey & Company expect that a meaningful fraction of new commercial and residential construction in advanced economies will incorporate significant autonomous robotics by the mid-2030s, with full-site autonomy in selected contexts (such as large repetitive residential projects) by the late 2030s. Adoption in less-developed markets will be slower and is likely to follow different paths, including off-site modular construction.
#construction#automation#robotics
MediumPrototype
▣Robotics2030–2035
Agricultural Robotics (Full Automation)
Agricultural robotics covers autonomous platforms that perform planting, weeding, spraying, monitoring and harvesting in field crops, orchards and protected cultivation. Several categories are already commercial: John Deere's autonomous 8R tractor, Naio Technologies' Oz, Dino and Ted platforms for vegetable weeding, Blue River Technology's See & Spray for selective herbicide application, and Monarch Tractor's electric autonomous platform. Harvest automation is harder but progressing rapidly for high-value crops: Advanced Farm's strawberry-picking robots, Tortuga AgTech's table-grape and strawberry harvesters, and FFRobotics's apple-picking platforms. The Food and Agriculture Organization (FAO) projects global agricultural production must increase significantly by 2050 to feed a population of approximately 10 billion, while labour shortages are acute in many advanced economies: the U.S. H-2A temporary agricultural worker programme has grown more than fourfold in a decade. Industry analysts at IDTechEx and Interact Analysis project that the agricultural robotics market will exceed USD 20 billion by 2035. Key technical challenges include perception in highly variable outdoor conditions, manipulation of delicate crops, integration with existing farm management systems, and the cost of platforms relative to human labour. The consensus expectation is that selective-task agricultural robots (weeding, spraying, monitoring) will be widely deployed in advanced-economy row crops by the early 2030s, with full-autonomy field operations for selected crops — particularly high-value specialty crops — by the mid-2030s. Adoption in less-developed agricultural economies will follow a different trajectory, with smaller-scale platforms and shared-use business models.
#agriculture#agtech#automation
MediumPrototype
⬡Materials2030–2035
Biodegradable Electronics at Scale
Biodegradable electronics are devices that perform electronic functions during a defined service life and then dissolve, degrade or compost without leaving persistent toxic residues. The motivation is the rapidly growing electronic-waste stream: the United Nations' Global E-waste Monitor 2024 reports that the world generated 62 million tonnes of e-waste in 2022, of which less than a quarter was formally recycled. Biodegradable electronics are based on substrates such as cellulose, silk fibroin, poly(lactic acid), poly(vinyl alcohol) and gelatin, combined with conductive inks using carbon, magnesium, zinc, molybdenum or tungsten — all of which are biocompatible in the quantities used. Several research groups, including John Rogers' at Northwestern University, have demonstrated implantable bioelectronic devices that dissolve in the body after a useful life, eliminating the need for surgical removal. Initial commercial applications include medical implants, agricultural sensors, environmental monitors and disposable consumer-electronics components. The Global E-waste Statistics Partnership, UNITAR and the Solving the E-waste Problem (StEP) Initiative have called for broader adoption of design-for-disposal principles. Industry analysts project that commercial-scale biodegradable electronics for single-use applications will become established in the 2030s, particularly in medical devices and agricultural IoT sensors. Wider adoption in consumer electronics will require development of biodegradable batteries, displays and packaging, and is expected to be gradual. Cost remains a barrier relative to conventional silicon electronics, but for applications where retrieval is impractical — environmental sensors, agricultural monitors, medical implants — biodegradable alternatives are competitive on a life-cycle basis.
#biodegradable#electronics#e-waste
MediumPrototype
➤Transportation2030–2035
Electric Aviation for Short-Haul Flights
Electric aircraft — fully battery-electric or hybrid-electric — are emerging as a viable option for short-haul regional routes of up to roughly 500 km. Several programmes have made significant progress: Heart Aerospace's ES-30 (30-seat, 200 km all-electric range with hybrid extension to 400 km) has orders from SAS, Air Canada and United; Eviation's Alice (9-seat all-electric) completed its first flight in 2022; Pipistrel's Velis Electro became the first FAA- and EASA-certified electric aircraft in 2020 and is in service for pilot training; and several startups including Wright Electric, Aura Aero and Bye Aerospace are pursuing adjacent niches. Airbus has delayed its zero-emission aircraft programme but continues hybrid-electric research; ATR and De Havilland are studying electric conversions of existing turboprops. The ICAO Long-Term Aspirational Goal of net-zero CO2 emissions from international aviation by 2050, agreed in 2022, and the EU ReFuelEU Aviation mandate for sustainable aviation fuel are pushing the industry towards electrification for short routes. Key technical challenges include battery energy density — current cells are around 250 Wh/kg, while practical regional electric aviation needs roughly 400–500 Wh/kg — together with certification standards for high-voltage electrical systems and battery fire safety. The consensus view of the International Air Transport Association (IATA) and ICAO is that the first commercial electric passenger services on short routes (up to ~30 seats, ~200 km) will begin in the late 2020s or early 2030s, with broader deployment through the 2030s as battery technology improves. Long-haul electric aviation remains implausible without fundamental breakthroughs in battery chemistry or alternative energy carriers such as hydrogen.
