Circul8 at London Tech Week

Our wonderful KTP Associate Bilal Khan had a change of pace this month when he attended London Tech Week on behalf of Techbuyer. In a busy week that included a conference presentation to the 7th IEEE Workshop on Smart Circular Economy [AW1] and a panel presentation to the Government Digital Sustainability Alliance, Bilal managed to pick up some excellent information from start-ups and keynote speakers at the event. Here are his highlights:

AI as National Infrastructure  

I was lucky enough to be in the audience when Prime Minister Keir Starmer and Nvidia President Jason Huang announced plans for growing AI capability in the UK. The most salient points for me were:

·         That AI could be “the great equaliser”, praising UK talent in this area.

 

·         There is a plan to up-skill 7.5 million workers in AI by 2030 via a £185 m education package co-developed with NVIDIA.

 

·         The infrastructure gap that exists may be resolved by Government pledges of £1 billion to expand sovereign compute capacity 20-fold.

 

·         Liquidity, an AI-driven private-credit firm, opens a 5 000 sq ft European HQ at 127 Charing Cross Road, Soho; having already invested £350 m in 12 UK scale-ups, it will inject a further £1.5 billion over the next five years.

AI Development Coming Thick and Fast

As you would imagine, talks on new AI product launches and updates are coming thick and fast. One of the stand-outs was when Arthur Mensch (Co-founder & CEO, Mistral AI) launched open-source reasoning model Magistral (both able to run on a laptop and multilingual). Mensch says most firms will start with giant models then shrink them and warns uneven AI uptake could widen global inequality. With this in mind, he urges Europe to build its own cloud and chip stack.

Another update I found interesting was on Agentic AI. Darren Hardman (CVP & UK CEO, Microsoft) announced that Barclays will scale Microsoft 365 Copilot to 100 000 employees worldwide, making Copilot the single “Colleague AI Agent” inside its productivity hub.

The public sector are also keen to capitalise on the agentic opportunity – he showed pilots in the NHS and UK Civil Service where Copilot shifts routine admin to AI, freeing staff for frontline work. He also covered the ethics side of AI, pledging that Microsoft will tackle security, privacy and inclusion head-on, aiming for “a fair AI economy where everyone moves forward together.

 

How AI Incorporated Products can Scale  

Having worked for the last two years on a decision support system[AW2]  for real time maintenance on laptops, I was interested in businesses who had been able to scale similar products.

Angelina Menchon Dyer, Pawel Kaminski, Maarten Ectors, Mario Tomic gave a very useful StartUp Product & Tech Q&A, as part of the StartUp-to-ScaleUp Club. Their advice included the following tips for both myself and the Circul8 product as it grows:

·         Validate the smallest testable slice (MVP) first; every extra feature multiplies scaling pain.

·         Use low-/no-code plus automated testing to iterate fast without ballooning the  team.

·         Make observability (logs, traces) part of the first sprint—retro-fitting costs 3-5× more

I also spoke with the team behind BatteryCheck, a predictive battery monitoring startup using AI to optimise battery-dependent systems. Their solution tracks current, voltage, temperature and energy throughput in time series to provide real-time health diagnostics and early failure alerts.

They support devices across vehicles, drones, power tools, and renewable energy systems. Their AI models analyse time-series telemetry including current, voltage, temperature, and energy throughput core metrics needed for accurate battery health forecasting. The system also ingests any additional data available from the device, enabling broader trend analysis and lifespan optimisation.

It's great to see more products that are using AI to predict performance based on sensor-based information. The feed from a drone or power tool is different from a laptop but the principle is the same. It definitely gives confidence that this type of approach will be the norm in the future!