• The GPU
  • Posts
  • Issue #9: The UK’s AI Push, Billions in NeoCloud Investments, & AI Diffusion

Issue #9: The UK’s AI Push, Billions in NeoCloud Investments, & AI Diffusion

Feat. The UK, Biden, Vantage Data Centres, Nscale, Kyndryl, Applied Digital, CoreWeave, Nvidia, Mistral, Microsoft, TSMC, and AWS

Author

Ben Baldieri
January 18, 2025 • Estimated Reading Time: 11 minutes

Week two of 2025 is trying to compress yet another decade of progress into just seven days.

The UK unveiled a £14B AI Opportunities Action Plan, Biden doubled down on export controls while signing an Executive Order to secure U.S. leadership in AI infrastructure, and NeoCloud providers announced billions in investments to scale globally. Meanwhile, TSMC enabled 4nm chips in Arizona, Mistral changed the open-source landscape, AWS opened a new region in Thailand, and Microsoft revealed AI breakthroughs from materials science to simplified enterprise adoption.

Every week, I break down the moves shaping GPU compute, AI infrastructure, and the data centres that power it all.

Here’s what’s inside this week:

  • The UK’s £14B AI Opportunities Action Plan to turbocharge growth.

  • Biden’s new chip export rules and a sweeping AI Executive Order.

  • CoreWeave, Nscale, and Applied Digital make major plays.

  • Mistral’s Codestral simplifies AI deployment with open-source innovation.

  • Microsoft’s AI meets materials science.

  • TSMC’s Arizona expansion amid Taiwan’s policy shifts.

  • AWS’s new region in Thailand, and Microsoft reveals Core AI Platform.

Let’s dive in.

The GPU Audio Companion Issue #9

Want the GPU breakdown without the reading? The Audio Companion does it for you—but only if you’re subscribed. Fix that here.

The UK’s AI Action Plan Unleashes £14B in Investment

The UK is all-in on AI.

Keir Starmer, UK Prime Minister, announced the AI Opportunities Action Plan. The bold initiative unveiled £14B in private investment and sweeping reforms to accelerate infrastructure development. The plan features AI Growth Zones to fast-track data centre builds, a twentyfold increase in public compute capacity, and a National Data Library for public data innovation. And with massive commitments from Vantage Data Centres, Nscale, and Kyndryl and the promise of over 13,000 jobs, the UK is joining the AI party in a big way.

Why this matters:

  • AI Growth Zones tackle planning and energy bottlenecks, enabling faster infrastructure scaling.

  • Nscale’s Essex project ensures sensitive AI workloads stay local, reducing reliance on foreign infrastructure.

  • The UK’s comprehensive plan challenges the US and EU in attracting AI investment.

Biden Tightens AI Chip Export Rules, Nvidia Counters

The U.S. doubles down on AI security with tighter chip export rules, prompting Nvidia to push back.

Biden’s administration has introduced stricter export controls for advanced GPUs, targeting nations deemed security risks. These measures aim to curb geopolitical adversaries’ use of bleeding-edge hardware and risk disrupting global supply chains. Nvidia has, naturally, passionately warned that overly restrictive policies could stifle innovation while advocating for balanced solutions to preserve collaboration.

NVIDIA Statement on the Biden Administration’s Misguided 'AI Diffusion' Rule

For decades, leadership in computing and software ecosystems has been a cornerstone of American strength and influence worldwide. The federal government has wisely refrained from dictating the design, marketing and sale of mainstream computers and software — key drivers of innovation and economic growth. The first Trump Administration laid the foundation for America’s current strength Read Article

blogs.nvidia.com/blog/ai-policy

Whether the restrictions remain in place with the incoming Trump administration remains to be seen.

Why this matters:

  • Restricted regions face limited access to downgraded chips, while some Western nations have unlimited access to the cutting edge.

  • If the path to AGI really is just a case of throwing more compute at the problem, this could be the edge the US needs to get there first.

  • There are likely heated conversations happening inside Nvidia’s walls right now as to what this means for their global accelerated computing paradigm.

Read about the new rules and Nvidia’s pushback below:

CoreWeave, Nscale, and Applied Digital Scale Up

NeoCloud players are stepping up, with billions in investments and multiple new data centres.

CoreWeave has launched two UK-based data centres powered by Nvidia Hopper GPUs, marking its first expansion outside North America.

Meanwhile, Nscale is committing $2.5B to build a sovereign AI data centre in Essex by 2026, part of a broader effort to localise sensitive AI workloads.

Applied Digital, backed by up to $5B from Macquarie, is also scaling its AI-focused data centre business to support generative AI and machine learning demands.

Why this matters:

  • Massive investment plans and debt facilities indicate rising confidence in GPU compute-related services as a significant long-term growth area.

  • CoreWeave is arguably the most successful NeoCloud provider, but the scale of Nscale’s plan in a new market for CoreWeave could provide some healthy competition.

  • Macquarie’s involvement with Applied Digital is also a pretty strong vote of confidence in the longevity of their business model (and their recent hires).

Read the full stories below:

Mistral’s Codestral and Microsoft’s AI for Materials Science

Mistral and Microsoft are taking bold steps to simplify AI adoption and accelerate breakthroughs.

