Olympia's AI decade, from promise to infrastructure
We attended the twelfth London Tech Week[12]Link to footnote[49]Link to footnote at Olympia London[5]Link to footnote from 8–12 June 2026, 30,000+ attendees, including 18,750 enterprise executives, 8,250 startup founders, and 1,500 venture capitalists, framed around "Europe's Decisive Decade".[3]Link to footnote Prime Minister Keir Starmer[22]Link to footnote announced a £400 million compute strategy; Technology Secretary Liz Kendall outlined a £1.[2]Link to footnote1 billion AI Hardware Plan.[4]Link to footnote
From the standpoint of the Impact Intelligence Lab, reading London’s AI arena for implications that travel to ASEAN, Africa, and other Global South contexts, LTW 2026 was less a product fair than a stress test of whether AI can become durable system-level infrastructure without reproducing extractive capital flows, voluntary safety theatre, and Northern-centric design defaults.[61]Link to footnote[78]Link to footnote Across sessions, AI was framed as a once-in-a-century general-purpose technology already reshaping workflows, sovereign compute, discovery labs, and industrial decarbonisation, yet panelists repeatedly returned to trust, inclusion, liability, and value retention as the variables that decide societal impact.[62]Link to footnote[69]Link to footnote
AI is moving from promise to system-level infrastructure that touches compute, sovereignty, labour markets, climate action, scientific discovery, and governance, but the ecosystem’s ability to retain value, build trust, and include diverse leadership will determine whether that promise translates into durable impact.
This briefing audits the floor through three lenses: AI as critical infrastructure, AI as socio-technical force, and AI as climate and bioscience accelerator, with session voices, Global South equity tests, and research-backed constraints on the claims that travelled loudest at Olympia.[9]Link to footnote[19]Link to footnote[35]Link to footnote[78]Link to footnote
London Tech Week 2026 attendance composition
On AI is the Computer, Aravind Srinivas (Co-Founder & CEO, Perplexity) argued that trustworthy “answer engines” must prioritise accuracy over engagement hacks, a product ethics claim that travels far beyond consumer search.[89]Link to footnote
Perplexity prioritised accuracy over entertainment, rejecting pressure to introduce hallucinations for user engagement, turning the computer into a “Ferrari for the mind.”
, Aravind Srinivas, Perplexity, AI is the Computer[89]Link to footnote
Session map: AI infrastructure, labour, and responsibility[57]Link to footnote
London Tech Week 2026 AI angle
| Three cross-cutting themes | |||
|---|---|---|---|
| Critical infrastructureCompute, capital, sovereignty | Socio-technical forceLabour, inclusion, safety | Climate & discoveryLabs, net-zero, dual-use | |
| Floor signalsSessions that set the AI agenda | • £1.1B Hardware Plan / AMD–Nebius • Tech Nation value-capture gap • Sovereign AI & Fifth Domain | • Copilots & future of work • Safe & Ethical AI (Ada/Apollo) • Women & responsible AI • Thrive / Responsible Innovation | • Intelligent Lab / EDEN • Greener Future dual-use • Geospatial AI & space infra • AI for Global Impact |
Sovereign compute and hyperscaler geopolitics
The core question in the AI Arena was: How can European nations build authentic technological and digital sovereignty while remaining structurally dependent on foreign hardware supply chains and hyperscaler cloud platforms?[3]Link to footnote[8]Link to footnote[9]Link to footnote
During the opening sessions, Prime Minister Keir Starmer and Technology Secretary Liz Kendall emphasized a state-backed approach to local compute capacity. The UK government announced a £400 million investment to buy specialized AI compute capacity directly for early-stage startups.[10]Link to footnote This sat alongside the state's £1.1 billion AI Hardware Plan, which allocated £750 million to a new national supercomputer due in 2030, alongside £120 million for a hardware innovation program and £150 million for next-generation inference chips. This package is designed to prevent domestic hardware pioneers from being acquired by foreign buyers. However, contributions from industry representatives revealed a persistent tension. AMD’s CEO, Dr. Lisa Su, announced a £2 billion five-year commitment to expand high-performance computing clusters in collaboration with Cambridge and Imperial College London. Concurrently, cloud provider Nebius pledged £1.7 billion to build regional AI capacity using Nvidia GPUs, aiming to scale its data center infrastructure to 65 megawatts by 2027.
On the Building the Compute Foundation for the AI Era stage, Carolyn Dawson OBE (Founders Forum Group) and Dr. Lisa Su (AMD) framed compute access as an equity problem, not only a competitiveness race.[79]Link to footnote
Areas with limited technological infrastructure were labelled “tech deserts”, bridging that gap is vital for equitable AI-driven growth.
, Carolyn Dawson OBE / Dr. Lisa Su, Building the Compute Foundation for the AI Era[79]Link to footnote
On The Emerging Case for Sovereign AI Development in Europe, Kanishka Narayan MP (Minister for AI and Online Safety), George Osborne (OpenAI for Countries), Judith Dada (Visionaries), and Matt Harris (HPE) argued that sovereignty is about democratic leverage, not stack isolationism.[80]Link to footnote
AI sovereignty matters for national security, economic advancement, and societal dignity. Sovereignty does not mean building every part of the stack domestically, it means strategic leverage over critical bottlenecks.
, The Emerging Case for Sovereign AI Development in Europe[80]Link to footnote
Capital commitments around London Tech Week 2026
These announcements illustrate that much of the region's compute expansion remains heavily reliant on US-designed silicon and infrastructure. While government representatives discussed local model building, enterprise leaders like Alison Kay, VP and Managing Director of AWS UK and Ireland, pointed out that the immediate barrier to sovereign adoption is not model availability, but a shortage of advanced operational skills. Only 24% of domestic organizations are deploying advanced AI systems, meaning standard corporate infrastructure remains tied to legacy US cloud environments.
Equal access to compute now determines which actors build frontier systems and which are left behind. Empirical mapping of 4,392 UK AI entities finds 41.3% concentrated in Greater London and over 55% in the London–Oxford–Cambridge "Golden Triangle," leaving regional ecosystems dependent on central infrastructure.[63]Link to footnote Geographic residency of workloads also does not equal stack independence: Nvidia and AMD silicon still sit underneath most "sovereign" capacity announced at Olympia, a sovereignty paradox that LTW's hardware rhetoric only partially confronted.[61]Link to footnote[63]Link to footnote
Material intelligence and molecular modeling for carbon solutions
On the Deep Tech Stage, the primary question was: How can generative AI and physics-based simulations accelerate the discovery of physical materials to solve global carbon-abatement and environmental-cleanup challenges?[25]Link to footnote[26]Link to footnote[27]Link to footnote
In a session titled Intelligent Matter and the Materials of Tomorrow, Dr. Chad Edwards, CEO of Cambridge-based CuspAI,[28]Link to footnote[31]Link to footnote presented a significant advance in computational materials design. CuspAI operates as a "molecular search engine," allowing users to input targeted physical properties, such as selective carbon dioxide adsorption or high thermal tolerance, and utilizing generative AI to design synthesizable molecular structures on demand.[7]Link to footnote The technical focus of CuspAI centers on the design of Metal-Organic Frameworks (MOFs) for Direct Air Capture (DAC) and Point-Source Carbon Capture. By combining generative model architectures with molecular dynamics simulations, the venture has achieved key technical milestones:[32]Link to footnote
- VUN (Valid, Unique, Novel) Modeling: CuspAI’s proprietary generative model, MOFGEN, achieves a 49% VUN generation rate for MOFs, significantly outperforming comparable models developed by Microsoft (10%) and Meta (16%).
- Discovery Timeline Compression: In partnership with industrial chemistry groups, the venture's "SkyVault" project compressed the end-to-end cycle of carbon-capture material design, physical synthesis, and laboratory validation down to six months.
