Yorkshire AI Labs LLP
Sheffield, England, United Kingdom
14K followers
500+ connections
About
Articles by David
Activity
14K followers
-
David Richards MBE shared thisSpaceX’s valuation has reached a staggering $1.5 trillion. For perspective: 3 years ago: $137 billion 5 years ago: $74 billion 10 years ago: $10 billion If SpaceX goes public at a $1.75 trillion valuation, it would rank as the 7th largest company in the S&P 500. An extraordinary example of scale, execution, and long-term value creation. Credit to Charlie Bilello for the valuation details.
-
David Richards MBE shared this📡 The AI power problem, sized. The IEA puts global data centre electricity use at 415 terawatt hours in 2024. By 2030, it projects 945. Goldman Sachs projects a 165 per cent increase in data centre power demand by then. McKinsey & Company puts the global capital required to build the physical capacity at nearly 7 trillion dollars. Now bring that home. UK data centre electricity demand is forecast to roughly quadruple by 2030 on NESO’s numbers. At the upper end of industry estimates, AI data centres in this country could be consuming the equivalent of the electricity used by Yorkshire and the Humber. Homes. Factories. Hospitals. Railways. Schools. Street lights. All matched by buildings full of silicon throwing off heat in the dark. Some UK data centre developers are now being quoted ten-year grid connection waits. But this whole debate rests on one assumption: that the machine is fixed and the only answer is to feed it more power. A British company in Oxford may already be challenging that assumption. The full breakdown is in this week’s Sunday Signal: https://lnkd.in/ennZ5u5vLight, the Grid, and the British Company That Might Break the AI Energy TradeLight, the Grid, and the British Company That Might Break the AI Energy Trade
-
David Richards MBE reposted thisDavid Richards MBE reposted thisDavid Richards MBE's Sunday Signal this week is worth reading for the Lumai story alone. A British photonics company in Oxford has built a server that runs large language model inference on optical compute rather than silicon, and if even half of what they're describing reaches production, the assumptions Wall Street has priced into the AI energy trade are wrong by a wide margin. The thread I want to draw out for the scale-up CEOs in my network is further down, in his layoff tracker. Standard Chartered's CEO told investors this week that the bank was replacing what he called "lower-value human capital" with AI. He walked the phrase back within 48 hours, but the version that lands with investors and journalists is always the version said on the call, not the memo sent afterward. The reason this matters to a scale-up selling into enterprise is that the buyers across the table from you are now under direct board pressure to produce their own version of that story. Their CFO is being asked to model AI-driven operating leverage. Their CEO is being asked how the cost base looks in 18 months. The board is asking why headcount is still growing if the productivity claims about AI are real. Those pressures change what enterprise buyers want from you. The questions in the sales cycle are shifting from "does your product solve this workflow" to "what does my team look like in two years if I deploy this at scale," and the answer needs to be ready, and it needs to be in a language the buyer can carry back into their own board room without setting off the reaction Bill Winters just triggered. Most GTM motions I'm seeing in this market are not ready for that conversation. The product marketing tells a productivity story. The sales team tells an ROI story. Neither is calibrated to the conversation the buyer is having internally, which is about workforce, accountability, and the awkward question of what the company is for if AI does most of the work. The scale-ups that build a clear and defensible answer to that question will win the next 18 months of enterprise budget. The ones that don't will lose deals they should have won. David's piece is worth your Sunday. The Lumai story is the headline. The Standard Chartered moment is the one I'd sit with the longest.Light, the Grid, and the British Company That Might Break the AI Energy TradeLight, the Grid, and the British Company That Might Break the AI Energy Trade
-
