AI Accelerator Icon

Cloud Server NPU

Artificial Intelligence is moving rapidly, together with chips and processor hardware technology. With an architecture that simulates neural networks, NPU – or Neural Processing Unit – can process high numbers of machine learning tasks saving energy and optimizing costs.
Our Cloud Server NPU are AI Accelerator services that improves your AI inference workloads enabling the highest computing power with the minimum environmental impact.

AI Accelerator CS NPU1

Neural Processing Unit

Tenstorrent Grayskull e75
8 GB RAM
96 Tensix Core
221 TeraFLOPs

Number of cards

1

RAM

64 GB

CPU CORE

16 vCore

Disk

100 GB SSD

Connectivity

10 Gbps

Hourly Cost

0.07
?
This order allows you to have at your disposal a cloud platform for creating and destroying all the NPU VMs you need, even for just one hour

Monthly Fee
(estimated use 24h x 30d)

49.00
AI Accelerator CS NPU2

Neural Processing Unit

Tenstorrent Grayskull e150
8 GB RAM
120 Tensix Core
332 TeraFLOPs

Number of cards

1

RAM

64 GB

CPU CORE

16 vCore

Disk

100 GB SSD

Connectivity

10 Gbps

Hourly Cost

0.10
?
This order allows you to have at your disposal a cloud platform for creating and destroying all the NPU VMs you need, even for just one hour

Monthly Fee
(estimated use 24h x 30d)

70.00

Take advantage of the Grayskull cards high performances that, differently from GPUs, offer easier scalability, ideal for sequential and parallel computation.

Designed for the inference tasks typical of artificial intelligence and machine learning, our NPU AI Accelerator speeds up and simplifies the development of SLM or Small Language Models.

  • What is Cloud Server NPU?

    Cloud Server NPU is an AI Accelerator service. It is a Cloud Server based on Linux Operating System that, integrated with the more standard CPUs, offers NPU technology. NPU stands for “Neural Processing Unit”. They are dedicated processors particularly designed for accelerating neural networks.

  • What is the difference between a GPU and an NPU?

    NPUs are designed for handling diverse and complex tasks related to AI and neural networks, while GPUs can be used for a wider variety of contexts and applications.

    NPUs custom hardware is particularly specialized for processing massive, data-driven calculations (for example, speech recognition, photo and video editing, objects recognition, and so on).

    AI Accelerator, our range of NPU-based cloud servers, guarantees an excellent infrastructure for the full stack development of the so-called Small Language Models, that are AI models having less training data and parameters than LLMs.

  • How to activate AI Accelerator NPU?

    Cloud Server NPU is a range of cloud server instances provided with Tenstorrent grayskull cards and able to satisfy your need of high power for artificial intelligence and neural network tasks. Access Seeweb Cloud Center panel and activate AI Accelerator, choosing the best NPU server for your workload.

    NPU Cloud Servers offer a pay per use billing model (cost/hour) and can be reserved for longer periods at more convenient prices.

  • How much bandwidth and traffic do I have with NPU servers?

    AI Accelerator offers 10 Gbps connectivity and unlimited traffic.

  • How can I access NPU AI Accelerator?

    It is possible to access the NPU AI Accelerator with SSH. The server can be delivered with specific setup when needed and can be integrated with the Seeweb cloud ecosystem.