#electric aircraft#aviation#decarbonisation
HighPrototype
➤Transportation2030–2035
eVTOL Air Taxi Services
Electric vertical takeoff and landing (eVTOL) aircraft are battery-electric aircraft that combine the vertical-takeoff capability of a helicopter with the cruise efficiency of a fixed-wing aircraft, intended for short-haul urban air-mobility services. Several companies have made significant progress towards certification: Joby Aviation has been conducting conforming flight tests of its S4 production prototype and has a target to begin commercial service in 2025–2026 (likely later in practice); Archer Aviation's Midnight aircraft is in flight testing and the company has announced planned service in the United Arab Emirates and the U.S. in 2025–2026; Lilium, despite financial difficulties, has continued certification flight testing; Volocopter has been demonstrating in Singapore, Paris and elsewhere; and EHang's EH216-S has received a type certificate from the Civil Aviation Administration of China. The European Union Aviation Safety Agency (EASA) and the FAA have established certification bases for powered-lift aircraft, and several type certificates are expected by 2026–2027. Industry analysts at McKinsey & Company and Morgan Stanley project that the eVTOL market could reach USD 1 trillion by 2040 under an upside scenario, though near-term adoption is likely to be limited by vertiport infrastructure, noise, public acceptance and battery technology. The consensus expectation is that commercial air-taxi services will begin in selected cities in the late 2020s, with broader deployment through the 2030s as costs fall and infrastructure expands. Initial routes are likely to connect airports to city centres and to serve high-income travellers.
#eVTOL#urban air mobility#aviation
MediumPrototype
✦Biotech2030–2040
Gene Drive for Malaria Eradication
A gene drive is a genetic engineering technique that biases inheritance, allowing an engineered trait to spread through a wild population faster than ordinary Mendelian inheritance. The most prominent application is the modification of Anopheles mosquitoes so that they can no longer transmit Plasmodium parasites that cause malaria, or so that female mosquitoes become sterile, collapsing local populations. Target Malaria, a research consortium funded by the Bill & Melinda Gates Foundation and the Wellcome Trust, has conducted contained-field trials in Burkina Faso and is preparing for additional studies. Similar work is being done by the University of California's Malaria Elimination Initiative. The WHO reported roughly 249 million malaria cases and 608,000 deaths in 2022, predominantly in sub-Saharan Africa and among children under five. Gene drives could, in principle, dramatically reduce or locally eliminate malaria transmission, but they raise serious ecological, ethical and governance questions. The Convention on Biological Diversity has been reviewing the technology, and the WHO released a guidance framework for testing genetically modified mosquitoes in 2021. Consensus timelines place the first release trials of gene-drive mosquitoes in the early 2030s, with broader deployment conditional on ecological monitoring, regional consent and the establishment of effective governance mechanisms. Implementation is expected to be uneven across jurisdictions, with significant capacity-building requirements in affected countries.
#gene drive#mosquitoes#public health
MediumResearch
△Energy2030–2040
Next-Generation Enhanced Geothermal Systems
Enhanced Geothermal Systems (EGS) extract heat from hot, dry rock by injecting water at high pressure, fracturing the rock and circulating water through the resulting fracture network to produce steam for power generation. The technology, in principle, makes geothermal energy available almost anywhere on Earth rather than only at naturally occurring hydrothermal sites such as Iceland, Kenya or the U.S. Geysers. Several companies are now demonstrating the technology at commercial scale, including Fervo Energy (Project Red in Nevada, which achieved 3.5 MW of commercial production in 2023), Eavor (closed-loop EGS in Germany) and Quaise Energy (millimetre-wave drilling technology originally developed at MIT). The U.S. Department of Energy's Energy Earthshots initiative targets a 90% cost reduction for EGS by 2035, to a levelised cost of USD 45/MWh. A particularly promising direction is super-hot-rock geothermal, where temperatures above 374°C yield dramatically higher power per well, but require drilling through very hard crystalline basement rock. Consensus projections place the first commercial-scale EGS plants in the late 2020s and early 2030s, with significant growth through the 2030s as drilling and reservoir-engineering costs fall. The technology has the potential to provide baseload renewable power in many regions without storage, complementing variable wind and solar. Induced seismicity, water use and surface footprint remain important regulatory and social-licence considerations.
#geothermal#renewable#drilling
MediumResearch
⬡Materials2030–2040
Carbon Nanotube Mass Production
Carbon nanotubes (CNTs) — rolled sheets of graphene with extraordinary tensile strength, electrical conductivity and thermal properties — have been produced in laboratory quantities since the early 1990s, but their use in bulk applications has been limited by cost, defect density, dispersion challenges and the difficulty of producing long, defect-free tubes at scale. Current global capacity is estimated at several thousand tonnes per year, mostly low-cost multi-walled CNTs used as additives in composites, batteries and plastics. Companies including Nanocyl, OCSiAl (TUBALL), Cabot Corporation, Thomas Swan and Zeon produce commercial grades. The next step is mass production of high-quality single-walled CNTs and CNT-based macroscopic materials, including yarns, fibres and sheets that could displace copper in electrical conductors, replace carbon fibre in structural composites, and enable lightweight electromagnetic shielding. Several announced projects, including Texas-based DexMat's spun CNT yarns and MIT-spinout Vecco's chemical vapour deposition process, are pursuing scale-up. The U.S. Department of Energy ARPA-E and the National Institute of Standards and Technology (NIST) continue to fund related R&D. The consensus expectation among materials-industry analysts is that the first mass-produced CNT conductors and structural reinforcements will reach commercial scale in the 2030s, with broader deployment in the 2040s once manufacturing costs fall below those of competing materials. Applications in aerospace, automotive, energy storage and electronics are expected to lead. Key challenges include quality control at scale, environmental health and safety (including the long-term toxicity of inhaled CNTs), and integration with existing industrial supply chains.
