Tailored Acceleration for Your Workload

Customized DPUs

From AI and networking to data processing and edge analytics, each Kalray DPU is co-designed and optimized for your specific workload, transforming your requirements into a validated, domain-specific accelerator that delivers efficiency, scalability, and control.

Kalray designs custom DPUs using its MPPA® architecture to meet your exact compute requirements.

1. Define The Acceleration Needs Of Your Use Case

Outcome: Clear workload definition and performance targets.

2. Translate insights into a modular architecture plan

Outcome: Validated architecture blueprint optimized for your workload.

3. Create And Validate Your Customized DPU

Outcome: Functional prototype ready for integration.

4. Deploy You DPU At Speed And Scale

Outcome: Full deployment with lifecycle support and optimization.

Benefits of building DPUs with MPPA

A Ready-Made Architecture For Building DPU Accelerators

No need to start from scratch

A shorter path to deployment

Reusable IP, pre-integrated subsystems, and standard toolchains

A flexible foundation that adapts to rapidly evolving workloads

Without sacrificing efficiency or control.

Kalray brings unmatched expertise in designing and delivering specialized DPU accelerators

Whether you need a fully customized DPU, a semi-custom acceleration card, or chiplets to integrate into your own SoC, Kalray’s co-design teams are ready to collaborate. Our proven track record includes successful deployments across AI, networking, storage, and secure computing—helping customers unlock domain-specific performance without the delays or costs of a full ASIC program.
  • 500+ engineering-years invested in MPPA development, covering silicon, boards, and software
  • Modular design approach, enabling rapid adaptation to new workloads and form factors
  • Shorter time-to-market than FPGAs or ASICs, with lower development risk
  • Flexible delivery options: standalone chips, chiplets, or fully integrated PCIe acceleration cards
  • Proven architecture deployed in AI, networking, security, and storage environments
  • End-to-end co-design support, from architectural definition to hardware, firmware, and SDKs
  • Developer-friendly programmability, with no RTL expertise required
  • Trusted by leading innovators across cloud, telecom, defense, and embedded systems

From Architecture to Real-World Impact

Discover how Kalray’s MPPA®-based DPUs power domain-specific acceleration for today’s most demanding workloads

AI & Compute

Accelerate AI inference, training, and compute-heavy workloads with tailored, energy-efficient DPUs.

Storage Acceleration

Optimize data throughput and reduce latency for next-generation storage systems.

5G Radio Access Networks (RAN)

Enable low-latency, high-bandwidth processing for software-defined 5G infrastructure.

Security Acceleration

Deliver scalable, high-performance encryption and secure data processing at the edge.

Kalray is recognized by industry analysts and certified by leading quality and interoperability organizations.

Analyst Recognition
GigaOm recognized Kalray as a Leader and Fast Mover in the 2023 Data Processing Units (DPU) Sonar Report — highlighting our innovative architecture and strong performance across AI and data workloads.
Quality Certifications
Kalray’s ISO 9001 certification and interoperability validations demonstrate our commitment to high-quality standards and sustainable performance in every solution we build.