Tuesday, 4 November, 2025 / Edition 83

Guys. When doing research for the newsletter this week, I came across this “windows button plus H” thing for voice to type. I’m still working through the kinks, but right now it looks like my fingers are going to get out of shape. I’m overly excited to try this on e-mails and IMs, though.

This week, we start a new series covering something the government is very interested in: fusion. This will be at least a two-part series, so let’s dig in!

Sarah Compton

Editor, Enspired

What’s New With NVIDIA?

Tada Images/Shutterstock.com

Recent reports suggest microchip giant NVIDIA has a lot of new projects in the works, and two of them could affect how we work as geos. Here’s the scoop.

Project #1: NVIDIA is building seven new supercomputers for the DOE.

What are they for: The new computers will reportedly be used to help the United States develop and maintain its store of nuclear weapons, as well as research alternative energy sources, such as fusion.

Here comes the competition: Fellow Silicon Valley chip leader Advanced Micro Devices also recently announced a $1 billion partnership with the DOE to construct two other supercomputers to “tackle large scientific problems ranging from nuclear power to cancer treatments to national security,” according to MSN.

Holy microchip: The biggest of these computers will be built in collaboration with Oracle and use 100,000 of NVIDIA’s chips.

Project #2: NVIDIA is collab-ing with several leading communications leaders to build out 6G and improve mobile AI usage by accelerating AI-RAN (radio access network) innovation.

Can you hear me now? In the current setup, AI is fighting for bandwidth against typical cellular service requirements, restricting the AI-based activities we can do/use with our phones, though the Skynet crowd might see that as a good thing 😜.

Why it matters: NVIDIA claims that almost 50 percent of Chat GPT’s 800 million weekly active users access the site via their mobile devices.

Who’s who: There are a lot of players in this partnership so let’s go over them:

• NVIDIA is the ringleader. They are investing $1 billion in Nokia to get the project off the ground.

• Nokia will provide the 6G infrastructure. The company is introducing its Arc Aerial RAN Computer, a 6G-ready telecommunications computing platform.

• T Mobile will attempt to integrate its 6G buildout with Nokia’s AI-RAN tech.

• Dell Technologies is providing its PowerEdge servers that will power this new AI-RAN solution.

Communication service providers will be able to launch AI-native 5G-Advanced as well as 6G networks on NVIDIA platforms.

What is AI-RAN?

• According to a PDF from NVIDIA, AI-RAN enables the deployment of cellular and AI workloads on a shared cloud native infrastructure.

• The underlying infrastructure for AI-RAN is built using a general purpose, accelerated computing platform without any RAN specific hardware, so it can run cellular and AI workloads at the same time and scale as needed.

• The software for AI RAN uses AI native principles to allow containerization and acceleration of AI and RAN workloads to optimize that underlying accelerated computing infrastructure.

Bigger picture: Support for mobile AI use is expanding, and the cellular providers, who are being outdone by Starlink in terms of remote coverage, want to lead the charge.

Goodbye laptops? Maybe your workstation will soon be downsized to just your phone, but don’t worry: I’m sure the six 40” monitors you use for your seismic line aren’t going anywhere.

For more information on the partnership, go here.

Sponsored

DIG Supports Your Geochemistry Needs

• Geochemistry and Isotope Measurement Laboratory
• Exploration, Development, Production and Environmental
• Oil - Gas – Water 
• Serving Global Energy Since 2006

The DoE’s Plans for Fusion

Darkfoxelixir/Shutterstock.com

Sustainable, abundant fuel is the name of the game for the future, and the DOE, led by former Liberty Oil CEO Chris Wright, seems to be hedging its bets on fusion. The agency recently released a Fusion Science and Technology Roadmap that it hopes will accelerate commercial uses of fusion power.

Big promise, low output: Fusion’s main problem is the fact it doesn’t actually work in any meaningful sense of the word (yet) and always seems to be “30 years out.”

The roadmap highlights three common areas—Build, Innovate, Grow—and sets specific goals within near (2–3 years), mid (3–5 years) and long (5–10 years) terms.

Build: The build portion describes how infrastructure for fusion science and technology will be built and calls for a buildout of an AI-Fusion Digital Convergence Platform.

That buildout includes AI for applications from automated controls to digital twins. The controls would keep reactions stable, whereas the digital twins would model various parameters within reactions and reactor design to high grade what is tested with real reactors.

Innovate: This section highlights the pursuit of innovative and transformative research in four key areas:

1. Stellarators (they use magnetic fields to confine plasma)

2. Liquid-metal PFCs, which can advance the cooling issue significantly

3. Alternate fusion concepts

4. Innovations in measurements, since it’s difficult to build tools capable of measuring exactly what’s going on in an environment that mimics the sun

Grow: This section goes over scalability. It covers:

• Expanding public-private partnerships, which can be tricky. There’s a lot of money that needs to come from somewhere, and ownership of the IP is no small thing. This section describes a variety of options spanning various company sizes, funding plans, and ownership splits.

• Securing a supply chain, though “a” supply chain is a misnomer, because several avenues need to be paved. This includes supplies of materials for reactors that can withstand the fusion environment, house facilities to pipe the power that’s generated, the fuel itself, and finally, its disposal.

• Building a talent pipeline. Tradesmen will be needed to build the facilities, and engineers, physicists, and other researchers are needed to suss out the science of it all.

• Leveraging previous fission work. Examples given include advanced manufacturing of high-temperature radiation-tolerant alloys, stress corrosion cracking measurement methodologies and durable, corrosion-resistant molten metal and molten salt system components

• Paths for fusion adoption and commercialization. Once this is all built out and working, people will need to be convinced and supported in their efforts to use it.

More to come: The roadmap goes on to describe some key challenge areas, which we will go over next week. We will also highlight major opportunities for people interested in getting involved.

To jump ahead on your own and read the report, go here.

Sponsored

Geothermal in Action on 13 Nov.

The final session of the AAPG Academy Geothermal Webinar Series brings together leading experts in geothermal development from Halliburton and Mazama to share practical insights and forward-looking strategies that are shaping the future of sustainable energy.

👍 If you enjoyed this edition of Enspired, consider supporting AAPG's brand of newsletters by forwarding to a friend or colleague and signing up for our other newsletters here.

➡️ Was this email forwarded to you? Sign up for Enspired here.

✉️ To get in touch with Sarah or suggest a topic for Enspired, send an email to editorial@aapg.org.