#carbon nanotubes#nanomaterials#manufacturing
MediumResearch
⬡Materials2030–2040
Self-Healing Materials (Commercial Use)
Self-healing materials repair damage — cracks, scratches, micro-fractures — autonomously or with minimal external trigger, extending service life and reducing maintenance cost. The field encompasses several mechanisms: microencapsulated healing agents released on crack propagation (originally developed at the University of Illinois), vascular networks that deliver healing agents (Brendan O'Flynn's group at Cork), reversible covalent chemistries (Diels-Alder, disulphide exchange, hydrogen bonding), shape-memory polymers, and biologically-inspired self-healing concretes using bacterial spores (Hendrik Jonkers' work at Delft University of Technology). Several products have reached commercial use: Profast's self-healing protective coatings, ComposiMold's thermoplastic elastomers, self-healing concrete being trialled by the Dutch Ministry of Infrastructure and the U.K.'s Resilient Materials for Life programme. The European Union's SMARTA and SHeMat projects have advanced the science and demonstration of self-healing materials. Key barriers to broader commercial use include the cost premium over conventional materials, validation of long-term performance, regulatory frameworks (especially for construction materials) and integration with existing manufacturing processes. Industry consensus, as reflected in reports by Lux Research and IDTechEx, is that self-healing materials will become a significant commercial category in the 2030s, initially in high-value applications such as protective coatings, marine paints, aerospace composites and infrastructure (particularly bridges, tunnels and offshore structures). The cumulative effect on infrastructure maintenance budgets, currently estimated at 3–4% of GDP in advanced economies, could be substantial over the long term.
#self-healing#polymers#durability
MediumPrototype
◉Society2030–2040
Universal Digital Identity
Universal digital identity refers to a system under which every individual has a verifiable digital identity that can be used to access government services, financial systems, healthcare, education and other online services. Several large-scale deployments already exist: India's Aadhaar system covers more than 1.4 billion people; Estonia's e-Residency and X-Road infrastructure underpins most government services; Singapore's SingPass provides a single identity for more than 4 million residents; and the European Union is rolling out the European Digital Identity Wallet under the eIDAS 2.0 regulation, with all member states required to offer it by 2026. The World Bank's ID4D programme reports that roughly 850 million people still lack any official identity document, mostly in sub-Saharan Africa and South Asia. The principal benefits of universal digital identity include access to financial services, government services and social protection programmes; reduced fraud and corruption; and more efficient administration. The principal risks include surveillance, exclusion of vulnerable groups, data breaches and the concentration of state power. The Open Identity Foundation, the World Wide Web Consortium's Decentralized Identifiers Working Group, and the Linux Foundation's OpenWallet Foundation are developing technical standards for interoperable, privacy-preserving identity systems. The consensus expectation is that the proportion of the global population with a verifiable digital identity will rise significantly through the 2030s, although the design of those systems — centralised, federated or decentralised; state-run, private or hybrid — will vary substantially across jurisdictions. By the mid-2030s the consensus view is that digital identity will be the universal default for accessing essential services in most countries.
#digital identity#governance#privacy
MediumResearch
◉Society2030–2040
Universal Basic Income Pilot Programmes
Universal basic income (UBI) is a regular, unconditional cash payment made to all members of a population, independent of income, employment or family status. Although full UBI has not been implemented at scale, many pilot programmes have been conducted, including Finland's two-year trial (2017–2018) with 2,000 unemployed citizens, Kenya's GiveDirectly trial covering 20,000 people across 200 villages for 12 years, Stockton, California's SEED programme, the United States' Mayors for a Guaranteed Income network of pilots, and South Korea's basic-income experiments in Gyeonggi Province. The OECD has analysed the fiscal and distributional implications of hypothetical UBI schemes in several countries. The World Bank and the ILO have published reviews of pilot evidence, generally concluding that unconditional cash transfers have positive effects on food security, mental health, school attendance and small-business formation, with limited effects on labour supply. The principal motivations for renewed interest in UBI include the possible future automation of substantial categories of routine work, the inadequacy of means-tested safety nets in many countries, and the perceived complexity of existing welfare systems. Counter-arguments include cost, the risk of crowding out targeted support for the most disadvantaged, the political sustainability of universal programmes, and the difficulty of financing at scale. The consensus expectation among policy researchers at the OECD, the World Bank and the ILO is that pilot programmes will continue and expand through the 2030s, with the most serious consideration of national-scale implementation likely in countries facing significant labour-market disruption from automation. No advanced economy is expected to implement full national UBI before the late 2030s at the earliest.
