News
DealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory’s NVIDIA & AMD Enterprise ExpansionDealsCorvex agrees USD 103 million take-or-pay deal to perform AI inference with Mid-Atlantic GPUsSovereignNeysa scales AI infrastructure in India with AMD Instinct MI350X GPUs on TytanInferenceMicrosoft Foundry adds Baseten, DeepSeek AI inference (MSFT:NASDAQ)ClustersDataOne and Rowan University sign engineering co-op program agreement for DataOne's Vineland AI factoryClustersLook inside Cal Poly's $3 million AI Factory (Video)DealsOpenAI Taps AMD Chips for 457 MW of AI Capacity in Six-Year TensorWave DealClustersCal Poly launches an AI data center on campus. Take a look insideInferenceArista takes aim at the next AI network battleground: distributed inferenceDealsOpenAI Signs 457 MW AI Data Center Deal as AMD Expands Its Challenge to NvidiaClustersBehind Dell AI Factory’s NVIDIA & AMD Enterprise Expansion
EveryNeocloud
Beam

Beam

Developer cloud for AI agents and GPU workloads

Beam is a developer-focused cloud platform that provides on-demand compute for AI agents and GPU tasks. It sources bare-metal servers globally to offer virtual machines, sandboxes, and serverless functions with fast startup times.

Share on WhatsApp

  • H100 GPUs available from $1.83 per hour
  • Sandboxes start in about half a second
  • No egress fees for data transfer
  • Billing stops when VMs are paused
  • Supports persistent VMs and serverless functions

What it is

Beam is a cloud infrastructure provider designed for developers building AI products. It abstracts away server management by offering virtual machines, isolated sandboxes, and serverless endpoints. The platform focuses on reducing wait times for compute resources, allowing code to run on demand while billing only for active usage. It supports both CPU and GPU workloads, with options for persistent virtual machines and ephemeral sandboxes that start in under a second.

The compute

The platform virtualises CPUs and GPUs directly on bare-metal hardware, using local NVMe storage for speed. It offers Nvidia H100 GPUs at rates starting from $1.83 per hour and RTX 4090s for lighter tasks. Sandboxes provide isolated environments for untrusted code, such as agent tool calls, with fast snapshot and restore capabilities. Clusters with InfiniBand networking are available by quote for larger training jobs.

Where the GPUs are

Beam sources bare-metal servers from providers around the world rather than owning its own data centres. This global distribution allows it to place compute close to users and reduce latency. The company states it buys hardware in bulk to pass savings to customers, though specific data centre locations are not listed in the provided material.

Customers and deals

The platform is used by thousands of developers to build applications without managing underlying infrastructure. Notable users include Jamie, whose co-founder Louis Morgner praised the developer experience. The service targets teams running coding agents, background jobs, and continuous integration builds who need scalable GPU access.

Money and ownership

Pricing is usage-based, billed by the second or millisecond depending on the resource type. There are no egress fees for data transfer. Storage pricing is planned at $0.0001 per GB-hour but is not currently billed. Volume pricing and bring-your-own-cloud options are available for high-volume users. Specific funding details or valuation figures are not provided.

Questions people ask about Beam

What types of compute does Beam offer?

Beam provides virtual machines, isolated sandboxes, task queues, serverless functions, and GPU instances. It supports both CPU and GPU workloads, including Nvidia H100s and RTX 4090s, with options for persistent state and fast scaling.

How does Beam handle billing?

Compute is billed by the second or millisecond while code is running. Costs stop when a virtual machine is paused. There are no egress fees, and storage billing is planned but not yet active. Users pay only for active resource usage.

Is Beam suitable for running untrusted code?

Yes, Beam offers sandboxes specifically for untrusted code like LLM outputs. These isolated environments start quickly, have their own filesystems, and can be paused or snapshotted. They provide a secure way to execute agent tool calls safely.

Good reading and sources

    Where this page comes from

    There is no English Wikipedia article on Beam yet, so the facts on this page come from the sources listed above: beam.cloud/, beam.cloud/about. Some are the subject's own pages, so its claims about itself are reported as its claims. EveryNeocloud wrote the text in its own words, with AI assistance, and compares it with the sources to make sure no passage is copied.

    It may contain mistakes, and figures change over time. Nothing here is investment or buying advice. If something is wrong, tell us. Last refreshed 2026-11-14.

    The weekly digest

    Every Monday: the provider of the day, new pages on EveryNeocloud, and the conferences and anniversaries of the week ahead.