- Economic Viability: By optimizing MOF pore geometry specifically to bind carbon dioxide selectively under varying environmental conditions, the platform aims to reduce the baseline cost of Direct Air Capture down to a commercial threshold of $100 per ton.
Traditional Computational Discovery
CuspAI "Synthesis-Aware" Pipeline
The discussions revealed consensus on the potential of AI to accelerate molecular discovery, but highlighted a clear bottleneck in physical synthesis.[30]Link to footnote[29]Link to footnote As noted by CuspAI’s CTO, Professor Max Welling, laboratory validation is frequently slowed because "recipes are very finicky," and small changes in ambient humidity or trace contaminants can disrupt chemical assembly. For clean tech entrepreneurs, this emphasizes that computational speed must be matched by automated, high-throughput physical synthesis and laboratory infrastructure to deliver real-world impact.[6]Link to footnote[5]Link to footnote
The same compression logic surfaced around intelligent labs more broadly, including foundation models such as Basecamp Research’s EDEN family for antimicrobial peptides and gene insertion, but peer-reviewed synthesis on AI in drug development still emphasises that clinical success rates hinge on empirical validation, interpretability, and regulatory adaptation rather than candidate generation alone.[64]Link to footnote[65]Link to footnote As experimentation automates, bottlenecks shift to biosecurity, trial integrity, and data equity; discovery speed without those controls is not responsible acceleration.
On The Intelligent Lab and the Future of Discovery, Dr. Oliver Vince and Francesco Farina (Basecamp Research), Dr. Simon Kohl (Latent Labs), Anthony Rowe (GSK), and Dr Angela Kukula (MedCity) stressed dual-use vigilance.[81]Link to footnote
Ethical concerns regarding misuse of AI in areas like transmissible diseases require careful consideration and proactive measures. Human oversight remains vital, particularly in addressing biases and ensuring meaningful outcomes.
, The Intelligent Lab and the Future of Discovery[81]Link to footnote
Deep tech scale-up financing and the patient capital gap
On the Deep Tech Stage, a panel of policymakers and venture partners addressed a persistent question: How can European deep tech and clean energy ventures secure long-term, domestic scale-up capital to prevent the migration of technology and talent overseas?[13]Link to footnote[52]Link to footnote[24]Link to footnote
Stephen Welton, Chair of the British Business Bank, opened the discussion by identifying a structural gap in the funding landscape.[33]Link to footnote While the region is highly efficient at seeding deep tech startups through academic spinouts, a lack of deep, long-term growth capital limits these companies from scaling locally. Welton detailed how the British Business Bank is expanding its investment capabilities to support larger domestic venture capital funds and increase institutional investment in high-growth technology. This strategy was reinforced by the bank's partnership with Playground Global, supported by a £150 million commitment to back advanced hardware and clean energy scale-ups.
Academic Spinouts
Late-Stage Growth Scale-Ups
The panel, which included Dr. Hermann Hauser of Amadeus Capital Partners, debated the role of state-backed initiatives compared to private capital. Hauser argued that without structured, region-wide sovereign investment funds, European innovations will continue to be acquired by better-capitalized US or Asian buyers. There was broad consensus that state-backed funds must act as anchor investors. However, some panelists expressed skepticism about whether public initiatives can match the speed and risk tolerance of private global funds, particularly in capital-intensive sectors like clean energy infrastructure and hardware-dependent deep tech.
Tech Nation’s LTW briefing sharpened the economic-ethics stake: the UK hosts more than 2,500 VC-backed AI startups in a market valued near £1.6 trillion, with £11 billion raised in six months, yet roughly half of that VC originates in the US, and 57 pence of every £1 of AI exit value accrues to US entities versus 9 pence retained in the UK.[66]Link to footnote Domestic innovation without domestic value capture is not merely a financing inconvenience; it constrains national agency to steer AI alignment toward public-interest goals, and it mirrors the dependency patterns Global South ecosystems already know too well.[78]Link to footnote
Startups prioritised government support in four areas: investment, AI talent, energy infrastructure for data centres, and advancing AI safety. Half of VC and 57p of every £1 of AI exit value flows to the US, leaving only 9p in the UK.
, Marco De Novellis, Founders Forum Group, UK Tech in 2026: The Tech Nation Report[66]Link to footnote[82]Link to footnote
On The Next Wave of AI, Jacomo Corbo (PhysicsX), Tamar Gomez (Ankar AI), and Ollie Ilot (UK Government Emerging Technology & AI) made the same point as industrial policy: early-stage access without growth-stage retention is a sovereignty failure.[82]Link to footnote
The UK is strong on early-stage access, but limited growth-stage capital often forces reliance on US or Asian investors to scale. Stronger domestic demand and local procurement incentives are needed to retain companies and talent.
, The Next Wave of AI[82]Link to footnote
Vertically integrated infrastructure and decentralized utility disruption
The Founders Stage addressed a vital operational question: How can a clean technology startup scale rapidly within a highly regulated, capital-intensive, and historically slow-moving energy market?[53]Link to footnote[40]Link to footnote[38]Link to footnote[36]Link to footnote[37]Link to footnote
In a session titled In Conversation with Alan Chang: How Fuse Energy Scaled in a Broken Market, the co-founder and CEO of Fuse Energy[39]Link to footnote[41]Link to footnote laid out a vertically integrated model designed to disrupt traditional utility structures. Chang (previously Chief Revenue Officer at Revolut) argued that the standard European clean energy retail model is structurally flawed. The majority of "green" energy suppliers do not own or build clean generation assets. Instead, they operate as purely financial intermediaries, trading Renewable Energy Certificates of Origin (REGOs) to hedge their reliance on fossil-fuel wholesale markets.[17]Link to footnote To address this, Fuse Energy integrates generation, retail distribution, and software optimization:
- Vertical Asset Integration: Fuse directly develops, owns, and operates its own clean generation assets, including solar farms in Southern England and wind farms in Scotland.
- Algorithmic Energy Trading: The venture utilizes a proprietary algorithmic trading platform that relies on probabilistic price models and intraday optimization to bid clean electricity directly into wholesale markets, increasing the revenue capture rate of its clean assets by low-double-digit percentages.
- Machine Learning Load Forecasting: Fuse integrates hybrid machine learning models leveraging real-time weather feeds, historical demand patterns, and localized grid signals to predict consumer demand, reducing forecast errors and cutting dispatch-related operational overhead by roughly 22%.
- Consumer Tariffs: By bypassing wholesale intermediaries, Fuse offers domestic tariffs that consistently undercut the state regulator's price cap by 10% to 15%, building consumer trust and a stable customer base.
With a Series B valuation of $5 billion and approximately 200,000 households under contract, Fuse represents a major shift toward treating clean energy not as a financial commodity, but as a vertically integrated technology platform. For a young entrepreneur, this highlights a vital strategic lesson: attempting to navigate fragmented utility regimes with pure-play software is a recipe for stagnation. To achieve scale, startups must integrate physical infrastructure with intelligent optimization software.[16]Link to footnote
Responsive control systems, including reinforcement-learning controllers for building and campus energy, adapt continuously to occupancy, weather, and grid signals, but reviews stress sample inefficiency, transfer fragility, and the risk that mis-specified rewards prioritise kilowatt-hours over occupant comfort and safety.[67]Link to footnote[68]Link to footnote Industrial claims at LTW of roughly 20% energy savings from basic controllers and up to 60% peak-load reduction in large facilities (e.g. Schneider Electric / AVEVA case studies) should therefore be read as deployment hypotheses: adaptive optimisation needs explicit safety envelopes, human override, and post-deployment monitoring to remain responsible infrastructure rather than opaque automation.[61]Link to footnote[68]Link to footnote
On Driving Innovation Toward a Greener Future (9 June), moderated by Mickey Carroll (Sky News), Philippe Rambach (Chief AI Officer, Schneider Electric), Arti Garg (Chief Technologist, AVEVA), and Julia Reinaud (Senior Director, Breakthrough Energy) framed AI as a decarbonisation lever, and named the dual-use problem directly.[83]Link to footnote Schneider Electric reported demand-reduction and peak-load savings on the order of 20% in basic room controllers and up to 60% in large facilities; AVEVA stressed process simulation for industrial optimisation and sustainable materials; Breakthrough Energy pointed to AI for critical-mineral discovery, geothermal siting, and Global South agriculture (weather forecasting through fertiliser optimisation). Across the panel, deployment success hinged on business value over tech novelty, cross-functional domain–AI squads, adaptive regulation, and transparent metrics for AI’s own energy footprint, still lacking standardised measurement, especially outside Europe.