#UBI#social policy#welfare
MediumResearch
◈Space2035–2040
Permanent Lunar Base
A permanent crewed lunar base would maintain a continuous human presence on the Moon, supporting scientific research, technology demonstration and the development of in-situ resource utilisation (ISRU) for water, oxygen and construction materials. The current international framework is the U.S.-led Artemis programme, which launched its first uncrewed test flight in 2022 and aims to land the first crewed mission of the programme (Artemis III) on the lunar south pole in 2026 or 2027. NASA's stated plan is to establish an Artemis Base Camp by the end of the 2030s, comprising a lunar foundation surface habitat, a habitable mobility platform and a power system. The Lunar Gateway, a small space station in lunar orbit, is being assembled from 2025 onwards with international contributions from ESA, JAXA and CSA. China and Russia are pursuing the parallel International Lunar Research Station (ILRS), with planned robotic precursor missions throughout the late 2020s and a crewed landing goal in the 2030s. The European Space Agency's Terrae Novae strategy supports a sustained European presence on the Moon. Key technical challenges include dust mitigation, surface power (likely small nuclear fission reactors such as NASA's Kilopower project), radiation protection, ISRU for water extraction from permanently-shadowed polar craters, and sustainable logistics. The consensus view across NASA, ESA, JAXA, CSA and equivalent bodies is that a continuously inhabited lunar outpost is plausible by the late 2030s, with growing permanence into the 2040s.
#Moon#exploration#habitation
MediumResearch
◈Space2035–2040
Crewed Mars Landing
A crewed mission to land on the surface of Mars and return safely to Earth is a long-standing objective of space exploration. The technical challenges are substantially greater than for the Moon: Mars is roughly 200 times farther from Earth than the Moon at closest approach, requiring transit times of six to nine months each way, life-support systems capable of operating autonomously for years, protection from cosmic radiation during transit and on the surface, precision landing of large payloads, ascent from the Martian surface, and either in-situ propellant production or pre-positioned return vehicles. SpaceX's fully reusable Starship-Super Heavy system is the most ambitious private-sector approach, with the stated goal of crewed Mars missions by the late 2020s, although most analysts consider a 2035–2040 window more realistic given the complexity of life-support and ISRU systems. NASA's Moon-to-Mars architecture uses Artemis lunar missions as preparation for eventual crewed Mars missions in the late 2030s or 2040s. China has announced plans for crewed Mars landings before 2033, with follow-up missions in 2035, 2037 and 2041. The consensus view of the international space community, expressed through bodies such as the International Space Exploration Coordination Group (ISECG), is that a first crewed Mars landing is plausible in the late 2030s, although schedule risk is high and the actual date could slip into the 2040s.
#Mars#deep space#exploration
MediumPrototype
◇AI2035–2045
Brain-Computer Interfaces for Consumer Use
Brain-computer interfaces (BCIs) record neural activity and translate it into digital commands, enabling direct communication between the brain and external devices. Clinical BCIs have existed for two decades in the form of Utah arrays used by patients with severe paralysis, but recent breakthroughs have shifted the field. Neuralink, Synchron, Paradromics and Blackrock Neurotech have all implanted devices in human patients as part of FDA-approved clinical trials. Synchron's Stentrode is inserted through the vasculature, avoiding open-brain surgery, while Neuralink's threads are placed by a surgical robot. Non-invasive approaches using EEG, fNIRS and focused ultrasound are also progressing. Consensus projections suggest that the first commercial BCIs aimed at patients with severe neurological conditions will reach the market in the early 2030s. Consumer-grade BCIs for able-bodied users — for gaming, productivity or accessibility — are expected later in the decade or in the 2040s, once implantable devices become safer, smaller and more capable. Significant barriers remain, including long-term biocompatibility, signal stability, surgical risk, regulatory frameworks, data privacy and ethical concerns about cognitive surveillance. The FDA has issued initial guidance on implantable BCIs and is expected to refine it as more trial data becomes available. By the mid-2030s the field is likely to be roughly where gene therapy was in the late 2010s: clinically validated for narrow indications, with broad consumer use still several years out.
#BCI#neural interfaces#Neuralink
MediumPrototype
◇AI2035–2045
Autonomous AI Scientists
Autonomous AI scientists are systems capable of formulating hypotheses, designing and running experiments, analysing results and writing papers with limited human supervision. Early precursors already exist. Sakana AI's 'AI Scientist' generates machine-learning ideas and writes them up, while systems such as Coscientist (CMU) and A-Lab (Berkeley) autonomously design and run chemistry experiments. In structural biology, AlphaFold and follow-on systems have effectively automated protein-structure prediction. However, fully autonomous discovery across a range of scientific disciplines requires the integration of multiple capabilities — literature comprehension, hypothesis generation, experimental design, robotic execution and statistical inference — and the ability to interface with laboratory equipment. Consensus timelines place broad autonomous AI scientists in the late 2030s to 2040s, with the precise window depending on progress in agentic planning, physical lab automation and validation standards. There are also serious institutional and incentive questions: who reviews AI-generated papers, how are findings reproduced, and how should scientific credit be allocated? Bodies such as the OECD, the Royal Society and the U.S. National Academies have begun discussing these issues. If realised, autonomous AI scientists could dramatically accelerate discovery in materials science, drug development and clean-energy research, with downstream effects on industrial productivity and global competitiveness.
#AI research#automated science#discovery
LowResearch
✦Biotech2035–2045
Patient-Specific Organ Bioprinting
Bioprinting uses 3D-printing techniques to deposit living cells, biomaterials and growth factors layer by layer, with the goal of producing functional tissues and eventually whole organs for transplantation. The need is enormous: the Global Observatory on Donation and Transplantation reports roughly 150,000 transplants per year against a waiting list of well over a million patients. Current progress includes 3D-bioprinted skin and cartilage that have reached early clinical use, vascularised tissue patches, and miniature organoids used in drug testing. Companies such as Organovo, CELLINK, 3D Systems and several academic centres are working on more complex structures. The principal scientific challenges are vascularisation (printing a network of blood vessels capable of sustaining larger tissues), innervation and long-term function. The Wyss Institute at Harvard has demonstrated thick vascularised tissue using SWIFT (Sacrificial Writing into Functional Tissue), and several groups have printed simplified kidneys and hearts that survive briefly in vitro. Industry consensus places patient-specific bioprinted simple organs, such as patches of cardiac tissue or sections of bladder and trachea, in the late 2030s, with fully vascularised solid organs (kidneys, livers, hearts) expected in the 2040s, contingent on progress in vascularisation and immune-matching. The combination with patient-derived iPSCs could eliminate the need for immunosuppression. Regulatory frameworks for bioprinted organs are still being developed by the FDA and EMA.
