At London Tech Week 2025, the future didn’t just arrive — it accelerated. Amid global uncertainty, rapid tech evolution, and mounting climate urgency, leaders from every industry gathered in London not just to talk about bold ideas, but to showcase real, actionable solutions. From Agentic AI and autonomous workflows to digital twins that anticipate disruption… from tackling the carbon cost of forgotten data to building AI-driven climate resilience — the event spotlighted the technologies already reshaping how we work, lead, and adapt.
Here are some of the key highlights and real-world use cases that show how today’s innovations are driving tomorrow’s business:
Infrastructure for the Next Industrial Age
London Tech Week 2025 made one message unmissable: artificial intelligence (AI) is no longer a tool; it is the infrastructure upon which the next industrial age will be built. Sir Keir Starmer, UK Prime Minister, opened the week by calling AI a catalyst for national renewal—announcing over £2.5 billion in AI and tech investment, including a £1 billion national compute upgrade and £1.5 billion HQ commitment from Liquidity. These aren’t tech-sector-only moves—they represent a whole-of-society transformation. The NHS, for example, is integrating AI into diagnostics and administration, using Microsoft Copilot to cut paperwork and reduce clinical backlogs.
Nvidia CEO Jensen Huang described AI as a force as consequential as electricity or the internet, urging the UK to build sovereign AI infrastructure, data centres, national language models, and secure data pipelines to maintain competitiveness. He highlighted the UK’s unique position: a top-tier academic ecosystem, high-calibre venture capital, and strong public-private alignment. The UK, he argued, must not become dependent on non-domestic models and infrastructure, likening sovereign AI to national energy or defence capabilities.
The aim isn’t just tech growth—it’s societal uplift: smarter systems, reskilled citizens, and more humane workflows where people focus on value while machines handle complexity.
Agentic AI – From Co-Pilots to Autonomous Colleagues
If AI is the engine of the future, Agentic AI* is the driver. Agentic AI systems are no longer just responding to prompts—they are setting their own goals, making decisions, and orchestrating entire workflows.
London Tech Week showcased how this shift is already reshaping enterprise operations. Microsoft’s Copilot, now deployed across 100,000 Barclays users and civil service teams, saves an estimated two weeks per employee per year. GitHub Copilot, with over 150 million users, is transforming software development into an AI-enhanced collaboration, rather than a purely manual task.
Amazon demonstrated its Bedrock Agents platform, featuring a “supervisor brain” that coordinates multiple AI agents across a task chain—writing, testing, migrating, and documenting code autonomously. Its Amazon Q Developer even handles architectural reviews and deployment checks.
AstraZeneca revealed ‘Milton’, a predictive health agent that monitors clinical and behavioural data to forecast illness before symptoms manifest. Additionally, their AR agents guide lab technicians through complex procedures autonomously, while the upskilling of 11,000 employees ensures these tools are embedded effectively across their workforce.
Start-up Mistral presented its “transcriptToPRDTicket” agentic pipeline, converting meeting transcripts into detailed product requirement documents and Jira engineering tickets cutting hours of manual work. Prosus discussed its domain-specific “superagents” operating like expert employees, complete with ethical protocols and multi-agent governance. These advances are not science fiction; they are active in global organisations.
However, with this autonomy comes accountability. Businesses must now define the rules of engagement, ensuring agents are auditable, their decisions are explainable, and their biases are minimised.
*Agentic AI refers to AI systems designed to autonomously make decisions and act to achieve complex goals with limited human supervision.
The Self-Healing Supply Chain – Anticipate, Reroute, Deliver
Nowhere is the impact of AI more profound than in supply chains. AstraZeneca showcased its “self-healing” supply chain powered by enhanced digital twins. These systems ingest real-time sensor data, simulate potential disruptions (from weather to inventory shortages), and automatically reconfigure supply and distribution routes. For example, when geopolitical risks disrupted a critical route, the system rerouted and triggered procurement alternatives without human intervention, cutting decision lag from days to minutes.
FedEx’s “fdx powered by FedEx” platform combines digital twins, AI routing, and predictive packaging to pre-empt delivery issues. It captures 200+ data points per package—from weather conditions to customs paperwork—to reduce errors and costs. Over three years, FedEx saved $6 billion through these digital transformation initiatives. Its use of warehouse AI robots and IoT integrations has become a benchmark for operational agility.
