The technology partners helping power the infrastructure, tools, and platforms behind production AI.
NVIDIA's GPUs and software stack power the vast majority of AI training and inference happening today, from foundation models to the infrastructure running production workloads at scale.
Pulumi is the open-source infrastructure as code platform for building and managing cloud infrastructure in real programming languages — TypeScript, Python, Go, C#, Java, and YAML — with the type checking, testing, and reusable components engineering teams already rely on.More than 4,000 companies run on Pulumi, including BMW, Snowflake, NVIDIA, Mercedes-Benz R&D, Atlassian, and Supabase.The platform spans infrastructure as code with 170+ providers across AWS, Azure, Google Cloud, Kubernetes, and AI/ML workloads; Pulumi Neo, an AI agent purpose-built for infrastructure engineering; centralized secrets and configuration through Pulumi ESC; and cloud-wide insights, compliance, and policy enforcement. SOC 2 Type II certified.As AI agents write more infrastructure code, agents and engineers share the same languages, tests, and guardrails.Open source, free for individuals, 25,000+ GitHub stars.
FriendliAI is the Frontier Inference Cloud for Agents, delivering high throughput, low latency, and reliability at scale for agentic workloads. Through vertically optimized inference infrastructure, it achieves up to 7× faster output token speed and 50–90% lower inference cost, backed by a 99.99% uptime SLA built for high-volume agent traffic. The platform powers low-latency streaming for real-time agents, reliable long-context inference, and robust tool calling.
Clockwork.io pioneered Software-Driven AI Fabrics™ — a programmable layer between hardware and workloads that delivers nanosecond-accurate telemetry, AI fault tolerance, and performance optimization across accelerators, networks, and deployment models. AI workloads require clusters to operate as one machine, yet failures and infrastructure bottlenecks strand capacity. Clockwork’s FleetIQ platform recovers that capacity, helping enterprises train, deploy, and serve demanding AI workloads faster, more reliably, and at lower cost across Ethernet, RoCE, and InfiniBand — without hardware lock-in. In independent testing, SemiAnalysis found TorchPass, Clockwork’s fault-tolerance product, to be the only solution that maintained full training throughput during failures, outperforming checkpoint-restart and leading open-source frameworks. Uber, Wells Fargo, DCAI, Nebius, NScale, and White Fiber trust Clockwork.io to power their AI infrastructure. Learn more at www.clockwork.io.
Liquid AI is a foundation model company building efficient, general-purpose AI at every scale. Founded out of MIT's Computer Science and Artificial Intelligence Lab, Liquid started from the observation that 99% of the world's processors live outside data centers, and they come with hard limits on latency, memory, power, and connectivity.
To address this, Liquid designs models around the hardware constraint first, rather than building a model and compressing it to fit afterward. The result: Liquid Foundation Models (LFMs) run on a hybrid architecture that moves past the limits of transformers, using a hardware-in-the-loop approach with our proprietary Automated Foundation Model Design (AFMD) to evolve architectures optimized for the exact silicon they'll run on.
That approach allows Liquid to deliver models that are private, fast, and reliable across both online and offline conditions. Our LFMs already power real deployments across automotive, consumer devices, drug discovery, and enterprise workflows, with partners including Mercedes-Benz, Shopify, Insilico Medicine, and more.
The VAST AI Operating System powers the world's most demanding artificial intelligence. It unifies infrastructure to deliver scalable performance, radical simplicity, and accelerated discovery, from model training to inference.
Core Scientific designs, builds, and operates large-scale, high-density data centers purpose-built for artificial intelligence and next-generation compute. The company delivers a fully integrated platform supporting every phase of the data center lifecycle.
Union.ai is the AI runtime for teams building production AI. Built on Flyte, the open-source orchestrator trusted by 4,000+ organizations and frontier AI labs, Union.ai unifies agent-native orchestration, durable self-healing workflows, and real-time and batch inference in one platform that runs on your own infrastructure. Union.ai helps ML, data, and platform teams move from experimentation to reliable, scalable AI—so scientists and engineers focus on the work, not the infrastructure.
NVIDIA pioneered accelerated computing to help solve the most challenging computational problems. NVIDIA is now a data-center-scale AI infrastructure company reshaping all industries.