#bioprinting#transplant#tissue engineering
MediumResearch
✦Biotech2035–2045
Synthetic Human Embryos from Stem Cells
Synthetic embryos, also known as embryo models or embryoids, are structures grown from stem cells that mimic key features of natural embryos without involving sperm, eggs or fertilisation. Researchers at the Weizmann Institute (Jacob Hanna's lab), the University of Cambridge (Magdalena Zernicka-Goetz) and others have produced mouse embryo models that develop through early organogenesis, and have extended the work to human stem-cell-derived embryo models that recapitulate aspects of gastrulation and early organ formation. The International Society for Stem Cell Research (ISSCR) revised its guidelines in 2021 to permit culture of human embryo models beyond the 14-day limit that applies to natural embryos, subject to oversight. The long-term goal is to provide an ethically more acceptable platform for studying human development, drug teratogenicity, and the causes of miscarriage and congenital disease. Industry consensus is that clinically useful applications — for instance, generating patient-specific tissues for transplantation, or modelling disease — are roughly a decade away. Direct clinical use of synthetic embryos to establish pregnancies is widely considered both technically remote and ethically impermissible under current guidelines. By the mid-2030s the consensus expectation is that human embryo models will be a routine laboratory tool, with applications in drug screening and developmental biology, but with substantial regulatory and ethical infrastructure still being built. The ISSCR, the U.S. National Academies and equivalent bodies are expected to refine governance in parallel with the science.
#stem cells#embryo models#developmental biology
LowResearch
△Energy2035–2045
Green Hydrogen Economy for Heavy Industry
Green hydrogen — produced by electrolysis of water using renewable electricity — is widely regarded as the most promising decarbonisation route for sectors that are difficult to electrify directly, including steel, cement, ammonia, shipping and long-duration aviation. The International Energy Agency's Global Hydrogen Review 2024 reports that the project pipeline for low-emission hydrogen has grown to roughly 38 Mt/year by 2030, although only a small fraction has reached final investment decision. Announced steel projects such as H2 Green Steel in Sweden, ThyssenKrupp's tkH2Steel programme and ArcelorMittal's Hamburg plant are pioneering hydrogen-based direct-reduced-iron processes. The European Hydrogen Bank and the U.S. Inflation Reduction Act's Section 45V production tax credit are subsidising the cost gap with conventional hydrogen, which currently stands at roughly USD 2–4/kg. The IEA's Net Zero Emissions by 2050 scenario requires roughly 150 Mt/year of low-emission hydrogen by 2030 and roughly 430 Mt/year by 2050. Consensus projections place cost parity with grey hydrogen in regions with abundant renewable resources, such as the Middle East, Australia, Chile and parts of Africa, by the early to mid-2030s, with broader industrial deployment through the late 2030s and 2040s. Significant barriers remain, including electrolyser manufacturing scale, transport and storage infrastructure, certification of green hydrogen, and the demand-pull needed to underwrite large-scale projects.
#hydrogen#decarbonisation#industry
MediumPrototype
△Energy2035–2045
Space-Based Solar Power Demonstration
Space-based solar power (SBSP) would collect solar energy in geostationary orbit, where panels receive continuous sunlight unaffected by weather or night, and beam the energy to Earth as microwaves or lasers. The concept was first proposed by Peter Glaser in 1968 and has been periodically reassessed by NASA, the U.S. Department of Energy, ESA, JAXA and China's space administration. Recent technology demonstrations include the Caltech Space Solar Power Project, which in 2023 demonstrated wireless power transfer in orbit, and the U.K. Space Energy Initiative's SOLARIS programme, which is conducting feasibility studies. ESA estimates that an operational demonstration in the 2030s and commercial plants in the 2040s are technically feasible, with levelised costs initially much higher than terrestrial renewables. China has announced plans for a megawatt-scale demonstrator in low Earth orbit by the early 2030s, scaling to a gigawatt-class commercial plant in geostationary orbit by 2050. The principal challenges are launch costs (which Starship-class vehicles could substantially reduce), robotic assembly of kilometre-scale structures in orbit, thermal management, and ground-rectenna infrastructure. Most analysts regard SBSP as a long-term option for stable, dispatchable clean power that complements intermittent terrestrial renewables, rather than a near-term climate solution. By the mid-2030s a credible technology demonstrator in orbit is plausible; commercial deployment is more likely in the 2040s contingent on launch-cost reductions.
