John J. Villa

General Compute is building the worlds fastest AI cloud. AI agents, coding tools, and voice assistants all share one constraint: token generation speed. The infrastructure serving them today was built for training, not inference. Were fixing that. General Compute is the first ASIC-first AI neocloud, purpose-built for high-speed inference. Our platform delivers 5-7x faster token generation than GPU-based clouds, serving the latency-sensitive workloads that are growing fastest: coding agents, voice AI, and autonomous systems.

We teach robots at scale

At Proximal, we build training environments that realistically reflect the work of software engineers. We work with frontier labs and enterprises to build simulations of the most complex engineering tasks to help them train and adapt the next generation of coding agents.

Compute is a commodity. We think people should buy it like one. Startups shouldn’t be forced to buy a year’s worth of compute time in order to get market rate and compute providers shouldn’t go bankrupt because they can’t fully book their clusters. At SF Compute , our goal is to solve this issue the same way this was solved for every other commodity — by building a venue where compute contracts are traded in real-time and by bringing a new kind of participant into the supply chain, traders. Buyers will be able to get a good price for any order, whether it’s 32 H100s for a month or 8,000 H100s for an hour, and sellers will instantly book out their clusters because traders will speculatively buy them, for a spread. Every FLOP will flow through us somewhere in the supply chain. What Brent is for oil, we will be for compute.

The Biological Computing Co. (TBC) is harnessing the brains intelligence to redefine computing. We use real neurons to make frontier AI models more stable, scalable and efficient as well as help shape what comes after language and transformers.

For decades companies have relied on archaic tools to inform decisions and make bets on the future. Until now. Fundamental empowers businesses to turn gambles into guarantees and determine their future with far greater accuracy than ever before. Built by DeepMind alumni and trusted by Fortune 100 enterprises, NEXUS is our most powerful Large Tabular Model (LTM). By revealing the hidden language of tables, NEXUS unlocks trillions of dollars of value by giving businesses the Power to Predict™.

Odyssey is an AI lab pioneering general world models: causal, multimodal systems that learn to understand and simulate the world. This foundational technology promises to revolutionize robotics, science, healthcare, education, gaming, and beyond.

Foundational AI for decoding the brain

Roboflow creates software-as-a-service products to make building with computer vision easy. Over 1,000,000 developers use Roboflow to manage image data, annotate and label datasets, apply preprocessing and augmentations, convert annotation file formats, train a computer vision model in one-click, and deploy models via API or to the edge. https://roboflow.com

World-class performance per parameter language models, from IoT to the data center.

We are revolutionizing the future of TV with the worlds first dual screen television offered for free! Visit www.freetelly.com for the ultimate free television upgrade for the living room.

Use Vellum to build and ship reliable AI solutions. Define, evaluate and monitor AI solutions through test-driven development for AI.

The Missing Google Colab Pro Tier. Fine-tune, train, or deploy. Use your own notebook, or one of ours. SSH too. CUDA, Python, Jupyter Lab, all set up.

CP Overture is an early-stage venture capital firm focused on software investments.

General Compute is building the worlds fastest AI cloud. AI agents, coding tools, and voice assistants all share one constraint: token generation speed. The infrastructure serving them today was built for training, not inference. Were fixing that. General Compute is the first ASIC-first AI neocloud, purpose-built for high-speed inference. Our platform delivers 5-7x faster token generation than GPU-based clouds, serving the latency-sensitive workloads that are growing fastest: coding agents, voice AI, and autonomous systems.

We teach robots at scale

At Proximal, we build training environments that realistically reflect the work of software engineers. We work with frontier labs and enterprises to build simulations of the most complex engineering tasks to help them train and adapt the next generation of coding agents.

Compute is a commodity. We think people should buy it like one. Startups shouldn’t be forced to buy a year’s worth of compute time in order to get market rate and compute providers shouldn’t go bankrupt because they can’t fully book their clusters. At SF Compute , our goal is to solve this issue the same way this was solved for every other commodity — by building a venue where compute contracts are traded in real-time and by bringing a new kind of participant into the supply chain, traders. Buyers will be able to get a good price for any order, whether it’s 32 H100s for a month or 8,000 H100s for an hour, and sellers will instantly book out their clusters because traders will speculatively buy them, for a spread. Every FLOP will flow through us somewhere in the supply chain. What Brent is for oil, we will be for compute.