Chris's AI supply chain startups

High-throughput chips for training, reinforcement learning, inference, and long-context workloads.

A national HVDC grid delivering power directly to AI and industry.

Rethinking the limits of AI, Lightmatter merges photonics and computing to build a future where speed, efficiency, and intelligence converge

AI infrastructure for training, inference, and agents, scaling to thousands of GPUs.

Train, evaluate, and deploy AI models you own.

Building factory-made nuclear reactors to supply power, hydrogen, and synthetic fuels

Inference that keeps getting better.

Optical connections that let AI clusters scale beyond the rack.

Memory-centric compute for AI inference.

A multi-silicon cloud for faster, more power-efficient AI inference.

Solid-state transformers for the AI and clean-energy grid.

Frontier AI inference systems built from specialized chips and optical interconnects.

Go where the energy is

Positron makes purpose-built hardware to accelerate generative AI.

Substrate is building America

We are creating a platform that makes it easy to turn any deep learning model into custom silicon quickly.

Programmable AI computing built around open software and RISC-V.

xLight builds powerful free-electron lasers for semiconductor manufacturing.

Factory-built modular nuclear power plants designed for AI data centers at scale.

fab2 builds the hardware and software needed to make chips, fabs, tools, and components.

Domestic uranium enrichment for America's conventional and advanced nuclear reactors.

GPU compute capacity, market data and financial infrastructure.

The San Francisco Compute Company. Buy and resell large-scale, vetted GPU clusters from the terminal.

Data Centers in Space

America’s Turbine Company. Rebuilding the capacity to manufacture turbines at scale, here in America.

AMD-powered cloud infrastructure for large AI models.

Modular turbine clusters for on-site generation

Redefining the future of energy, one atom at a time.

The first U.S. derivatives exchange purpose-built for the artificial intelligence economy.

Armada’s platform combines connectivity, compute, and real-world AI to solve your toughest challenges right where your data is generated.

Transform ships into autonomous combat data centers.

Building Overture, a faster and more efficient supersonic airliner.

The AI-native operating partner for the built world.

Compact fusion plants built around powerful magnets, working toward abundant energy.

Self-healing infrastructure for AI compute.

Power for AI. Flexibility for the Grid.

Adaptive Phase Cooling for data centers.

On-premises AI infrastructure for regulated industries: private, secure, and examiner-ready. Deploy in under 24 hours with zero data egress.

GridCARE is the essential one-stop power solution for AI data centers, delivering energy security and resilience at unprecedented speed.

Lamb Labs is building MPUs (Model Processing Units) that hardcode entire AI models into silicon.

AI supercomputers for training and inference, from on-demand GPUs to dedicated clusters.

Connectivity Platform for AI

Optical systems built with in-house factories and software tuned to the hardware.

Community-funded research and image and video tools that expand human imagination.

A new paradigm of AI hardware to power human progress.

Research at Oak Lab

Discover PRISM, the full photonic network from Oriole Networks. Improve AI performance, cut energy use and simplify data centre architecture

The cloud you own. Hardware, with the software baked in, for running infrastructure at scale.

AI agents for AI factories

Portable one-megawatt nuclear generators designed to replace off-grid diesel power.

Frontier AI methods for chip design and self-improving systems.

Frontier-scale open models in hosted infrastructure, with RunAnywhere accelerators and SDKs for local inference.

One GPU cloud from model experimentation to production inference.

The execution layer where AI agents get faster and safer.

Efficient inference and persistent computers for long-running AI agents.

The Forefront of Semiconductor Integration

Building the new default chip for computation.

Leading innovations in nuclear applications for reactor and storage technology, while advancing methods to transform the fight against the1?

The Biological Computing Co. (TBC) turns discoveries from real neurons into software that makes AI faster and cheaper, without retraining.

A simple and powerful neural network framework

Real-time GPU infrastructure with API-native provisioning and per-second billing.

The intelligence layer between your assets and the grid, built for the AI era.

The Utility of Compute.

Manufactures rare earth magnets for electric vehicles, defense, and other advanced technologies.

Expand ambition in the physical world.

AI silicon for recursive self-improvement and superintelligence.