Resources

Discover insights on the GPU cloud market, hear from real Shadeform users, and explore tutorials on a variety of use cases.

Serving GLM-5.3-Flash on Eight L40s at 256k Context

An inference engineering report on running a 320B MoE model built for Hopper on an 8× L40 node with no GPU peer-to-peer: what changed, what was measured, and what it traded away.

Article|October 5, 2026

How Mosies keeps training pipelines running during a global GPU shortage with Shadeform

Moises is the creative suite for musicians trusted by Grammy-winning artists, prestigious institutions, and more than 73 million artists and producers across the world. Moises runs on several proprietary models that the company trains in-house using licensed materials. By working with Shadeform, Moises is able to consistently secure compute for training with unified access to 30+ providers, reduce their infrastructure complexity with a single consolidated platform, and accelerate training timelines with access to better hardware at lower prices.

Customer Story|July 2, 2026

How Parasail turns around any customer GPU requirement in under an hour using Shadeform

Parasail is an AI inference platform that orchestrates workloads across a global supply of GPUs to hit a target combination of price, performance, reliability and scale for their customers. By working with Shadeform, Parasail can seamlessly deploy the right hardware for their customer needs across 30+ vetted providers, both instantly for burst workloads and within an hour for complex cluster requirements.

Customer Story|April 9, 2026

How Tinfoil sources confidential computing infrastructure 15x faster with Shadeform

Tinfoil is a private AI platform that runs inference, fine-tuning, and AI agent workloads inside secure hardware enclaves so no one, not even Tinfoil, can access sensitive user data. To enable this, Tinfoil requires secure hardware features only available in select modern GPUs and CPUs, as well as specific firmwares and BIOS configurations. By working with Shadeform, Tinfoil can easily access machines with these uncommon specs at scale through a single platform, rather than hunting down small allocations across individual providers.

Customer Story|March 26, 2026

Blackwell Ultra B300s, Now Available on Shadeform

NVIDIA’s latest accelerator, the Blackwell Ultra B300, is now available to deploy on-demand on Shadeform, with reserved clusters for both HGX B300 and GB300 available via inquiry.

Article|February 12, 2026

Model Guide: Kimi K2.5

Explore the latest flagship model release from Moonshot AI, how it compare to similar models, recommended hardware, and how to deploy it yourself.

Tutorial|January 29, 2026

Model Guide: Deepseek V3.2

Explore two new model releases from DeepSeek, how they compare to similar models, which GPUs make the most sense to run them on, and how to deploy them yourself.

Tutorial|December 5, 2025

Evaluate Clouds with Shadeform + Langfuse

Learn how to evaluate self hosted model performance across multiple cloud environments using Shadeform and Langfuse.

Tutorial|August 13, 2025

NVIDIA B200 GPU Price Guide

Learn about the NVIDIA B200 GPU, its features, architecture, AI performance, and pricing across the market.

Article|August 11, 2025

Benchmarking GPUs for MD Simulations: Speed and Cost Insights from SimAtomic & Shadeform

Simatomic partnered with Shadeform to provide extensive insights into the speed and cost of running MD simulations on a variety of GPUs from different cloud providers.

Article|August 8, 2025

NVIDIA H200 GPU Price Guide

Learn about the NVIDIA H200 GPU, its features, architecture, AI performance, and pricing across the market.

Article|April 28, 2025

NVIDIA H100 GPU Price Guide

Learn about the NVIDIA H100 GPU, its features, architecture, AI performance, and pricing across the market.

Article|April 25, 2025