The dual challenge is to use AI to reduce emissions while managing AI’s own energy consumption. AI in agriculture, from weather forecasting to fertiliser optimisation, is vital for addressing climate challenges in the Global South.
, Philippe Rambach / Arti Garg / Julia Reinaud, Driving Innovation Toward a Greener Future[83]Link to footnote
Geospatial AI and the space infrastructure race
On 10 June, Olympia’s space and Earth-observation track made the same dual-use and access questions concrete for climate MRV, agriculture, and sovereign infrastructure.[94]Link to footnote[95]Link to footnote[96]Link to footnote[97]Link to footnote[98]Link to footnote[99]Link to footnote
On Why Enterprises Need to Think about Geospatial AI, Today, Will Marshall (Co-Founder & CEO, Planet), in conversation with Professor Helen Czerski, described a constellation of 200+ satellites imaging Earth daily, feeding “large Earth models” that process more than 4 million images per day for agriculture, journalism, disaster response, and security.[99]Link to footnote Hyperspectral sensing for methane leaks and vegetation typing, AI-enabled transparency credited with helping cut deforestation in Brazil by about 60%, and NATO hybrid-warfare monitoring illustrated dual-use Earth intelligence. Marshall stressed Planet’s public-benefit corporation model and ethics-committee oversight, plus ambitions to democratise access for farmers and journalists and to trial space-based compute with Google, constant solar power and simpler cooling as cost advantages for on-orbit processing.[99]Link to footnote
Geospatial AI bridges the gap between data and actionable insights, ethical business models and transparent practices are foundational to responsible use.
, Will Marshall, Planet, Why Enterprises Need to Think about Geospatial AI, Today[99]Link to footnote
From Launchpad to Orbit linked that sensing stack to UK scale-up policy. Baroness Lloyd of Effra CBE (Minister for Digital Economy) framed space as national infrastructure; the UK Innovation Science Seed Fund’s £9.25 million commitment via Future Planet Capital was pitched as de-risking early space ventures and catalysing private capital roughly sixfold.[94]Link to footnote Dr Marco Rocchetto (SpaceFlux) and Vishal Soomaney Vijaykumar (Messium) credited ESA BIC and UK Space Agency programmes plus early government contracts for international expansion, Messium’s hyperspectral satellites targeting precision agriculture and fertiliser optimisation as an environmental-impact use case.[94]Link to footnote
The Space Infrastructure Race, Rob Desborough (Seraphim VC), Carissa Bryce Christensen (BryceTech), Giorgio Taylor (Xona Space Systems), Steve Young (ICEYE), cast venture-funded dual-use players as the new infrastructure layer: ICEYE’s SAR for near-real-time monitoring; Xona’s next-generation navigation against GPS jamming and spoofing; government defence and intelligence demand as the primary growth driver; and procurement reform so primes are not the only path to sovereign capability.[98]Link to footnote
The Age of the New Space Entrepreneur pushed democratised launch as a market-structure problem. Maureen Haverty (Seraphim), Dr Katie King (BioOrbit), Rita Rinaldo (ESA), and Manu Nair (Ethereal Exploration Guild) argued that reusable rockets and re-entry capabilities are unlocking microgravity pharma and space-data applications in insurance and logistics, but over-reliance on a single launch provider risks monopolising orbital “real estate.” India’s geographic launch advantages and Europe’s IRIS2 sovereignty track were named as counterweights.[97]Link to footnote
How to Solve the Space Debris Problem treated orbital pollution as an everyday-infrastructure risk. Andrew Faiola (Astroscale), Prof. Jonathan Eastwood (Imperial College London), Dr Alex Barber (ESSI), and Joanne Wheeler MBE (Alden / ESSI) called for removing large derelicts first to prevent cascade fragmentation, mandatory materials inventories for atmospheric burn-up toxicity, licensing and insurance incentives, and international regulatory alignment, with the UK positioned to lead on legal, financial, and scientific capacity if action is timely.[95]Link to footnote
Preparing for Space: The Making of an Astronaut, Dr Meganne Christian (ESA Astronaut Reserve / UK Space Agency) with Libby Jackson OBE (Science Museum), closed the human-spaceflight loop: ESA’s reserve model and commercial stations (Axiom, VAST, Starlab) broaden participation, while microgravity research (protein crystallisation, organ bioprinting) and Artemis lunar pathways were framed as Earth-benefit science, not only exploration prestige.[96]Link to footnote
On The Fifth Domain and the Future of Intelligent Infrastructure, Rory Daniels (techUK), Max Buchan and Ryan Radloff (Valarian), Prof. Maire O’Neill (Queen’s University Belfast), and Tyler Edwards (Overmind) extended responsibility into embodied and agentic systems.[84]Link to footnote
Sovereignty is the ability to control, modify, and run systems without external interference. Kill switches are needed in AI systems to manage agentic workflows without disrupting entire operations.
, The Fifth Domain and the Future of Intelligent Infrastructure[84]Link to footnote
Policy implications and geopolitical positioning
The outcomes of London Tech Week 2026 highlighted a widening gap between the UK's bilateral technology policy and the European Union's regulatory, ecosystem-focused industrial strategy.[20]Link to footnote
The UK's muscular bilateral policy
Post-Brexit, the UK is executing a highly directed, bilateral technology policy designed to position London as a global hub for deep tech and sovereign computing. Under Starmer and Technology Secretary Liz Kendall, the state is shifting away from pure horizontal regulatory oversight toward a vertical, interventionist industrial strategy. This approach includes direct interventions in capital markets, such as legal reforms aimed at channelling up to £25 billion in pension-fund assets directly into high-growth British deep tech and clean energy ventures.[10]Link to footnote Simultaneously, the UK is moving quickly to establish bilateral trade partnerships across the Asia-Pacific region. This is driven primarily by two key developments:
- The CPTPP Accession: Following its formal entry into the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP), the UK has gained access to a vast, low-tariff digital trade bloc spanning Canada, Japan, Australia, and Southeast Asia. This framework establishes clear parameters for cross-border data flows, bars localized data-residency mandates, and bans import duties on cross-border digital transmissions.
- The UK-Japan Frontier Technology Partnership (FTP): Launched during Downing Street bilaterals, the FTP establishes direct R&D and commercialization corridors in advanced semiconductors, quantum computing, and materials science, bypassing standard multilateral European programs.[47]Link to footnote
The EU's regulatory ecosystem framework
In contrast, the European Commission's newly unveiled European Technological Sovereignty Package (ETSP), comprising the revised Chips Act 2.0 and the Cloud and AI Development Act (CADA), pivots toward structured, region-wide industrial protectionism.[43]Link to footnote[59]Link to footnote[60]Link to footnote While the EU is attempting to streamline corporate compliance and reduce administrative burdens for SMEs, its primary mechanism is a strict, criteria-based sovereignty assessment framework. CADA introduces four distinct tiers of "Union assurance," forcing public sector procurement teams to systematically prioritize European cloud infrastructure that is legally insulated from foreign extraterritorial demands, such as the US CLOUD Act.[42]Link to footnote
United Kingdom
European Union
On Europe Is Rising to the Challenge, Ned Baker (Helsing UK), Dr. Hayaatun Sillem CBE (Argentic Associates), and Kyle Thomas (SAIF Autonomy) tied sovereignty to public mandate rather than isolationism.[88]Link to footnote
Sovereignty is not isolationism but the ability for Europe to act independently if necessary. Autonomy and AI can reduce human risk, but need public trust and a societal mandate.
