Saturday, July 19, 2025

DV25003 Data Centres V01 190725

 The Times Copyright acknowledge.

UK plays catch-up in the data centre race
America is rolling out giant facilities at pace to feed the AI boom. Can Britain compete, asks Emma Powell

On a former Second World War airbase in Cambridgeshire plans are afoot to construct one of Europe’s largest data centres, providing the computing energy needed to cope with increasingly powerful artificial intelligence technology.

The positioning of the 100-acre scheme, roughly the size of about 75 football pitches, outside the traditional west London data centre heartland is as much to do with power access as its proximity to the advanced computing and research hubs of Oxford and Cambridge.

“The grid capacity in the Slough area and west London area is very constrained, so getting new connections to the grid is very limited,” said Niall Brunker, co-head for the UK at Icona Capital, the investment firm that is part of the joint venture behind the scheme.

The development, which will have a capacity of 330 megawatts, is one of several energy-hungry data centre projects being planned across the UK in the next few years, designed to power the AI boom. American technology giants and private capital are pouring billions into building the facilities.

However, the UK remains far behind the US in its data centre rollout. Total installed compute capacity in North America stands at 20 gigawatts (GW), according to data from Cushman & Wakefield, the real estate services group.

Even after taking into account projects in the pipeline, the capacity in the UK market stood at just 3.6GW at the end of last year.

Higher energy costs and an aged grid put Britain at risk of falling even behind further, industry insiders have warned.

In January, Sir Keir Starmer backed an action plan to make Britain “one of the great AI superpowers” and an “AI maker” rather than an “AI taker”.

Data centres have been designated as critical infrastructure, to be provided with additional protection from cyberattacks and IT outages in an effort to bolster investor confidence in building the facilities in the UK.

It is the first such designation for almost a decade after defence and space were awarded the status in 2015.

Yet the UK’s data centres are dwarfed by the so-called hyperscale facilities being rolled out in the US.

The AI arms race intensified this week after Mark Zuckerberg, the billionaire founder and chief executive of Meta Platforms, said that his company — which owns Facebook, WhatsApp and Instagram — would spend hundreds of billions on rolling out “multi-gigawatt” data centres.

Amazon is planning to build 30 data centres across a vast stretch of farmland in Indiana.

Amazon Web Services, the cloud computing arm of the ecommerce giant, has said it will invest £8 billion in the UK over the next five years on building, operating and maintaining data centres. Google is set to open its first UK data centre in Hertfordshire later this year. Blackstone, the world’s largest private equity firm, plans to spend $10 billion on transforming the site of Britishvolt’s ill-fated attempt to build a gigafactory in Northumberland.

As demand for the facilities — which are filled with thousands of specialised computer chips required to handle more sophisticated AI workloads — has risen sharply, the land grab for sites in the UK with access to reliable power and water supply has intensified.

Data centres require large amounts of energy to run and keep cool.

“The data centre providers want to make sure they have two good suppliers of power and a good steady source of water, and those are hard to come by in the same place at the same time”, said Alex McMullan, chief international technology officer at Pure Storage, a technology provider that is assisting Meta with its global data centre rollout.

“What that really means is we are a little bit behind the curve here.”

A typical AI-focused data centre consumes as much electricity as 100,000 households, but the hyperscalers under construction will consume 20 times as much, according to the International Energy Agency, which represents member nations of the Organisation for Economic Cooperation and Development.

In a report released in November, the national systems operator in the UK projected that demand for electricity to power data centres would grow fourfold by the end of the decade, as requirements from electric vehicles and home heating also rise sharply.

However, the UK’s energy grid is already under pressure. The design and capacity of the nation’s electricity network, which has evolved around coal-fired assets, has failed to keep pace with the rapid expansion of renewables, notably the wind power heavily concentrated in Scotland and on the east coast of England but needs to be delivered to the south of the country.

So-called zombie projects have also led to lengthy grid connection queues, with some projects waiting up to 15 years to connect.

The delays threaten to jeopardise Labour’s aim to rapidly expand the build-out of clean power in an effort to decarbonise the UK’s energy system by the end of the decade.