David Richards MBE shared thisTwo weeks ago I sat in the Royal Academy of Engineering watching a quiet Oxford physicist make the most aggressive claim I have heard in British deep-tech this year. Ninety-five per cent less energy than an NVIDIA GB200 by 2029. Ninety per cent lower cost. A British company doing AI inference with light instead of electrons. I followed up. The result is this week's The Sunday Signal. The company is Lumai. The CEO is Dr Xianxin Guo. The Iris Nova server, launched last month, is currently running Meta's Llama 8B and 70B on optical compute. The only system in the world doing it. Three lines from the interview that I cannot stop thinking about. "Cost per token. That is the only metric that matters." "Jevons' paradox is real, and we will see more AI used, not less." "Where the company is built is the question that matters, and the answer is the UK." His ask of the British state was not subsidy. It was procurement. A committed government anchor purchase of Lumai hardware into the AI Research Resource, this financial year, or watch the next generation of AI compute be created outside of the UK again. Britain has watched Google DeepMind, Magic Pony, VocaliQ, Latent Logic and Oxford Ionics walk out of the door. Light is the next one. Buy the thing, or read the press release. 📡 The full issue: https://lnkd.in/eWXTnQBk 📡 The full interview with Dr Xianxin Guo: https://lnkd.in/exVGx33tThe full Sunday Signal interview with Dr Xianxin Guo, CEO of LumaiThe full Sunday Signal interview with Dr Xianxin Guo, CEO of Lumai
-
David Richards MBE shared thisEveryone is asking the wrong question about AI and energy. The question is not just: how do we build enough power stations, data centres and grid connections to feed the machine? The better question is: what if the machine itself changes? This week’s The Sunday Signal is built around my interview with Dr Xianxin Guo, CEO of Lumai, an Oxford spinout building optical compute for AI inference. Their claim is extraordinary: 90% less energy at the rack appliance level, 10% of the cost, and more AI tokens from the same power budget. If they are right, the implications go far beyond data centres. It changes the grid debate. It changes the economics of AI. It challenges the trillion-dollar “AI energy trade”. And it raises the British question we keep getting wrong: when we invent the future, do we own it, or do we watch someone else buy it? Google DeepMind. Magic Pony. VocaliQ.Latent Logic. Oxford Ionics. Britain has seen this film before. Lumai’s answer is simple: not subsidy, procurement. If the British state wants sovereign AI, it needs to become a serious first customer for British AI hardware before someone else becomes the owner. This week’s issue is about light, the grid, Lumai, and whether Britain finally learns to keep the companies it creates.Light, the Grid, and the British Company That Might Break the AI Energy TradeLight, the Grid, and the British Company That Might Break the AI Energy Trade
-
David Richards MBE shared this📡 This week in The Sunday Signal. A British company in Oxford says it can do AI inference for ninety per cent less energy than the hardware the entire industry currently runs on. Light, not electrons. I sat down with Dr Xianxin Guo, CEO of Lumai, this week, after watching him pitch at the Royal Academy of Engineering a fortnight ago. The Iris Nova server he has just launched is the only system in the world running billion-parameter LLMs on optical compute, currently doing real-time inference on Meta's Llama 8B and 70B. His ask of the British state was not subsidy. It was procurement, this financial year. The full breakdown is in this week's LinkedIn newsletter. Physics, power, the trillion-dollar energy trade that may be priced wrong, and the British procurement question that keeps a company like this here rather than letting another one walk to the US. → Read the full issue on LinkedIn Britain has watched Google DeepMind, Magic Pony, VocaliQ, Latent Logic and Oxford Ionics walk out of the door. Light is the next one. Buy the thing, or read the press release. Subscribe on LinkedIn https://lnkd.in/esttJkCj
-
David Richards MBE reposted thisDavid Richards MBE reposted thisThe smartest deal is sometimes the one you walk away from: David Richards https://trib.al/7eNZgTMThe smartest deal is sometimes the one you walk away from: David RichardsThe smartest deal is sometimes the one you walk away from: David Richards
-
David Richards MBE shared thisIndeed a great evening. Have to say though it would have been better had the governer actually answered the real exam question “how do we get out of this” rather than the Sunday supplement guide to AI…David Richards MBE shared thisGreat to meet the Governor of the Bank of England, Andrew Bailey, and so many great friends and business leaders at the Cutlers Feast, at the Cutlers Hall in Sheffield. A big thank you to British Chambers of Commerce board member and Sheffield Chambers CEO Louisa Harrison-Walker OBE and team for making me so welcome. Great to see David Richards MBE Morton Bell Dan Fell MBE and more!