, Europe Is Rising to the Challenge[88]Link to footnote
On Scaling Sovereign Innovation through Frontier Partnerships, Dr. Rich Drake (Anduril UK), Erin Hallock (NATO Innovation Fund), Charlotte Warburton (Deloitte), and Robyn Staverley (ARX Robotic) argued that Europe’s barrier is adoption speed and procurement design, not capital scarcity.[90]Link to footnote
Europe does not face a capital shortage but an adoption crisis. Outdated procurement, monolithic multi-year programmes, hinders startup scalability and innovation. Cross-border collaboration and interoperability are critical; countries must stop working in silos.
, Scaling Sovereign Innovation through Frontier Partnerships[90]Link to footnote
For a climate tech venture based in Brussels with Asian ties, this regulatory divergence presents both a barrier and an opportunity. While CADA's strict sovereignty mandates restrict the use of global public clouds for sensitive workloads in Europe, the UK's CPTPP integration and newly concluded digital trade deals, such as the ASEAN Digital Economy Framework Agreement (DEFA), provide clear corridors to export advanced climate analytics, localized grid optimization software, and materials databases directly into rapidly growing Asian markets.[23]Link to footnote[45]Link to footnote
AI as infrastructure: ethics, labour, and responsive governance
At Olympia, artificial intelligence was framed less as a speculative frontier than as connective infrastructure powering compute, capital, scientific discovery, and industrial decarbonisation.[61]Link to footnote Moving AI from localised promise to durable system-level infrastructure creates acute socio-technical friction. Grounding LTW’s floor narratives in empirical research, and reading them from Global South impact standpoints, shows that technical performance alone will not dictate long-term value; value retention, trust, and inclusive, responsive governance will.[62]Link to footnote[69]Link to footnote[78]Link to footnote
Global South equity test: who designs, who bears cost?
AI for Global Impact, with Sir John Lazar (Royal Academy of Engineering), May Habib (Writer), and Rowland Manthorpe (Sky News), was one of the few LTW sessions that explicitly asked whether London’s AI story travels equitably beyond high-income hubs.[85]Link to footnote
Equity, inclusion, and citizen trust require accessible AI solutions tailored for underrepresented regions and communities. Innovation driven by constraints, such as low-cost AI applications in Africa, could inspire global solutions.
, AI for Global Impact: Solving the Defining Challenges of Our Time[85]Link to footnote
That framing matches CIGI Policy Brief No. 225 (Maheshwari, 2026): the Global South holds 88% of humanity across 134 countries and 90% of the world’s population under 25, and is projected to drive 65% of global growth by 2035, yet Africa accounts for roughly 18% of humanity and under 2% of global AI investment.[78]Link to footnote Facial-analysis bias research cited in the brief still shows error rates near 34.7% for dark-skinned females versus 0.8% for light-skinned males, a reminder that “global” AI products often fail the people they claim to serve.[78]Link to footnote
Africa: population weight vs AI investment share
- Africa18% people<2% AI capital
Facial-analysis error rates by demographic group
- Error rate34.7%0.8%
Equity
Maheshwari trinity
Ethics
Maheshwari trinity
Ecological sustainability
Maheshwari trinity
Global AI governance frameworks by status (illustrative count)
Compute sovereignty and regional equity
The £1.1 billion Hardware Plan, AMD’s £2 billion HPC commitment, and Nebius’s £1.7 billion / 65 MW build-out treat compute as national infrastructure.[4]Link to footnote[10]Link to footnote Equal access to that infrastructure determines who builds frontier systems. Firm-level evidence on the UK AI economy documents high geographic and capital consolidation in London, leaving regional ecosystems reliant on central hubs.[63]Link to footnote
Geographic concentration of UK AI entities
Work, Copilots, and labour-market ethics
Enterprise deployments, notably NHS England’s Microsoft 365 Copilot rollout to 505,000 clinicians and support staff, with reported pilots reclaiming ~43 minutes/day (~5 weeks/year), are reconfiguring administrative workflows and expanding non-expert access to complex tasks.[70]Link to footnote Governance must address deskilling, worker surveillance, and equitable productivity distribution rather than pilot efficiencies alone.[62]Link to footnote[69]Link to footnote
Darren Hardman (Corporate Vice President & CEO, Microsoft UK and Ireland), in Shaping the Future of Work and Opportunity in the UK, cast AI literacy as a foundational capability, and regional opportunity as uneven.[86]Link to footnote
Organisations move from information work to intelligence work, where AI acts as a collaborator rather than merely a tool. While talent is distributed evenly across the UK, opportunities are not.
, Darren Hardman, Microsoft UK and Ireland[86]Link to footnote
On AI That Helps Humans Thrive (10 June), Michelle He (Abound), Toyin Ajayi (Cityblock Health), and Serena Dayal (Athena Capital) measured AI’s value by personalisation, efficiency, and fairness for populations traditional systems underserve.[92]Link to footnote Abound’s open-banking credit decisioning cut default rates by roughly 70%, improving pricing and access; Cityblock’s models supported care for about 150,000 high-risk patients through cost prediction and scalable personalisation. Both stressed human–AI collaboration in regulated settings, AI for 100% quality-assurance coverage and back-office optimisation; humans for trust, exceptions, and empathy, and a forward vision of 24/7 personalised health coaches and AI financial assistants that only works if access is equitable.[92]Link to footnote
AI helps humans thrive when it improves efficiency, personalisation, and fairness, ensuring its advantages reach all societal segments.
, Michelle He / Toyin Ajayi, AI That Helps Humans Thrive[92]Link to footnote
Comparative US firm-level evidence from Ramp and Revelio Labs (n = 21,559) complicates mass-layoff narratives: high-intensity AI adopters show +10.2% total headcount and +12.0% entry-level headcount over 24 months versus not-yet adopters, while low-intensity adopters show no detectable headcount gain.[71]Link to footnote Gains concentrate among already larger, more technical firms, an equity warning for UK public deployments that automate junior pathways, and for Global South labour markets exposed to remittance and offshoring channels.
AI adoption measurement benchmarks (firm surveys)
US firm AI adoption by sector (Dec 2025)
AI adoption by firm size and engineering intensity
Post-adoption headcount path by AI intensity (illustrative ATT)
High-intensity AI adopter headcount ATT by role (0–24 months)
Sector ATT for high-intensity AI adopters (0–24 months)
Inclusive leadership remains a governance variable, not only a pipeline story. The EQL:Lounge session Why Women Are the Future of Responsible AI, hosted by Tech She Can with Sheridan Ash MBE (Tech She Can), Zehra Chatoo (Code For Good Now), Karen Blake MBE (Women in Tech Taskforce / Tech Talent Charter), and Alena Frankel (Faculty), framed diverse development teams and women’s leadership in ethics, governance, and product as central to trustworthy AI, not peripheral diversity targets.[91]Link to footnote Adjacent Closing the AI Gender Gap coalition framing noted women comprise 22% of the AI workforce, 18% of AI researchers, and 14% of AI executives, yet occupy more than 80% of roles most exposed to disruption, arguing the gap is driven by trust and structural barriers (harsher judgement for AI use, algorithmic bias, concerns about how systems are built), not a pure skills deficit.[72]Link to footnote
Women’s leadership and diverse teams are central to responsible AI, not peripheral. Non-traditional pathways, career switchers, returners, self-taught talent, are vital to inclusive AI ecosystems; inclusion is a systems-change question rather than a purely individual challenge.