In April, the government put forward plans to reform the grid connection process, prioritising clean energy assets and kicking out the zombie schemes in order to unblock the queue.

In America, Amazon led a $700 million funding round in X- Energy, a nuclear power developer, last year. Microsoft has struck a 20-year power purchase agreement that will bring the Three Mile Island nuclear plant in Pennsylvania back online.

In the UK, nuclear power has dwindled to about 14 per cent of the UK’s electricity mix, down from about a quarter in the late 1990s. Efforts to revive the industry have been drawn out, with Hinkley Point C, the only nuclear power plant currently being built, running years behind and billions of pounds over budget. Small modular reactors, which can be built in a factory and assembled offsite have been put forward as a more expedient option to power data centres, but the first of those is not expected to come online until the mid-2030s.

The power supply issue has become more acute as much of the existing data centre infrastructure predates the AI boom inspired by the breakthrough of ChatGPT in 2022. “We’re already seeing workload demands from a power perspective that are three, four, five times what a data centre has been designed for,” McMullan said.

Britain’s gas distribution networks have claimed that data centre operators are seeking alternatives, having received more than 30 inquiries related to data centre connections in the past six months alone. However, any plans by data centres to connect to Britain’s gas pipelines and build their own gasfired power plants would threaten to put Britain’s AI ambitions on a collision course with its decarbonisation goals.

The government has said it will create “AI growth zones” where projects have better access to the grid.

These zones will be crucial for dealing with land and energy scarcity issues, said Karl Havard, chief operating officer at Nscale, and could prove “a catalyst for getting renewable energy into the grid”. Nscale plans to spend £2 billion on building data centres in the UK.

Rising geopolitical tensions could give UK-domiciled cloud storage providers like Nscale an edge, Havard said, as “the need to take control of the infrastructure within the sovereign territory” becomes more important.

Earlier this year, European groups raised concerns that the US could use the continent’s reliance on cloud technology provided by US technology companies as leverage in trade talks.

Proximity to the capital means Slough and the broader area just west of London has become the world’s second-biggest data centre hub.

Of all the data transferred in activities, from online shopping and the use of social media to playing video games and streaming TV programmes, by people in London and surrounding areas, nearly half will go through the Slough trading estate.

“It is becoming increasingly untenable to build in London,” Kevin Restivo, head of European data centre research at CBRE, the real estate services company, said. He added that people are looking to build data centres in locations they wouldn’t have “dreamed of putting [them] on” even four years ago.

Thursday, July 17, 2025

DV25002 New UK Supercomputer V01 180725

 AI supercomputer primed for a quantum leap


Mark Sellman - Technology Correspondent 

Simon McIntosh-Smith with the £225m taxpayer-funded Isambard-AI in Bristol

Britain’s most powerful AI supercomputer has been switched on to help the nation’s tech sector and accelerate breakthroughs in drug discovery, climate change and material science.

Isambard-AI, a £225 million taxpayer-funded facility in Bristol, is central to the government’s aim of building “sovereign” capacity in artificial intelligence.

Named after Isambard Kingdom Brunel, the engineer who designed the city’s Clifton Suspension Bridge, the supercomputer is the 11th most powerful in the world and sixth in Europe.

It has the capacity of 600,000 laptops and uses the energy of 16,000 kettles, but its cooling system means it is the world’s fourth greenest supercomputer.

Some of the energy generated at Isambard, which was built in less than a year on the site of a car park, will go towards heating local homes.

Labour recently committed itself to building a £750 million supercomputer called Archer2 at the University of Edinburgh, which will be the UK’s biggest when it goes live in 2027.

Isambard is a dedicated AI facility, together with the smaller Dawn supercomputer in Cambridge, but Archer2 will do other scientific work. 

The government is planning to spend another £1 billion to expand AI-dedicated computer capacity twentyfold by 2030, having spent £350 million on the Bristol and Cambridge facilities.

Professor Simon McIntosh-Smith, director of the Bristol Centre for Supercomputing, said: “Isambard is big enough to train ChatGPT — the latest version — in the UK. We could not do that before. We would know exactly what was run, what data went into it and doing it all here. That puts us in a much safer, stronger position.”