Mistral’s new open-source framework, Codestral, streamlines the deployment of large language models (LLMs) and generative AI systems. By optimising compute resources and integrating seamlessly with cloud providers like AWS and Azure, Codestral lowers the barrier to entry for enterprises adopting AI.

Meanwhile, Microsoft’s research team is using AI to accelerate materials discovery, combining large language models with domain-specific datasets. This initiative focuses on developing and optimising materials for industries like renewable energy, electronics, and healthcare, aiming to cut the traditionally lengthy R&D cycles in materials science.

Why this matters:

  • Codestral reduces the operational complexity of deploying LLMs, letting enterprises focus on outcomes, not infrastructure.

  • Microsoft’s MatterGen raises interesting questions about the future of AI usage in energy storage, healthcare, and advanced manufacturing, and driving novel innovations in these industries.

  • Both innovations lower barriers for smaller enterprises to leverage advanced AI, expanding the reach of these technologies.

Read the full stories on Mistral and Microsoft below:

Taiwan Lifts Chipmaking Restrictions as TSMC Scales in Arizona

Taiwan is reshaping its chipmaking strategy as TSMC ramps up production in the U.S.

The Taiwanese government has lifted restrictions on foreign chipmaking as TSMC begins 4nm production at its Arizona plant. This is TSMC’s first advanced-node factory outside Taiwan, serving U.S. hyperscalers and enterprise clients. The move diversifies supply chains while preserving Taiwan’s leadership in semiconductor innovation.

Why this matters:

  • The Arizona plant reduces the U.S.’s reliance on Asian supply chains for critical AI hardware.

  • Taiwan’s policy change allows TSMC to expand abroad while maintaining domestic R&D strength.

  • Given the broader geopolitical picture of the USA/Taiwan/PRC threeway, this is likely part of a complex, ongoing exercise in risk management.

Biden’s AI Executive Order Targets Leadership in Infrastructure

The U.S. is doubling down on AI infrastructure with sweeping policies enabling GW+ scale developments on federal lands.

President Biden’s Executive Order sets the stage for scaling U.S. AI infrastructure with a focus on massive compute power and sustainability. It includes funding for AI-specific supercomputers and programs to modernise power delivery for data centres. Crucially, the order allows for data centre projects on federal lands, enabling GW-scale developments to meet the energy-intensive demands of AI workloads while integrating clean energy sources like wind, solar, and small modular reactors (SMRs).

Why this matters:

  • Opening federal lands to data centre projects allows for unprecedented scale.

  • Integrating renewable energy and SMRs aligns infrastructure growth with long-term environmental goals.

  • This move, along with the new chip restrictions, cements the US as the premier location for AI infrastructure.

AWS Expands in Thailand, Microsoft Launches Core AI Platform

AWS expands its reach in Southeast Asia, while Microsoft simplifies enterprise AI.

AWS has launched its Asia Pacific (Thailand) Region, providing low-latency cloud access to one of Southeast Asia’s fastest-growing economies. Meanwhile, Microsoft introduced its CoreAI Platform, offering pre-trained models, agent orchestration, no-code tools, and Azure integration to accelerate enterprise AI adoption.

Why this matters:

  • AWS’s Thailand region gives it a new foothold in Southeast Asia’s cloud market.

  • Microsoft’s CoreAI platform and tools massively lower the barriers that enterprises face when integrating AI into workflows.

  • Given that SE Asia is doubling down on sovereign capabilities and enterprise adoption is what everyone is waiting for, both moves give a strong indication of where future battlegrounds are likely to be.

Read the full stories from AWS and Microsoft below:

The Rundown

I’ve said it before, and I’ll likely say it again: what a week.

The UK announced its £14B AI Opportunities Action Plan, aiming to turbocharge AI development with Growth Zones, sovereign AI data centres, and massive private investment. Meanwhile, Biden tightened global chip export rules, triggering strategic pushback from Nvidia, while also signing an Executive Order to cement U.S. AI infrastructure leadership. CoreWeave, Nscale, and Applied Digital collectively revealed billions in new projects, signalling they’re scaling aggressively to challenge hyperscalers. Mistral’s open-source Codestral framework simplifies AI deployment, while Microsoft is leveraging AI to accelerate materials discovery. Taiwan lifted chipmaking restrictions as TSMC scaled up 4nm production in Arizona, strengthening U.S. chip supply chains and reducing dependence on Asia. AWS launched a new region in Thailand to capture Southeast Asia’s booming digital economy, and Microsoft unveiled its Core AI Platform to drive enterprise AI adoption.

Ladies and gentlemen, please keep your arms and legs inside the ride at all times.

The AI race is accelerating and fast.

Keep The GPU Sharp and Independent

Good analysis isn’t free. And bad analysis? That’s usually paid for.

I want The GPU to stay sharp, independent, and free from corporate fluff. That means digging deeper, asking harder questions, and breaking down the world of GPU compute without a filter.

If you’ve found value in The GPU, consider upgrading to a paid subscription or supporting it below:

☕ Buy Me A Coffee → https://buymeacoffee.com/bbaldieri

It helps keep this newsletter unfiltered and worth reading.

Reply

or to participate.