, Why Women Are the Future of Responsible AI / Closing the AI Gender Gap (Tech She Can; Women in Tech Taskforce)[91]Link to footnote[72]Link to footnote
Ada Lovelace research on inclusive governance argues that participation improves legitimacy and trustworthiness of deployed systems, a finding that applies equally to Global South civil-society capacity gaps when standards bodies interpret fundamental rights without public legitimacy.[73]Link to footnote
Gender representation in UK AI roles
Ecosystem value capture
Domestic startup density without domestic value capture creates economic vulnerability. High dependence on foreign capital and cloud providers limits agency to direct AI toward public-interest goals.[66]Link to footnote[63]Link to footnote
UK AI exit-value retention per £1
- Exit value share9p UK57p US
Intelligent labs and dual-use discovery
Foundation-model discovery (including EDEN-style metagenomic and peptide design narratives on the LTW floor) compresses early candidate selection, but clinical translation, biosecurity, and data equity remain the binding constraints.[64]Link to footnote[65]Link to footnote[61]Link to footnote Rational molecule design introduces dual-use risk; Indigenous and biodiversity sampling raise data-sovereignty questions that vendor demos rarely surface.
Safety, evaluation, and binding governance
Ada Lovelace Institute and Apollo Research panelists at LTW criticised the absence of a unified safety roadmap and over-reliance on voluntary, company-led evaluation, calling for supply-chain liability and mandatory pre- and post-deployment testing.[61]Link to footnote[74]Link to footnote On Securing the Future of AI: The Roadmap to Safe & Ethical AI, Gaia Marcus (Ada Lovelace Institute), Marius Hobbhahn (Apollo Research), Patricia Clarke (The Observer), and Jacomo Corbo and Robin Tuluie (PhysicsX) put the failure mode plainly.[87]Link to footnote
No clear governance structure exists, leaving key decisions in the hands of tech companies, which exacerbates risks. Industry-led voluntary safety measures are insufficient. Harm arises not just from technical flaws but from the interaction between systems and their deployment contexts.
, Gaia Marcus / Marius Hobbhahn et al., Securing the Future of AI[87]Link to footnote
Adjacent enterprise practice on 10 June reinforced the same speed-and-control tension. On Assessing the Business Impact of Responsible Innovation, Ingrid Verschuren (Dow Jones), Pooja Bagga (Guardian Media Group), Devesh Raj (Sky), and Elizabeth Seger (Tony Blair Institute) described steering committees, iterative evaluation, Dow Jones’s “authentic intelligence” blend of human oversight with aggregation tools, and Sky’s Athena Labs sandbox for rapid experimentation on synthetic data before full compliance, arguing that clear guardrails enable faster adoption, not slower.[93]Link to footnote Transparency about automated translation and investigative tooling, plus coalitions such as SPER on fair content licensing, were treated as external-trust infrastructure.[93]Link to footnote
Effective governance can enable faster adoption by establishing clear guardrails, allowing safe experimentation and reducing risks.
, Ingrid Verschuren / Pooja Bagga / Devesh Raj / Elizabeth Seger, Assessing the Business Impact of Responsible Innovation[93]Link to footnote
Ada’s Regulate to Innovate frames regulation as the missing link for an “AI superpower”: clear rules, regulator capacity, and transparency beyond the state.[74]Link to footnote Policy and Society synthesises generative-AI risks (hallucination, opacity, labour impacts, power imbalance) and argues adaptive, participatory governance over purely technocratic or voluntary models.[69]Link to footnote
Clear AI rules
Ada pillar 1
Regulator capacity
Ada pillar 2
Transparency & accountability
Ada pillar 3
Developers
Adapters
Deployers
Ada regulatory objectives
Regulate to Innovate
Lifecycle toolkit
Ex ante → ex post
Governance positioning: voluntary safety vs responsive oversight
Control-theoretic work on a Social Responsibility Stack treats fairness drift, autonomy preservation, cognitive burden, and explanation clarity as explicit engineering constraints in a closed loop among AI behaviour, human behaviour, and governance intervention, a concrete vocabulary for responsive AI in critical infrastructure and public systems.[75]Link to footnote
Climate dual-use and net-zero balance
Deploying AI for climate mitigation requires a strict net balance: infrastructure energy draw must not erase abatement gains.[76]Link to footnote[68]Link to footnote Stern et al. (npj Climate Action, 2025) estimate AI could abate 3.2–5.4 GtCO₂e/yr by 2035 across power, meat/dairy, and light road vehicles, outweighing an estimated 0.4–1.6 GtCO₂e rise from global data-centre and AI power use under their scenarios, moving progress roughly 36% closer to an ambitious path versus BAU.[76]Link to footnote
AI climate dual-use balance by 2035 (GtCO₂e/yr)
AI abatement by sector by 2035 (GtCO₂e/yr)
Alternative protein adoption rates under AI scenarios
Emissions trajectories to 2035: BAU, AI scenario, ambitious path
Five AI climate impact areas
| Stern et al. climate-transition framework | ||
|---|---|---|
| Impact area | FocusMitigation / adaptation / resilience | |
| 1. Complex systems | Transforming complex systems | Cities, land, transport, industry, energy, live-data redesign and operation |
| 2. Discovery & efficiency | Technology discovery and resource efficiency | Materials, processes, and productivity that cut emissions intensity |
| 3. Behaviour | Nudging and behavioural change | Demand-side shifts in mobility, diet, and energy use |
| 4. Modelling & policy | Climate modelling and policy interventions | Better forecasts, risk pricing, and policy design support |
| 5. Adaptation | Adaptation and resilience | Early warning, resilient infrastructure, and climate-risk management |
UK practice evidence reinforces the same dual-use logic. The UCL / Royal Academy of Engineering AI for Decarbonisation report maps 90 case studies, Energy 26, Transport 24, Built Environment 22, Industry 14, Multi-sector 4, with forecasting (66), planning (61), and discovery (34) as dominant capabilities.[77]Link to footnote
UK AI-for-decarbonisation case studies by sector (n=90)
AI capabilities identified across decarbonisation case studies
AI pathways to net zero
Olawade et al. (2024) synthesis
Ethics and governance guardrails
Same review, risk framing
Literature mapping: LTW themes × ethics × evidence
| Responsible and responsive AI synthesis | |||
|---|---|---|---|
| LTW 2026 focus | Ethics & governance lens | Evidence base | |
| Compute & infrastructure | Sovereign hardware, supercomputing, private capital | Access equity, green energy sourcing, geographic concentration | Nature Climate Action (2025); UK AI firm performance (2025/26) |
| Future of work | Enterprise Copilots, admin automation, skills initiatives | Deskilling, displacement, surveillance, productivity distribution | Ramp–Revelio (2026); Policy and Society (2025) |
| Intelligent discovery | Zero-shot design, metagenomic / materials foundation models | Dual-use biosecurity, data sovereignty, clinical validation | Nature Medicine (2025); ACS Omega (2025) |
| AI governance & safety | Safety roadmaps, supply-chain liability debates | Failure of voluntary testing; continuous post-deployment eval | Ada Regulate to Innovate; Social Responsibility Stack (2025) |
| Climate & net zero | Industrial simulation, building peak-load reduction, Global South agriculture AI | Net climate impact: infrastructure draw vs abatement gains; electricity access equity | Stern et al. (2025); UCL AI for Decarbonisation (2024); Maheshwari (2026) |
| Global South impact | AI for Global Impact session; constraint-driven design | Capital under-allocation, bias, frugal AI, child protection | Maheshwari CIGI PB 225 (2026); GIRAI / jagged-economy framing |
Stakeholder perspectives and ecosystem fragmentation
The discussions at London Tech Week 2026 revealed significant fragmentation across European tech ecosystems. This friction was particularly apparent between academic hubs, government agencies, corporate partners, and emerging founders.[1]Link to footnote
Academic hubs versus scaling capital