Britain still lags behind other European countries, such as Germany, Italy, Switzerland and Finland. The US is said to be home to 75 per cent of the world’s AI supercomputers, and China second.

David Hogan, a senior executive at Nvidia, said a challenge for the UK was the price of energy. “It’s cheaper because of renewables in the Nordics, solar in Spain, and then there’s a mix in the middle,” he said. “France is fortunate because it’s stuck with nuclear. There are approaches to address that [in the UK] like small nuclear reactors.”

Isambard has piloted 60 research projects, including a ChatGPT-style large language model called BritLLM, as well as AI-driven drug discovery and advanced climate modelling.

University College London researchers are using it to develop AI imaging for prostate cancer screening and Bristol academics are designing a system to help dementia patients at home.

Jon Lees, a lecturer in bioinformatics at Bristol University, has been using the system to predict how proteins interact with each other in patients with Alzheimer’s. “It’s been huge for us and I think it’s really democratising for UK science,” he said. “It enables you to try out ideas faster. So you can sort of say ‘I’m just going to give this a go’, and you can get that preliminary data.”

Researchers will be able to apply to the Department for Science, Innovation and Technology and the funding agency UK Research and Innovation for access. Isambard, which was built by Hewlett Packard Enterprise using 5,400 GH200 chips from Nvidia, is funded for five years.

Thursday, July 10, 2025

DV25001 Nanotechnology V01 100725

 K. Eric Drexler is an American engineer and author best known as a pioneering advocate and theorist of molecular nanotechnology—the idea of building materials and machines at the scale of individual atoms and molecules.


Key Facts:

Full Name: Kim Eric Drexler

Born: April 25, 1955

Education:

B.S. in Interdisciplinary Sciences from MIT

M.S. in Astro/Aerospace Engineering from MIT

Ph.D. in Molecular Nanotechnology from MIT (the first ever awarded in this field)


Major Contributions:


1. Founding Nanotechnology as a Field


Drexler is widely regarded as the father of molecular nanotechnology. His early work laid the theoretical foundation for the idea of atomically precise manufacturing.


2. Influential Books:

“Engines of Creation: The Coming Era of Nanotechnology” (1986)

A popular science book that introduced the concept of nanotechnology to a broader audience.

He proposed ideas like assemblers—molecular machines that could build complex objects atom by atom.

Also raised concerns about “grey goo,” a hypothetical scenario where self-replicating nanobots run amok.

“Nanosystems: Molecular Machinery, Manufacturing, and Computation” (1992)

A highly technical, peer-reviewed book that provides detailed theoretical models for building molecular machines.

Helped move nanotechnology from science fiction into serious scientific discourse.


3. Controversy and Criticism

His ideas, especially regarding self-replicating nanobots and the “grey goo” scenario, were criticized by some scientists, including Nobel laureate Richard Smalley, who argued that Drexler underestimated the complexity of chemistry at the nanoscale.

The debate spurred more detailed examination of what is feasible in molecular engineering.


4. Foresight Institute

Drexler co-founded the Foresight Institute in 1986 with Christine Peterson.

The institute aimed to educate the public and policymakers about the benefits and dangers of nanotechnology.


Later Work:


In recent years, Drexler has shifted his focus to Artificial Intelligence and advanced technologies more broadly. He has worked with organizations like the Future of Humanity Institute at Oxford University, researching safe and beneficial long-term uses of advanced technologies.


Legacy:

Drexler’s vision helped establish nanotechnology as a field of legitimate scientific inquiry.

Though mainstream nanotechnology today focuses more on materials science and less on molecular assemblers, his theoretical groundwork remains influential in long-term technology foresight and policy.


Would you like a summary of his debate with Richard Smalley or more on his later AI-related work?

DV26001 Book Summary - Critical Theory and the Digital. V01 220126

  Critical Theory and the Digital. David M.Berry (2014). Bloomsbury Academic ISBN 9781501310966 Introduction   In the Introduction to Critic...