#SBSP#satellites#renewable
LowResearch
◈Space2035–2045
Asteroid Mining Demonstration
Asteroid mining refers to the extraction of water, metals and other materials from near-Earth asteroids and eventually main-belt asteroids. Water extracted from carbonaceous asteroids could be split into hydrogen and oxygen to refuel spacecraft in orbit, dramatically reducing the mass that must be launched from Earth. Platinum-group metals from metallic asteroids are a longer-term resource for terrestrial industry. NASA's OSIRIS-REx and JAXA's Hayabusa2 missions have successfully returned samples from asteroids Bennu and Ryugu, validating the navigation, sampling and return technologies. NASA's Psyche mission, launched in 2023, will study the metal-rich asteroid 16 Psyche and provide the first close-up data on a metallic body. Several private companies — including TransAstra, AstroForge and Karman+ — are pursuing commercial asteroid mining, with AstroForge launching its first demonstration mission in 2023. The Luxembourg Space Agency, through its SpaceResources.lu initiative, has built a legal and regulatory framework for space-resource utilisation. The U.S. Commercial Space Launch Competitiveness Act of 2015 recognises the right of U.S. citizens to own, transport, use and sell asteroid resources. Consensus projections place the first credible demonstration of asteroid resource extraction (likely water extraction and propellant production) in the late 2030s or 2040s, with commercial-scale operations later in the century. Key challenges include automation, propulsion for deep-space missions, regulatory frameworks, and the uncertain economics of returning materials to Earth.
#asteroid mining#ISRU#deep space
LowResearch
▣Robotics2035–2045
General-Purpose Household Robots
A general-purpose household robot would be capable of performing most routine domestic chores — folding laundry, loading dishwashers, tidying rooms, food preparation, light maintenance — in unmodified home environments. The challenge is enormous: homes are cluttered, unstructured and vary widely across cultures, and the required manipulation, perception and planning are beyond today's specialised robots. Recent advances have rekindled optimism. Tesla's Optimus, Figure's Figure 02, Agility Robotics' Digit, 1X Technologies' NEO, Sanctuary AI's Phoenix, Apptronik's Apollo, and Boston Dynamics' Atlas (electric version) are all-purpose humanoid platforms under active development, with pilot deployments in warehouses and factories beginning in 2024–2025. The path to homes, however, is widely expected to take a decade or more. Key technical challenges include robust manipulation in clutter, robust grasping of soft and deformable objects, human-robot interaction safety, battery life, and end-effectors that combine the dexterity of a human hand with the durability of industrial grippers. Cost is also a major barrier: current humanoid platforms cost USD 30,000–200,000, well above consumer price points. Industry analysts at ABI Research and Interact Analysis project that the first credible consumer-grade household robots will appear in the late 2030s, with broader adoption in the 2040s. The deployment will likely begin with constrained tasks — such as tidying or single-room cleaning — and expand as capabilities and consumer trust grow. Privacy and safety regulation will shape deployment.
#humanoid robots#domestic#AI
LowResearch
➤Transportation2035–2045
Hyperloop Operational
Hyperloop refers to a proposed mode of passenger and freight transport in which pods travel through low-pressure tubes at aircraft-equivalent speeds — conceptually between 500 and 1,200 km/h. The concept was popularised by Elon Musk's 2013 white paper, but commercial development has been carried out by independent companies including Virgin Hyperloop (which demonstrated passenger transport in 2020 but subsequently pivoted to freight), Hyperloop Transportation Technologies, Hardt Hyperloop, Zeleros, and the Indian government-backed IIT Madras project. Several feasibility studies have been completed for routes including Mumbai-Pune, Chicago-Columbus-Pittsburgh, the UAE, and the Pune-Mumbai Expressway. The European Hyperloop Centre in Groningen opened a 420-metre test track in 2024. China Aerospace Science and Industry Corporation is developing a maglev test vehicle targeting 1,000 km/h. The principal engineering challenges include maintaining near-vacuum conditions in tens or hundreds of kilometres of tube, managing thermal expansion, ensuring pod safety during emergency deceleration, and the cost of land acquisition and infrastructure. Industry analysts at McKinsey and the Railway Industry Association consider that the first commercial passenger route is unlikely before the mid-2030s, with significant risk of further delay. Several prominent venture-backed startups, including Virgin Hyperloop One (now dissolved) and Hyperloop Transportation Technologies, have faced financial challenges. The consensus view among transport infrastructure analysts is that hyperloop is more likely to be a niche high-speed connection between specific city pairs than a general replacement for high-speed rail, and that commercial operation in the late 2030s is the most plausible near-term scenario.
#hyperloop#high-speed rail#transportation
LowResearch
◉Society2035–2045
Carbon Removal at Scale (1 Gt/year)
Carbon dioxide removal (CDR) at the gigatonne-per-year scale — removing 1 billion tonnes or more of CO2 from the atmosphere annually — is widely regarded as essential to limiting warming to 1.5°C or 2°C, complementing aggressive emissions reductions rather than substituting for them. The IPCC Sixth Assessment Report (AR6, 2023) concludes that most pathways consistent with limiting warming to 1.5°C require substantial CDR, with deployment scaling to several GtCO2/year by mid-century. CDR approaches include afforestation and reforestation, soil-carbon sequestration, bioenergy with carbon capture and storage (BECCS), direct air carbon capture and storage (DACCS), enhanced rock weathering, ocean alkalinity enhancement and blue-carbon management. Climeworks' Orca and Mammoth DAC plants in Iceland, Carbon Engineering's STRATOS plant under construction in Texas (designed for 500,000 t/year when fully operational) and Charm Industrial's bio-oil sequestration represent the first commercial-scale operations. The State of Carbon Dioxide Removal report (2024) estimates current global CDR at roughly 2 Gt/year, almost entirely from conventional afforestation. The consensus expectation among the IEA, the IPCC and the International Energy Agency is that achieving 1 Gt/year of durable, engineered removals will require significant policy support, carbon pricing or procurement, and is plausible by the mid-to-late 2030s in the most ambitious scenarios. The U.S. Department of Energy's Carbon Negative Shot targets USD 100/t CO2 for durable removals by 2032. Key challenges include energy requirements for DAC, land competition for BECCS, monitoring and verification, and the cost of durable versus nature-based removals.