Academic hubs, represented at the event by Imperial College London and its "WestTech London" innovation district in White City, demonstrate exceptional "discovery velocity". Spinouts like Jelly Drops combine design and technical innovation to solve complex societal challenges. Alyssa Gilbert, Co-Director of the School of Convergence Science at Imperial, noted on the Innovation Arc Day[34]Link to footnote that academic environments excel at early-stage discovery, but are often disconnected from industrial deployment.[1]Link to footnote This early-stage excellence contrasts with a persistent capital gap at the scale-up stage.[13]Link to footnote Despite the presence of state-backed entities like the British Business Bank, venture capital remains highly risk-averse, particularly for deep tech hardware and infrastructure projects. As a result, many European innovators are forced to seek late-stage capital from US or Asian growth funds, leading to an asymmetric migration of intellectual property and talent.[10]Link to footnote
Government agencies and border silos
Government agencies at the event, including the UK's Department for Science, Innovation and Technology (DSIT) and the Department for Business and Trade (DBT), heavily promoted national support schemes. The DBT soft-launched its "Concierge Service" to streamline regulatory pathways for high-growth firms in eight priority industrial sectors, including clean energy and digital technology. Additionally, the Global Talent Taskforce was strengthened to offer visa fee reimbursements for scale-ups in clean energy.[50]Link to footnote However, these initiatives remain structurally confined within national borders. For a startup based in Brussels, navigating these national silos requires setting up separate entities, complying with disconnected tax regimes, and managing redundant regulatory requirements. This lack of policy coordination across European borders continues to hinder startups from achieving region-wide scale.[14]Link to footnote
Brussels Vantage
London Hub
Emerging founders and the generation gap
The event also exposed a clear generational divide. LTW’s physical VIP spaces and core investor match-making lounges, such as Founders Fuse, were heavily dominated by established, older-generation founders and senior partners from major VC funds. Grassroots fringe programs, such as NexTech Hub’s Bridging Access virtual panels, championed accessible, inclusive pathways for underrepresented and younger tech professionals.[54]Link to footnote However, these initiatives remained largely separated from the main stages where high-value deals are made.[51]Link to footnote This generational exclusion is particularly challenging for young climate entrepreneurs who operate with an acute sense of climate urgency. Because traditional European funding networks remain highly gatekept, younger founders are increasingly looking outside the continent for partners. This trend was highlighted by the active presence of Asian investment hubs, such as InvestHK’s Founders Fuse Lounge, which pitched Hong Kong as a low-barrier, highly capitalized launchpad for expansion into the broader Asia-Pacific market.[55]Link to footnote
Critical analysis: gaps, contradictions, and hypocrisies
A critical assessment of London Tech Week 2026 reveals four major contradictions at the intersection of climate action, advanced compute, public procurement, AI safety, and Global South equity.[11]Link to footnote
The grid mismatch: compute intensity versus net-zero mandates
The most glaring hypocrisy of the event was the disconnect between the climate-action commitments championed by techUK’s Climate Action Hub and the reality of the AI compute boom discussed in the AI Arena. Throughout the panels, industry leaders confidently asserted that digital technology will enable large emissions cuts by 2030 through smart agricultural monitoring, grid management, and supply chain optimization.[56]Link to footnote Peer-reviewed bottom-up work is more precise, and more demanding: Stern et al. estimate 3.2–5.4 GtCO₂e/yr of AI-enabled abatement by 2035 in three major sectors, contingent on aggressive management of a 0.4–1.6 GtCO₂e data-centre and AI electricity overhead.[76]Link to footnote Floor rhetoric that treats climate AI as an unqualified win without net accounting fails that test.
Yet, the compute infrastructure required to power this transition is driving an unprecedented surge in grid demand. Nebius’s £1.7 billion compute expansion, designed to deploy 65 megawatts of Nvidia hardware by 2027[10]Link to footnote, illustrates the scale of the issue. The carbon intensity of this advanced compute can be modeled as CI = P_compute × PUE × CI_grid, where:
- P_compute represents the raw electrical power consumed by the GPU clusters during model training and inference.
- PUE is the Power Usage Effectiveness of the hosting data center.
- CI_grid represents the carbon intensity of the local grid.
As organizations scale autonomous, multi-step agentic AI systems, which execute continuous, looped inference cycles rather than single-prompt queries, both P_compute and inference cycle frequency scale exponentially. Despite these metrics, the structural barriers to securing clean, localized power for data centers were treated as a secondary issue. In private roundtables, data center operators noted that planning and land use laws, combined with severe grid interconnection backlogs, mean that deploying sustainable, localized power infrastructure in Europe will take years. This reality clashes directly with corporate Scope 3 decarbonization commitments[58]Link to footnote, revealing a critical gap: the European tech ecosystem is rapidly scaling carbon-intensive computing while lacking a credible, synchronized strategy to decarbonize the underlying grids.
Open-source idealism versus public sector lock-in
The second major hypocrisy lies in the tension between open-source policy advocacy and the reality of public sector procurement. Both the EU's Open Source Strategy and the UK's compute initiatives heavily promote open-source AI as a tool for digital sovereignty. The UK government, for instance, pledged £500,000 in direct compute credits for open-source AI builders.[14]Link to footnote Yet, when analyzing high-value, long-term public sector contracts, this open-source rhetoric is consistently undermined by proprietary lock-in. While startups face intense capital constraints, massive public contracts are routinely awarded to non-European, closed-source legacy giants. For example:
- Anthropic was commissioned to develop the generative AI assistant for the UK's central portal, GOV.UK.[10]Link to footnote
- Palantir continues to hold a massive, highly controversial £330 million NHS data integration contract, despite sustained pressure from domestic privacy advocates and tech groups to transition to open-source alternatives.[10]Link to footnote
This reveals a deep-seated policy contradiction. European and UK policymakers publicly advocate for open-source innovation, GDPR compliance, and domestic digital sovereignty.[42]Link to footnote However, when deploying critical public infrastructure, they repeatedly fallback on proprietary, US-headquartered platforms.[10]Link to footnote This practice locks in public systems, starves local open-source startups of critical early-stage validation contracts, and reinforces the very infrastructural dependencies they claim to be dismantling. It also weakens inclusive governance: when procurement concentrates decision power in closed vendors, civil-society and SME auditors lose the practical ability to contest system behaviour, the opposite of Ada’s call for transparency beyond the state.[73]Link to footnote[74]Link to footnote
Voluntary safety versus responsive socio-technical evaluation
The third contradiction sits in AI safety itself. LTW’s Ada/Apollo panels argued that voluntary, company-led evaluation cannot mitigate systemic risk, algorithmic bias, or agentic failure, yet much of the UK’s principles-based posture still privileges soft guidance over binding lifecycle duties.[61]Link to footnote[74]Link to footnote True safety requires socio-technical evaluation of human–AI interaction in operational environments, supply-chain liability across developers, adapters, and deployers, and continuous post-deployment monitoring rather than one-off launch checklists.[69]Link to footnote[75]Link to footnote Without those instruments, “responsible AI” remains a brand posture adjacent to the AI Arena, not an operating constraint on sovereign compute and public Copilot rollouts.