Retail giants Tesco and Unilever are building digitally integrated ecosystems using GenAI and agentic tools to enhance demand prediction, automate inventory, and reduce packaging waste.
DHL spoke about building resilience through shared logistics intelligence, integrating AI insights across partners.
ESRI contributed by highlighting its cloud GIS service scaling at 50% annually, enabling spatial analysis of supply chain risks.
Through location intelligence, real-time asset tracking, and digital simulations, ESRI supports logistics partners in anticipating bottlenecks and navigating climate-related challenges. Across the board, the shift is from reactive management to predictive, autonomous systems, built to adapt in real-time.
Big Data’s Secret – Tackling the Carbon Cost of the Cloud
AI’s rapid adoption has a cost—one that’s largely invisible. Data storage now contributes more carbon emissions (4% globally) than the entire aviation sector. London Tech Week pulled no punches: 68% of stored enterprise data is effectively “junk,” yet consumes energy for storage, cooling, and backup. Unlike physical infrastructure, obsolete data isn’t decommissioned. It lingers in dark corners of servers, invisible but resource-hungry.
“68% of stored enterprise data is effectively “junk,” yet consumes energy for storage, cooling, and backup.”
New regulations are turning this risk into a boardroom priority. The EU’s Data Act (in force September 2025) and the UK’s Data Use and Access Act mandate efficient, interoperable, and sustainable data practices including reporting on carbon impact. The Corporate Sustainability Reporting Directive (CSRD) further demands clarity on how digital infrastructure supports (or undermines) environmental targets. This isn’t just about legal compliance. It’s a reputational and cost-saving issue.
These insights are a call to action for businesses: proactively assess data lifecycle policies, shut down redundant storage, and implement tiered cloud usage.
Smart data governance now means more than security and access—it’s about energy, efficiency, and ethics. Cleaning up “data waste” could help organisations cut IT costs while meeting ESG goals. And in an AI-powered future, leaner data = faster AI = lower emissions.
Climate Innovation – AI’s Unlikely Role as a Sustainability Engine
AI isn’t just a carbon culprit—it’s also a climate solution. One of the most exciting themes at LTW 2025 was how AI is powering decarbonisation. The UK’s Met Office is using AI to improve weather forecasting accuracy by 40%, giving governments and businesses earlier warnings and response options for extreme weather.
Schneider Electric applies AI to optimise industrial energy systems, reducing emissions in real-time across grid-connected buildings and data centres.
Start-ups like Monumo and Fugro are pushing boundaries further, modelling underground infrastructure and ocean-floor systems for energy safety and resilience. This has direct implications for utilities, transport networks, and renewable energy development. AI also plays a central role in battery storage optimisation, helping organisations buy less energy and reduce costs during peak periods.
For businesses, action plan is clear: embed AI across operations to predict energy loads, reduce waste, and protect infrastructure from climate volatility. Crucially, fostering a workforce that understands and embraces AI tools can amplify this impact. The message from the event: “AI won’t take your job, but someone using AI might.” Embracing this mindset means equipping teams to act as innovators, not just implementers—turning sustainability into a shared, scalable capability.
“AI won’t take your job, but someone using AI will”
Conclusion
The closing message of London Tech Week 2025 was nuanced and necessary: AI is simultaneously overhyped and underappreciated. The short-term buzz can obscure long-term breakthroughs. Agentic AI, digital twins, and data sustainability aren’t just buzzwords; they’re the tools & trends reshaping work, reducing emissions, and unlocking systemic productivity. But this future is fragile unless supported by deliberate leadership.
Investing in AI skills, building sovereign digital infrastructure, and embedding ethics into every layer of transformation emerged as strategic imperatives. Success is less about speed, and more about structure. Companies that move beyond experimentation and adopt AI as a core strategic capability, while remaining accountable, will define their industries for the next decade.
In short, AI is no longer a feature it’s the foundation. Businesses must now ask: not “what can AI do?” but “what must we become to lead in an AI world?” London Tech Week 2025 didn’t offer hype. It offered hard truths, hopeful strategies, and a window into what bold, human-centred digital transformation really looks like.
Authors:
Natalie Gyarfas, DX Consultant
Dave Crossley, DX Consultant