-
David Richards MBE shared thisA bad marriage can at least end in divorce. A bad deal, a bad contract, a bad investor? Much harder. That is the argument in my Yorkshire Post column today. Founders are often told that raising money is the victory. It is not. The wrong capital can cost you the company. I learned that lesson the hard way. The cheque was welcome. The terms were not. Control did not disappear in one dramatic boardroom coup. It went gradually, politely, until the company was being run by people who had not built it. Equity is not control. If you do not control the board, you do not control the company. This matters in the North now more than ever. Founders in Leeds, Sheffield and Manchester are being offered capital that was unavailable a decade ago. That is progress. But gratitude is a ruinous negotiating position. The sinking feeling is data. Read my column in today’s Yorkshire Post Business. And in this week’s The Sunday Signal I’m looking at a very different constraint on ambition: the AI energy problem, the grid, and the British company trying to change the trade-off between computing power and electricity demand. Subscribe here: https://lnkd.in/eTgwVvmY
-
David Richards MBE liked thisDavid Richards MBE liked thisThis week a junior showed me an amazing vibe coded product he had created. It got me thinking. Stuart Nuttall Bradley Wilson David Richards MBE
-
David Richards MBE liked thisDavid Richards MBE liked thisDavid Richards MBE's Sunday Signal this week is worth reading for the Lumai story alone. A British photonics company in Oxford has built a server that runs large language model inference on optical compute rather than silicon, and if even half of what they're describing reaches production, the assumptions Wall Street has priced into the AI energy trade are wrong by a wide margin. The thread I want to draw out for the scale-up CEOs in my network is further down, in his layoff tracker. Standard Chartered's CEO told investors this week that the bank was replacing what he called "lower-value human capital" with AI. He walked the phrase back within 48 hours, but the version that lands with investors and journalists is always the version said on the call, not the memo sent afterward. The reason this matters to a scale-up selling into enterprise is that the buyers across the table from you are now under direct board pressure to produce their own version of that story. Their CFO is being asked to model AI-driven operating leverage. Their CEO is being asked how the cost base looks in 18 months. The board is asking why headcount is still growing if the productivity claims about AI are real. Those pressures change what enterprise buyers want from you. The questions in the sales cycle are shifting from "does your product solve this workflow" to "what does my team look like in two years if I deploy this at scale," and the answer needs to be ready, and it needs to be in a language the buyer can carry back into their own board room without setting off the reaction Bill Winters just triggered. Most GTM motions I'm seeing in this market are not ready for that conversation. The product marketing tells a productivity story. The sales team tells an ROI story. Neither is calibrated to the conversation the buyer is having internally, which is about workforce, accountability, and the awkward question of what the company is for if AI does most of the work. The scale-ups that build a clear and defensible answer to that question will win the next 18 months of enterprise budget. The ones that don't will lose deals they should have won. David's piece is worth your Sunday. The Lumai story is the headline. The Standard Chartered moment is the one I'd sit with the longest.Light, the Grid, and the British Company That Might Break the AI Energy TradeLight, the Grid, and the British Company That Might Break the AI Energy Trade
Experience & Education
View David’s full experience
By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.
Volunteer Experience
-
Co-Founder & President
David & Jane Richards Family Foundation
- Present 8 years 7 months
Children
We founded the David & Jane Richards Family Foundation to educate, empower, and improve the lives of children. We will achieve this through hands on programs, and targeted assistance. We aim to encourage children to fulfill their potential and make a positive impact on the world around them.
Publications
-
Examining Tech Solutions To Natural Disasters
Teccrunch
See publicationWhen Cyclone Pam hit Vanuatu, it struck a very vulnerable target. The Pacific nation is considered one of the world’s poorest and is ranked by the United Nations’ WorldRiskIndex as the most disaster-prone country.
-
Building A Data-Driven White House
Techcrunch
See publicationWorking in government IT doesn’t mean what it used to. In the past, computers were only used to carry out simple office work or as a substitute for pens and paper. And now, organizations are fast realizing they require new leadership to get the most out of their data. This is more than just another fickle trend; a quick search on LinkedIn reveals that “data scientist” now appears in roughly 36,000 help-wanted posts.