#carbon removal#climate#CDR
MediumPrototype
◉Society2035–2045
Aging Population Care Automation
Aging population care automation refers to the use of robotic, AI and monitoring technologies to support the care of older adults at home and in residential facilities, addressing labour shortages in the care workforce and enabling older people to live independently for longer. The demographic context is significant: the United Nations' World Population Prospects 2024 projects that the global population aged 65 and over will rise from roughly 800 million in 2024 to 1.6 billion by 2050. In Japan, more than 29% of the population is already over 65; in many European countries the figure is above 20%. Several technologies are being deployed or piloted: assistive exoskeletons (Cyberdyne's HAL, Atoun Model Y) for carers lifting residents; socially assistive robots such as PARO the therapeutic seal and SoftBank's Pepper in care homes; monitoring systems using non-intrusive sensors and AI to detect falls and changes in routine; companion robots (ElliQ from Intuition Robotics); and AI-driven conversational assistants for cognitive engagement. The WHO's Global Strategy on Digital Health 2020–2025 and the OECD's Health at a Glance reports both identify ageing-related care technology as a priority. The principal challenges include the cost of devices, integration with care workflows, ethical issues around consent and surveillance, and the difficulty of replicating human empathy. Industry analysts at ABI Research and Japan's Ministry of Economy, Trade and Industry project a meaningful deployment of eldercare automation in advanced economies from the mid-2030s, scaling through the 2040s as the demographic imperative intensifies. Cultural acceptance varies substantially across societies.
#eldercare#demographics#robotics
MediumResearch
△Energy2040–2045
First Commercial Fusion Power Plant
A commercial fusion power plant would generate electricity by fusing light nuclei — typically deuterium and tritium — and converting the released energy into grid power. The field was transformed in December 2022 when the U.S. National Ignition Facility achieved scientific breakeven, producing more energy from fusion than the laser energy delivered to the target. Public-sector projects such as ITER (France), the U.K. STEP programme, China's CFETR and South Korea's K-DEMO continue to advance magnetic-confinement designs. Private firms including Commonwealth Fusion Systems, Helion Energy, TAE Technologies and Tokamak Energy have raised billions of dollars and are building demonstration devices. The U.S. Department of Energy's Bold Decadal Vision for Commercial Fusion Energy targets a demonstration plant in the 2030s, with commercial deployment in the 2040s. The Fusion Industry Association's 2024 report estimates cumulative private investment of more than USD 7 billion. Key remaining challenges include sustained plasma confinement at high gain, materials that survive high neutron flux, tritium breeding, remote maintenance, and the capital cost of first-of-a-kind plants. The consensus view of the IEA and the U.S. DOE is that the first grid-connected commercial fusion plant is plausible by the early 2040s, with broader deployment through the late 2040s and 2050s. Fusion is widely regarded as a long-term complement to renewables, fission and storage rather than a near-term climate solution, given the slow pace of building first-of-a-kind capital projects.
#fusion#clean energy#power generation
MediumPrototype
◇AI2040–2050
Artificial General Intelligence
Artificial General Intelligence (AGI) refers to systems capable of performing most economically valuable cognitive work at or above the typical human level across a wide range of domains. Unlike today's narrow models, an AGI would transfer knowledge between tasks, plan over long horizons, and adapt to novel situations without per-task retraining. Consensus projections from major AI laboratories and policy bodies suggest that scaling current architectures, combined with reinforcement learning from verifiable rewards, may produce systems that approach general competence sometime in the 2040s. However, no lab has yet demonstrated reliable out-of-distribution generalisation at human level, and several unsolved problems remain, including long-horizon planning, sample-efficient learning, and persistent value alignment. The timing estimate assumes continued growth in compute budgets, breakthroughs in memory architectures, and incremental progress on alignment. Many researchers regard the 2040–2050 window as plausible but not certain; a minority argue for shorter timelines, while others expect AGI may take substantially longer if current scaling laws plateau. The societal implications, including labour-market disruption, security risks and concentration of power, are the subject of active work by the OECD AI Policy Observatory, the UK AI Safety Institute and the US AI Safety Institute. The next decade's milestones will likely include autonomous AI scientists, agentic systems that complete multi-day projects, and standardised evaluation suites for general competence.
#AGI#machine learning#automation
MediumResearch
✦Biotech2040–2050
Reversing Aging in Humans (Early Trials)
Therapies that aim to slow, halt or partially reverse aspects of human aging are a small but active field of biomedical research. The biology is grounded in the 'hallmarks of aging' framework — including cellular senescence, epigenetic drift, mitochondrial dysfunction, telomere attrition and loss of proteostasis. Several interventions are in human clinical trials, including senolytics (drugs that selectively kill senescent cells), rapamycin analogues, NAD+ precursors, and partial epigenetic reprogramming using Yamanaka factors. Animal studies, including recent work in mice showing extended lifespan and partial rejuvenation, have driven substantial private investment in companies such as Altos Labs, Calico, Unity Biotechnology and Loyal. The American Federation for Aging Research and the FDA have begun discussing regulatory frameworks for 'aging-related' indications, and the first generation of approved drugs targeting specific aging hallmarks is plausible within the 2030s. However, translating this into measurable 'reversal' of aging in humans is far harder: clinical endpoints such as all-cause mortality are slow to measure, and human biology is more complex than the mouse models that motivate most interventions. Consensus projections place meaningful, validated human trials of multi-hallmark 'rejuvenation' therapies in the 2040s, with broader clinical use conditional on those trials succeeding. The societal implications — for pensions, healthcare, work and inequality — are likely to be substantial even if the therapies extend healthspan rather than maximum lifespan.
