Issue #3: xAI’s $6B Raise, AWS’s $10B Bet & Core42's New Partnership

I’m Ben Baldieri, and welcome to The GPU.

Every week, I deliver sharp, no-fluff insights on the moves shaping GPU compute, AI infrastructure, and the data centres powering it all.

From Musk’s $6B Christmas raise to AWS’s next billion-dollar expansion and Intel teasing a potential rival to Nvidia, this is where you’ll find the updates that matter.

Here’s what’s in this issue:

• Elon Musk’s xAI raises another $6B to grow its AI supercomputer.

• Nvidia’s acquisition of Run:ai gets the EU’s green light.

• Core42 partners with TeraWulf for a GPU-powered NeoCloud push into the US.

• Genesis promises generative AI worlds powered by real physics.

• Intel teases “Jaguar Shores” in its fight to catch up to Nvidia.

• Digital Realty’s $3B equity plan to expand its data centre empire.

• AWS announces a $10B investment in Ohio to meet AI demand.

Let’s dive in.

The GPU Audio Companion Issue #3

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

xAI Secures $6B to Power 100,000 More GPUs

Elon Musk’s xAI just raised $6 billion in a Series C round.

This brings its total funding to $12B in less than a year. Backers include A16Z, Nvidia, AMD, Blackrock, and more. Musk plans to double the size of xAI’s Memphis supercomputer, which already hosts up to 100,000 Nvidia H100 GPUs, with an eye on scaling to 1 million GPUs.

Why this matters:

• Musk’s supercomputer is shaping up to be one of the largest AI training infrastructures in the world.

• Nvidia and AMD investing in xAI signals confidence in the long-term GPU demand Musk is driving.

• xAI’s $50B valuation cements it as a direct challenger to OpenAI, with a focus on Musk’s vision of “truth-seeking” AI models.

Read the full story on DCD here.

Nvidia Wins EU Approval for Run:ai Deal

The green company receives the green light.

Nvidia’s $700M acquisition of Israel-based AI resource manager Run:ai has cleared EU regulatory hurdles. Run:ai helps optimise GPU workloads, making it a critical addition of Nvidia’s burgeoning AI software ecosystem.

Why this matters:

• This deal strengthens Nvidia’s dominance in AI infrastructure, integrating GPU hardware with software that boosts efficiency.

• Run:ai’s longstanding collaboration with Nvidia ensures seamless integration into its platform.

• Despite Nvidia’s market power, the EU’s approval highlights its faith in competition remaining intact.

Read the full story on Tech Crunch here.

Core42 Partners With TeraWulf

Core42 has signed long-term data centre lease agreements with TeraWulf.

This unlocks 70MW of turn-key data centre infrastructure for the G42 company at TeraWulf’s Lake Mariner facility in Upstate New York. The deal will support Core42’s GPU clusters, including Dell IR5000 racks with liquid-cooled PowerEdge XE9680L GPU servers.

The infrastructure will come online in phases, starting in Q1 2025, with full operation expected by Q3 2025. The agreement also includes options to expand Core42’s hosting capacity by an additional 135MW, equivalent to 108MW of critical IT load.

Why this matters:

• TeraWulf could be positioning itself as a key player in AI infrastructure hosting by tailoring its facilities to GPU-intensive deployments.

• Core42’s use of advanced Dell liquid-cooled servers reflects growing demand for sustainable, high-performance computing.

• With scalability built into the deal, potential expansion could solidify Core42’s role in the evolving AI compute ecosystem, especially in the US.

Read the full story in the G42 Newsroom here.

Genesis: Generative AI Meets Physics

GenAI is colliding with real physics in a new physics engine.

Genesis is an open-source generative physics engine. It can simulate 4D dynamical worlds for robotics, physical AI, and more. The engine is 10–80x faster than existing platforms like Isaac Gym, with speeds ~430,000x real-time. This means it can train robotics policies in just 26 seconds on an RTX 4090.

Wild.

Why this matters:

• Genesis bridges generative AI and physics, opening new doors for robotics and virtual worlds.

• Its open-source approach could democratise access to cutting-edge simulation tools.

• Faster training speeds slash costs, making advanced robotics and physical AI more accessible.

Read about Genesis on the project webpage here.

Intel Teases Jaguar Shores

Intel has hinted at a new AI chip, “Jaguar Shores,” as a successor to its upcoming Falcon Shores GPU.

Jaguar Shores aims to challenge Nvidia and AMD, signaling Intel’s intent to re-enter the high-performance AI race after an unending series of setbacks. Potentially challenging with a headless organisation. And with shareholders demanding the recently fired CEO return 3 years of salary. But at least they’re not out of the race.

…yet?

Why this matters:

• Intel has struggled to compete in AI, but Jaguar Shores shows it’s still betting on GPUs to fight back.

• With AMD and Nvidia dominating, Intel will need Jaguar Shores to deliver serious performance gains to close the gap.

• Intel’s future may hinge on leveraging its upcoming 18A manufacturing node to gain an edge over rivals.

Read the full story on Tech Radar here.

Digital Realty Launches $3B Equity Plan

Digital Realty has unveiled a $3B at-the-market (ATM) equity offering to fund its next phase of growth.

With 312 data centres globally, the company plans to use proceeds to expand its portfolio, repay debt, and fund new developments. Given that rack densities and power and cooling requirements are only moving in one direction, it’s not hard to imagine where and how the funds raised from this offering will be allocated.

Why this matters:

• Digital Realty’s 83.9% occupancy rate shows continued strong demand for co-location and hyperscale data centres.

• This equity plan gives it the financial flexibility to capitalise on AI-driven growth.

• As one of the largest data centre REITs, Digital Realty’s moves often set trends for the broader sector.

Read the full story on The Tech Capital here.

AWS Bets $10B on Ohio for AI

AWS announced plans to invest $10 billion into expanding its data centres in Ohio by 2030.

The investment will fund up to eight new sites. This means hundreds of direct AWS jobs and thousands more across the supply chain. Fayette County alone will see $5B go toward cutting-edge data centre builds.

Proper money.

Why this matters:

• AWS’s investment highlights the growing demand for AI infrastructure as workloads explode.

• This expansion solidifies Ohio’s position as a key hub for AI development in the US.

• By integrating renewable energy projects, AWS is aiming to balance sustainability with AI’s massive power needs.

Read the full story on AWS here.

The Rundown

Not even a Christmas slowdown this week.

xAI’s $6B raise underscores the race to dominate AI infrastructure. Nvidia’s Run:ai acquisition enriches the green ecosystem. TeraWulf’s Core42 deal adds another piece to the NeoCloud puzzle, and Genesis is redefining generative physics for robotics, enriching the software layer. Intel’s Jaguar Shores hints that rumours of its death may yet be greatly exaggerated. Digital Realty’s $3B equity plan positions it for long-term growth, and AWS’s $10B Ohio expansion highlights how hyperscalers are not playing around.

Never a dull moment!

Where do you think the AI infrastructure race goes next? Reply to this email and let me know!

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.