-
Healthcare’s Big Data Opportunity
Techcrunch
See publicationThe hospitals where my grandchildren are born will be unrecognizable to those we know today; they’ll be safer, cleaner and more efficient. And what is considered world class in 2014 will tomorrow be viewed with the same disdain we reserve for medical practice in the Dark Ages.
Patents
-
Method for managing proposals in a distributed computing system
Issued US 10,481,956
In one embodiment of the present invention, a replicated state machine comprises a proposal manager, an agreement manager, a collision/back-off timer and a storage reclaimer. The proposal manager facilitates management of proposals issued by a node of a distributed application for enabling coordinated execution of the proposals by all other nodes of the distributed application. The agreement manager facilitates agreement on the proposals. The collision/back-off timer precludes repeated…
In one embodiment of the present invention, a replicated state machine comprises a proposal manager, an agreement manager, a collision/back-off timer and a storage reclaimer. The proposal manager facilitates management of proposals issued by a node of a distributed application for enabling coordinated execution of the proposals by all other nodes of the distributed application. The agreement manager facilitates agreement on the proposals. The collision/back-off timer precludes repeated pre-emptions of rounds in attempting to achieve agreement on the proposals. The storage reclaimer reclaims persistent storage utilized for storing at least one of proposal agreements and the proposals.
Other inventors -
-
Distributed computing systems and system components thereof
Issued US 8,364,633
In one embodiment of the present invention, a replicated state machine comprises a proposal manager, an agreement manager, a collision/back-off timer and a storage reclaimer. The proposal manager facilitates management of proposals issued by a node of a distributed application for enabling coordinated execution of the proposals by all other nodes of the distributed application. The agreement manager facilitates agreement on the proposals. The collision/back-off timer precludes repeated…
In one embodiment of the present invention, a replicated state machine comprises a proposal manager, an agreement manager, a collision/back-off timer and a storage reclaimer. The proposal manager facilitates management of proposals issued by a node of a distributed application for enabling coordinated execution of the proposals by all other nodes of the distributed application. The agreement manager facilitates agreement on the proposals. The collision/back-off timer precludes repeated pre-emptions of rounds in attempting to achieve agreement on the proposals. The storage reclaimer reclaims persistent storage utilized for storing at least one of proposal agreements and the proposals.
Other inventors -
Honors & Awards
-
Hon LittD (Honorary Doctorate)
University of Sheffield
The University of Sheffield confers honorary degrees (or degrees honoris causa – as a ‘mark of honour’) on people who have given distinguished service or brought distinction to the University, the City of Sheffield or the region
-
Most Excellent Order of the British Empire: M.B.E.
Her Majesty The Queen
For services to the Information Technology Sector and to Young People, particularly during Covid-19.
-
Listed Company Entrepreneur of the Year.
2018 Enterprise Awards in association with the WCIT
The Enterprise Awards in association with the WCIT is celebrating and recognizing the very best of the UK’s technology entrepreneurial talent with ‘The Oscars of the Technology Industry’.
Uniquely focused on the founder rather than the businesses themselves, the Enterprise Awards boasts previous winners that have achieved prominence on both private and public markets. The judging panel will be looking for examples of vision, growth, innovation, funding, use of capital, use of resources…The Enterprise Awards in association with the WCIT is celebrating and recognizing the very best of the UK’s technology entrepreneurial talent with ‘The Oscars of the Technology Industry’.
Uniquely focused on the founder rather than the businesses themselves, the Enterprise Awards boasts previous winners that have achieved prominence on both private and public markets. The judging panel will be looking for examples of vision, growth, innovation, funding, use of capital, use of resources, strategy, execution, ambition and determination to succeed. -
Hon. DSc. (Honorary Doctorate)
Sheffield Hallam University
Honorary Doctorate
-
Innovative Company of the Year
City AM
The City A.M. Awards 2013
Languages
-
English
Native or bilingual proficiency
-
French
-
Organizations
-
Tech London Advocates
-
Recommendations received
1 person has recommended David
Join now to viewWebsites
- Personal Website
-
Http://www.DJRFF.org
Other similar profiles
Explore more posts
Explore collaborative articles
We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.
Explore More