An alternative to Modal.
Your Python. European GPUs.

01 / French expertise, European infrastructure

Put a French cloud partner behind your workload.

Blade is Shadow’s AI infrastructure brand. Shadow is a French company registered in Paris, identified in our service terms. Blade’s own inference GPUs are operated by Shadow in Europe. That combines a French cloud partner with a concrete European hosting option for Blade-hosted models.

The team brings ten years of cloud experience to the service. Running GPU workloads at scale means managing capacity, availability and responsiveness as everyday engineering work. You can read more about the cloud experience behind Blade.

Years of cloud experience

10

Instances per year

16M

GPUs in the fleet

12,000+

These figures describe Shadow’s cloud operations, the experience behind Blade.

For AI Fabric, discuss your required execution location, runtime and data paths during onboarding. This brings the infrastructure discussion into the design of your application: what will run, what capacity it needs and which processing requirements apply.

Modal also supports regional controls. Its data residency documentation distinguishes container location, request routing and storage; several logs and storage products are documented as US-based. Blade should likewise be assessed service by service. Our hosting and data handling guide describes the scope currently documented for inference and optional tools.

02 / Code and compute together

Start from your Python and size the compute around it.

AI Fabric is designed for your own models and Python workloads: serving an endpoint, processing a batch or running the work behind an application. The team helps you confirm the runtime, GPU memory and deployment options for a representative workload.

That is useful when you want to discuss the fit between your code and the compute before committing production traffic. Bring your dependencies, sample inputs, expected concurrency and target response time. Use those to agree what to test and how to judge the result.

Modal’s platform documentation describes Python-defined environments, managed scaling, GPU execution and per-second usage billing. It already provides a self-service path for those workloads. AI Fabric is opening gradually; it is an option to evaluate with the Blade team, not a claim of immediate feature parity.

03 / Capacity that follows the work

Make bursty workloads part of the design.

Document processing, evaluations and scheduled jobs do not always need a permanently running GPU. AI Fabric is designed around autoscaling and scale to zero. Confirm the available settings, startup behaviour and commercial terms with your access.

The right test includes an idle period, a burst of requests and the time needed to load your model. Measure both the cost and the response time over that whole cycle. This makes the discussion about the service your application receives, rather than a headline rate for a GPU.

Modal also provides managed serverless execution. Compare the resource allocation, scaling configuration and billing scope of the actual workloads you plan to run; this article does not establish a price or cold-start advantage for Blade.

04 / Two ways to build with Blade

Use a hosted API where your own deployment adds no value.

Some parts of an application need custom Python; others simply need a model response. Models is available today for the second case. It gives you a hosted catalogue through an OpenAI-compatible API without deploying those models yourself.

For an assistant that needs current sources, supported models can use Flash by Staan web search on enabled accounts. Blade handles the built-in search loop; executed searches are billed separately. You can review the search integration while preparing the custom parts of your workload for AI Fabric.

This lets you separate the work you need to own from the model serving you can hand over. Hosted inference uses the published model rates; AI Fabric’s workload support and terms are confirmed during onboarding.

Blade and Modal at a glance.

Scroll sideways to compare both providers →