echoloc
Get started
echoloc/ Companies/ Rebellions

Rebellions Tech Stack

AI inference chip and full-stack software platform for on-premises deployment

Semiconductor Manufacturing Seoul 201–500 employees Founded 2020 Privately Held
rebellions.ai LinkedIn

Rebellions designs purpose-built inference accelerators (Rebel 100) paired with full-stack software (PyTorch and vLLM native support) targeting enterprises and governments seeking sovereign AI capability. The tech stack reveals deep hardware specialization—Verilog, SystemVerilog, UVM, PCIe Gen 6, RDMA, InfiniBand, TPU/GPU integration—coupled with infrastructure software (Kubernetes, RTOS, SR-IOV, IOMMU). Active projects center on NPU communication stacks, packaging solutions, and automated design flows, while pain points cluster around power delivery, crosstalk mitigation, and communication bottlenecks—typical of early-stage chip companies scaling from prototype to production.

Tech Stack 57 technologies

Core StackPython PyTorch Kubernetes C++ TensorFlow HubSpot Verilog SystemVerilog UVM TCL ANSYS HFSS SIwave Keysight ADS PCIe Gen 6 C/C++ RDMA NCCL MPI NVLink GPU TPU vLLM TensorRT-LLM IOMMU RTOS KVM SR-IOV RISC-V JTAG InfiniBand+27 more

What Rebellions Is Building

◆Challenges

  • Stable power delivery under high-frequency loads
  • Mitigating crosstalk and reflection
  • Implementing next-generation packaging solutions
  • Performance bottlenecks in npu communication
  • System-level bottlenecks
  • High availability failover
  • Ai inference performance
  • Streamlining partner operations processes
  • Accelerating deal speed
  • Driving npu infrastructure growth

▲Active Projects

  • Next-generation packaging solutions implementation
  • Automated design flows development
  • Collective communication library for npu
  • Ai inference infrastructure design
  • Ai solution adoption
  • Performance benchmarking
  • Communication software stack development for npu
  • Performance optimization of communication stacks
  • D2d subsystem development
  • Ucie ip firmware

Hiring Activity

Accelerating10 roles · 9 in 30d

Department

Engineering
7
Finance
1
Marketing
1
Ops
1

Seniority

Senior
7
Manager
2
Principal
1
Company intelligence

Find more companies like Rebellions by tech stack, pain points and active projects

Get started free

About Rebellions

Rebellions builds AI inference hardware and software for on-premises deployment, founded in 2020 and based in Seoul with operations across South Korea and the United States. The company's core product is the Rebel 100, a chiplet-based accelerator optimized for inference workloads, paired with a software stack supporting PyTorch and vLLM out of the box. They position the system around three competitive angles: performance-per-dollar, power efficiency, and the ability for enterprises and governments to retain control over AI infrastructure without cloud dependency. Active deployments span commercial and government sectors.

HeadquartersSeoul
Company Size201–500 employees
Founded2020
Hiring MarketsSouth Korea, United States

Frequently Asked Questions

What tech stack does Rebellions use?

Hardware design: Verilog, SystemVerilog, UVM, ANSYS HFSS, Keysight ADS. Infrastructure: Kubernetes, KVM, RTOS, IOMMU, SR-IOV, PCIe Gen 6, RDMA, InfiniBand, NVLink. Software frameworks: PyTorch, TensorFlow, vLLM, TensorRT-LLM.

What is Rebellions working on?

Next-generation chiplet packaging, automated design flows, NPU communication software stacks, performance benchmarking, and D2D subsystem development. Internal challenges include power delivery stability under high-frequency loads and NPU communication bottlenecks.

Similar Companies in Semiconductor Manufacturing

Other companies in the same industry, closest in size

V
Vishay Siliconix Itzehoe GmbH
Semiconductor Manufacturing

How this profile is built

Rebellions's technology stack, projects, and hiring signals are inferred from public hiring and company data — career pages, public listings, and company web presence — then clustered and de-duplicated. Figures are estimates that refresh over time. Read our full methodology →

This is not an official vendor or customer list. It is a technology-adoption signal inferred from public data, intended for B2B research.