Global South as afterthought versus design bar
A fourth gap is geographic: Olympia’s loudest infrastructure narratives (sovereign compute, hyperscaler PPAs, NHS Copilots) still centre high-income deployment contexts, while AI for Global Impact, Greener Future agriculture remarks, and geospatial / space sessions (Planet’s large Earth models; Messium’s hyperspectral fertiliser optimisation; multipolar launch and debris commons) remain thinner than the capital and chip announcements.[85]Link to footnote[83]Link to footnote[99]Link to footnote[94]Link to footnote Maheshwari’s equity–ethics–ecology trinity shows why that imbalance matters: 88% of humanity and most of the world’s youth live in the Global South, Africa’s AI investment share is under 2%, and 750 million people still lack electricity, so “AI for climate” and “AI for work” claims that ignore energy poverty, frugal sensing, and multilingual pivots risk exporting Northern stack defaults as development policy.[78]Link to footnote Constraint-driven solutions (Bhashini’s 22-pair pivot; >85% accuracy low-cost traffic sensing) should set the design bar for responsible global AI, not appear as charity-stage footnotes to London’s Decisive Decade.[78]Link to footnote[85]Link to footnote
Forward-looking assessment
Over the next 1–3 years, London’s AI posture will matter less for chip counts than for whether UK and European institutions export responsible, responsive, and equitable AI practices, or export dependency, voluntary safety theatre, and Northern design defaults into ASEAN, Africa, and other Global South markets.[46]Link to footnote[78]Link to footnote[85]Link to footnote From an Impact Intelligence Lab standpoint, five priorities follow:
1. Treat Global South constraint-driven AI as a design bar, not a catch-up story
Low-cost, low-electricity, multilingual systems (and frugal model strategies such as language pivots) should set product requirements for climate, health, and public-service AI, not be treated as afterthought localisation.[78]Link to footnote[85]Link to footnote ASEAN DEFA and CPTPP digital corridors are useful trade rails, but only if deployments respect local agency, data sovereignty, and child-protection norms.[45]Link to footnote[46]Link to footnote
2. Pair every climate-AI claim with net accounting
Vendor energy-savings numbers and Stern-style abatement ranges only become responsible when paired with measured inference/training load, renewable procurement, and standardised footprint metrics, the gap Schneider/AVEVA panels already flagged at LTW.[76]Link to footnote[83]Link to footnote Co-locating clean generation with compute (CADA acceleration zones, microgrids) is a delivery pattern, not a substitute for disclosure.[48]Link to footnote
3. Make socio-technical evaluation mandatory in public deployments
NHS-scale Copilots and agentic infrastructure need pre- and post-deployment evaluation, kill-switch / override design, and liability mapping across developers, adapters, and deployers, Ada’s regulate-to-innovate toolkit, not company-led checklists.[74]Link to footnote[75]Link to footnote[87]Link to footnote[84]Link to footnote
4. Close value-capture and “tech desert” gaps together
UK exit-value leakage (57p/£1 abroad) and Global South capital under-allocation (<2% of AI investment to Africa) are the same structural problem at different scales: activity without retained agency.[66]Link to footnote[78]Link to footnote Domestic procurement, regional compute access, and patient capital are ethics instruments as much as industrial policy.[79]Link to footnote[82]Link to footnote
5. Centre inclusion as system design, not pipeline PR
Women’s underrepresentation in AI roles alongside overrepresentation in AI-exposed occupations means responsible AI requires inclusive governance, not only STEM recruitment campaigns.[72]Link to footnote[73]Link to footnote Diverse leadership and civil-society audit capacity are prerequisites for trustworthiness in both London and Global South deployments.
Cloud and AI Development Act (CADA)
European Commission, European Technological Sovereignty Package
Chips Act 2.0
European Commission, revised semiconductor framework
- 1.Imperial at London Tech Week | About
- 2.Reviews for London Tech Week 2026, UK - Eventible
- 3.London Tech Week 2026 Agenda Goes Live as New Wave of Global Founders and Enterprise Leaders Announced
- 4.London Tech Week 2026: Britain puts people at the heart of the AI revolution - Fintech Circle
- 5.London Tech Week 2026 - Olympia
- 6.London Tech Week 2026
- 7.London Tech Week 2026 Opens Registration Announcing First Global Speakers As We Enter “Europe's Decisive Decade”
- 8.London Tech Week 2026 agenda puts sovereign AI centre stage - IT Brief UK
- 9.London Tech Week 2026: Sovereign AI Takes Centre Stage - Generation Digital
- 10.London Tech Week 2026: the AI billions, the US build-out, and a royal first - TNW
- 11.London Tech Week 2026: The headlines and highlights beyond the hype - Infinite Global
- 12.London Tech Week 2026 | June 8-12 | Olympia London - Relve
- 13.Scaling UK Deeptech: Building Global Companies from Britain - London Tech Week 2026
- 14.UK tech sovereignty: Insights from London Tech Week 2026 | TLT LLP
- 15.3 Trends Shaping the Startup Landscape: Insights from London Tech Week 2026
- 16.Meet extraordinary Fuse Energy I HSBC Innovation Banking UK
- 17.What is Brief History of Fuse Energy Company? – businessmodelcanvastemplate.com
- 18.CuspAI: The Cambridge Startup Rewriting the Rules of Materials Discovery | Content Hub
- 19.techUK's Market Access Brief: International Opportunities for tech companies
- 20.UK-APAC Tech Forum | London Tech Week Event
- 21.techUK's Market Access Brief: International Opportunities for tech companies
- 22.Prime Minister's speech at London Tech Week 2026 - GOV.UK
- 23.UK Sovereign AI: A Big Week for British Compute - University of Bristol
- 24.Top speakers at London Tech Week 2026 | See who's confirmed
- 25.London Tech Week 2026 to Showcase Global Deep Tech Innovators in Space, Robotics, Sciences, Quantum and AI
- 26.Stages | London Tech Week 2026
- 27.Intelligent Matter and the Materials of Tomorrow - London Tech Week 2026
- 28.Chad Edwards - London Tech Week 2027
- 29.CuspAI | The heart of talent development in AI and digital innovation - LAB42
- 30.AI Accelerates Discovery of Next-Gen Carbon Capture Materials
- 31.Cusp.ai - Studio Chong
- 32.CuspAI: When AI Meets Carbon Capture - Amsterdam Science Park
- 33.Stephen Welton - London Tech Week 2026
- 34.Innovation Arc Day | London Tech Week 2026
- 35.London Tech Week 2026 virtual event
- 36.Alan Chang - London Tech Week 2027
- 37.In conversation with Alan Chang: How Fuse Energy Scaled in a Broken Market - London Tech Week
- 38.Fuse | Lowercarbon Capital
- 39.Fuse Energy - Wikipedia
- 40.From Fintech to Green Tech | News - Cranbrook Connects
- 41.Fuse Energy Review 2026: The Fintech-Built UK Supplier - SwitchPilot
- 42.Europe's Technology Sovereignty Package: What Do the Cloud & AI Development Act and Chips Act II Mean for UK Tech? - techUK
- 43.The European technological sovereignty package – a change in the EU's approach to digital autonomy - Wolf Theiss - Leading Lawyers in CEE&SEE
- 44.UK-APAC Tech Forum 2025 | Event Round-Up - techUK
- 45.UK-APAC Tech Forum - techUK
- 46.ASEAN Concludes the Digital Economy Framework Agreement (DEFA) – What's in it for tech? | techUK | Official Press Release - WiredGov
- 47.International policy and trade - techUK
- 48.The EU just made sovereign, open source AI official policy. - Xinity AI
- 49.London Tech Week 2026 - techUK
- 50.London Tech Week 2026: Government unveils new scale-up support package - techUK
- 51.Experience | London Tech Week 2026
- 52.The Untapped Talent Wave Powering Global Growth - London Tech Week 2026
- 53.London Tech Week 2026: Unveils Its Founders Stage As Europe's Builders Navigate An Era of AI, Capital Shifts and New Routes to Scale - EntrepreNerd
- 54.Bridging Access: Empowering the Next Generation Through Technology and Innovation - London Tech Week 2027
- 55.London Tech Week 2026 - Founders Fuse Lounge - InvestHK
- 56.London Climate Action Week 2026 - techUK plans
- 57.Purpose, Power, Policy: Key Takeaways from London Tech Week 2026
- 58.The companies leading on climate aren't waiting for 2050 - City AM
- 59.Europe advances digital sovereignty as EU Commission unveils measures to strengthen technological independence - ak europa
- 60.European Technological Sovereignty Package: European Commission adopts Chips Act 2.0 | Practical Law, https://uk.practicallaw.thomsonreuters.com/w-050-3569?transitionType=Default&contextData=(sc.Default)
- 61.London Tech Week 2026 AI session synthesis, compute-as-infrastructure framing, Ada/Apollo safety critique, intelligent-lab and climate case studies (Tech Nation / Founders Forum floor reporting).
