48+ AI Hardware & Silicon

Secure the silicon behind every AI workload.

From GPUs and custom accelerators to HBM memory and foundries — Safeguard maps the hardware and silicon supply chain that powers modern AI, so you know exactly what your models run on.

GPUs & AI accelerators

13

NVIDIA

United States

GPUs and AI accelerators (H100, Blackwell)

AMD

United States

Instinct GPUs and EPYC CPUs

Intel

United States

Xeon CPUs and Gaudi AI accelerators

Google TPU

United States

Tensor Processing Units for training and inference

AWS Trainium / Inferentia

United States

Custom training and inference silicon

Microsoft Maia

United States

Azure custom AI accelerator

Meta MTIA

United States

Custom training and inference accelerator

IBM Telum / AIU

United States

On-chip AI inference acceleration

Cerebras

United States

Wafer-scale AI engine

Groq

United States

LPU inference accelerators

SambaNova

United States

Reconfigurable dataflow units (RDU)

Graphcore

United Kingdom

Intelligence Processing Units (IPU)

Qualcomm

United States

Cloud AI 100 and on-device NPUs

Custom silicon, IP & EDA

7

Broadcom

United States

Custom AI ASICs and networking

Marvell

United States

Custom AI silicon and interconnect

Arm

United Kingdom

CPU architecture and IP licensing

Apple Silicon

United States

M-series chips with Neural Engine

Imagination

United Kingdom

GPU and neural network accelerator IP

Synopsys

United States

EDA tools and semiconductor IP

Cadence

United States

EDA tools and design IP

Foundry, memory & equipment

9

TSMC

Taiwan

Leading-edge semiconductor foundry

GlobalFoundries

United States

Specialty semiconductor foundry

Samsung Semiconductor

South Korea

HBM memory and foundry

Micron

United States

HBM and DRAM memory

SK Hynix

South Korea

HBM memory for AI accelerators

ASML

Netherlands

EUV lithography systems

Applied Materials

United States

Semiconductor fabrication equipment

Lam Research

United States

Wafer fabrication equipment

KLA

United States

Process control and inspection

Edge, photonics & emerging

14

Tenstorrent

Canada

RISC-V AI processors

Etched

United States

Transformer-specialized ASIC (Sohu)

d-Matrix

United States

In-memory compute inference chips

Hailo

Israel

Edge AI processors

SiMa.ai

United States

Edge MLSoC for embedded AI

Mythic

United States

Analog in-memory compute

Lightmatter

United States

Photonic compute and interconnect

Celestial AI

United States

Photonic fabric for AI memory

Ayar Labs

United States

Optical I/O interconnect

Astera Labs

United States

Connectivity for AI infrastructure

BrainChip

Australia

Neuromorphic edge AI (Akida)

EdgeCortix

Japan

Energy-efficient edge AI processors

Axelera AI

Netherlands

Edge AI inference acceleration

Esperanto

United States

RISC-V AI inference silicon

Regional

5

FuriosaAI

South Korea

RNGD inference accelerators

Rebellions

South Korea

ATOM AI inference chips

Untether AI

Canada

At-memory compute inference

Huawei Ascend

China

Ascend AI processors

Cambricon

China

AI training and inference chips