#longevity#senolytics#epigenetic reprogramming
LowResearch
➤Transportation2040–2050
Hydrogen-Powered Long-Haul Aviation
Hydrogen-powered aviation — using either hydrogen fuel cells for electric propulsion or direct combustion of hydrogen in modified gas turbines — is regarded by most analysts as the most credible long-term path for decarbonising medium and long-haul flight, where battery-electric propulsion is implausible due to the specific energy limits of batteries. Airbus's ZEROe programme has committed to introducing the world's first hydrogen-powered commercial aircraft by 2035, with three concept aircraft (a turbofan, a turboprop and a blended-wing-body) under study. Rolls-Royce and easyJet have demonstrated a hydrogen-fuelled AE2100 turboprop engine on the ground; CFM International and Airbus are pursuing a hydrogen-combustion A380 flying testbed by the mid-2020s. The U.K. Jet Zero Council, the EU Clean Aviation partnership and NASA have all funded hydrogen-aviation R&D. Significant challenges remain: hydrogen's volumetric energy density is much lower than kerosene, requiring new aircraft architectures such as blended-wing bodies or cryogenic tank integration; airport hydrogen infrastructure would need to be built at scale; and hydrogen production itself must be low-carbon (green hydrogen) to deliver emissions benefits. The Royal Society's 2023 report on synthetic aviation fuels and hydrogen concluded that hydrogen is unlikely to make a major contribution to aviation emissions before 2040, with broader deployment in the 2040s and beyond. The consensus view among industry analysts at IATA and ICAO is that hydrogen will play a meaningful role in commercial aviation from the late 2030s onwards, complementing sustainable aviation fuels in the near term and gradually expanding to longer ranges as technology matures.
#hydrogen aviation#decarbonisation#long-haul
LowResearch
⬡Materials2040–2060
Room-Temperature Superconductor
A room-temperature superconductor — a material that exhibits zero electrical resistance at or near ambient temperatures — has been a goal of condensed-matter physics for more than a century. Such a material would transform power transmission, magnetic energy storage, electric motors, maglev transport and high-field magnets for fusion and particle accelerators. The current record, established by the University of Rochester group (Ranga Dias) in 2023, claimed superconductivity at 294 K (21°C) in a lutetium-hydride system, but the result was retracted by Nature in late 2023 amid data-integrity concerns, illustrating the difficulty of validating claims in this field. Other research groups, including Hyowon Hyun's group at Japan's National Institute for Materials Science and the Max Planck Institute for Chemistry, are pursuing hydride systems under high pressure, including sulphur-hydride and yttrium-hydride compounds. The 2023 announcement of LK-99 by a Korean group generated enormous public interest but was subsequently not reproduced. The U.S. Department of Energy's Basic Energy Sciences programme continues to fund superconductivity research, and the Condensed Matter Theory programme supports computational searches for new candidate materials. The consensus view of the international materials-science community, expressed through the American Physical Society and the Materials Research Society, is that a practical, ambient-pressure, room-temperature superconductor remains a research goal, unlikely to be realised before the 2040s and possibly later. Even if a candidate is identified, the path to commercial application — including manufacturability, cost, mechanical properties and stability — typically requires one or two further decades.
#superconductors#physics#materials science
LowResearch
⬡Materials2050–2070
Programmable Matter (Early Demonstrations)
Programmable matter refers to a substance composed of many tiny, networked units (sometimes called 'catoms' or 'claytronic atoms') that can rearrange themselves into arbitrary 3D shapes on command. A piece of programmable matter could be a chair one moment, a table the next, then dissolve back into a heap of raw material to be reused. The concept was articulated by Toffoli and Margolus in 1991 and developed by Carnegie Mellon's Claytronics project (Seth Goldstein and Todd Mowry) in the mid-2000s. Recent advances in micro-robotics, including the demonstration of 1-million-strong swarms of microbots at MIT (Marina Waltz, 2024) and centimetre-scale modular robots at Penn State and EPFL, are gradually approaching the physical substrates needed for programmable matter at meaningful scales. The most ambitious early work, such as the MIT Programmable Matter Project funded by DARPA, explored applications in shape-shifting antennas and reconfigurable structures. Consensus in the materials science and micro-robotics communities, expressed through the European Future and Emerging Technologies (FET) programmes and the U.S. National Science Foundation, is that meaningful demonstrations of programmable matter at the scale of small everyday objects are unlikely before 2050, and broader commercial applications are more plausibly mid-century. The principal scientific challenges include micro-actuation, power distribution across millions of micro-units, inter-unit communication, and computational planning for shape reconfiguration. Ethical and security implications are also non-trivial and are likely to require new governance frameworks well in advance of deployment.
#programmable matter#claytronics#robotics
LowResearch