- 62.Governance of Generative AI, Policy and Society (2025)
- 63.Ashraf, Coyle & Debnath, Code, Capital, and Clusters: Understanding Firm Performance in the UK AI Economy
- 64.Artificial Intelligence in Drug Development, Nature Medicine (2025)
- 65.AI-Driven Drug Discovery: A Comprehensive Review, ACS Omega (2025)
- 66.Tech Nation Report 2026 findings presented at London Tech Week (Marco De Novellis / Founders Forum), >2,500 VC-backed UK AI startups; £11bn raised in six months; ~50% VC from US; 57p of £1 exit value to US entities vs 9p retained in UK.
- 67.A Review of Reinforcement Learning for Controlling Building Energy Systems
- 68.Artificial intelligence potential for net zero sustainability, Olawade et al., Next Sustainability (2024)
- 69.Governance of Generative AI, Policy and Society (2025), adaptive, participatory governance of generative systems
- 70.Microsoft 365 Copilot for NHS England staff (505,000-user rollout / pilot time-reclaim reporting)
- 71.Ramp & Revelio Labs, AI Jobs Impact (2026), high-intensity adopters +10.2% headcount / +12.0% entry-level over 24 months (US firm-level evidence).
- 72.Closing the AI Gender Gap coalition framing at London Tech Week 2026
- 73.Inclusive AI governance, Ada Lovelace Institute (2023)
- 74.Regulate to innovate, Ada Lovelace Institute (2021)
- 75.The Social Responsibility Stack: A Control-Theoretic Architecture for Governing Socio-Technical AI
- 76.Stern et al., Green and intelligent: the role of AI in the climate transition, npj Climate Action (2025)
- 77.AI for Decarbonisation: Capability, Current Practice and Trends, UCL / Royal Academy of Engineering (2024), https://www.ucl.ac.uk/engineering/sites/engineering/files/4final_report-_ai_for_decarbonisation.pdf
- 78.Maheshwari, AI’s Global South Pivot, CIGI Policy Brief No. 225 (2026).
- 79.Session summary, Building the Compute Foundation for the AI Era (Carolyn Dawson OBE, Founders Forum Group; Dr. Lisa Su, AMD), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 80.Session summary, The Emerging Case for Sovereign AI Development in Europe (Kanishka Narayan MP; George Osborne, OpenAI for Countries; Judith Dada, Visionaries; Matt Harris, HPE), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 81.Session summary, The Intelligent Lab and the Future of Discovery (Basecamp Research; Latent Labs; GSK; MedCity), London Tech Week, 9 June 2026, London Tech Week 2026 virtual event.
- 82.Session summary, The Next Wave of AI / Tech Nation Report briefing (Marco De Novellis, Founders Forum Group; Jacomo Corbo, PhysicsX; Tamar Gomez, Ankar AI; Ollie Ilot, UK Government), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 83.Session summary, Driving Innovation Toward a Greener Future (Philippe Rambach, Chief AI Officer, Schneider Electric; Arti Garg, Chief Technologist, AVEVA; Julia Reinaud, Senior Director, Breakthrough Energy; moderated by Mickey Carroll, Sky News), London Tech Week, 9 June 2026, London Tech Week 2026 virtual event.
- 84.Session summary, The Fifth Domain and the Future of Intelligent Infrastructure (techUK; Valarian; Queen’s University Belfast; Overmind), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 85.Session summary, AI for Global Impact: Solving the Defining Challenges of Our Time (Sir John Lazar, Royal Academy of Engineering; May Habib; Rowland Manthorpe, Sky News), London Tech Week, 9 June 2026, London Tech Week 2026 virtual event.
- 86.Session summary, Shaping the Future of Work and Opportunity in the UK (Darren Hardman, Microsoft UK and Ireland), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 87.Session summary, Securing the Future of AI: The Roadmap to Safe & Ethical AI (Gaia Marcus, Ada Lovelace Institute; Marius Hobbhahn, Apollo Research; Patricia Clarke, The Observer; PhysicsX), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 88.Session summary, Europe Is Rising to the Challenge (Ned Baker, Helsing UK; Dr. Hayaatun Sillem CBE, Argentic Associates; Kyle Thomas, SAIF Autonomy), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 89.Session summary, AI is the Computer (Aravind Srinivas, Perplexity), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 90.Session summary, Scaling Sovereign Innovation through Frontier Partnerships (Dr. Rich Drake, Anduril UK; Erin Hallock, NATO Innovation Fund; Charlotte Warburton, Deloitte; Robyn Staverley, ARX Robotic), London Tech Week, 8 June 2026, London Tech Week 2026 virtual event.
- 91.Why Women Are the Future of Responsible AI, Tech She Can EQL:Lounge (Sheridan Ash MBE; Zehra Chatoo; Karen Blake MBE; Alena Frankel, Faculty), London Tech Week, 8 June 2026; Closing the AI Gender Gap coalition framing.
- 92.Session summary, AI That Helps Humans Thrive (Michelle He, Abound; Toyin Ajayi, Cityblock Health; Serena Dayal, Athena Capital), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
- 93.Session summary, Assessing the Business Impact of Responsible Innovation (Ingrid Verschuren, Dow Jones; Pooja Bagga, Guardian Media Group; Devesh Raj, Sky; Elizabeth Seger, Tony Blair Institute), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
- 94.Session summary, From Launchpad to Orbit: Supporting Space Ventures Through the Scale-Up Gap (Baroness Lloyd of Effra CBE; Dr Marco Rocchetto, SpaceFlux; Shruti Iyengar, Future Planet Capital; Vishal Soomaney Vijaykumar, Messium), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
- 95.Session summary, How to Solve the Space Debris Problem (Andrew Faiola, Astroscale; Prof. Jonathan Eastwood, Imperial College London; Dr Alex Barber, ESSI; Joanne Wheeler MBE, Alden / ESSI), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
- 96.Session summary, Preparing for Space: The Making of an Astronaut (Dr Meganne Christian, ESA Astronaut Reserve / UK Space Agency; Libby Jackson OBE, Science Museum), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
- 97.Session summary, The Age of the New Space Entrepreneur (Dr Katie King, BioOrbit; Rita Rinaldo, ESA; Manu Nair, Ethereal Exploration Guild; Maureen Haverty, Seraphim), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
- 98.Session summary, The Space Infrastructure Race (Rob Desborough, Seraphim VC; Carissa Bryce Christensen, BryceTech; Giorgio Taylor, Xona Space Systems; Steve Young, ICEYE), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
- 99.Session summary, Why Enterprises Need to Think about Geospatial AI, Today (Will Marshall, Planet; Professor Helen Czerski), London Tech Week, 10 June 2026, London Tech Week 2026 virtual event.
