Your computer chips supplier for data-center GPUs, server and desktop CPUs, memory, networking, microcontrollers and power semiconductors. Prices are in USD with the date they were last checked, quantity pricing follows real price breaks, and anything without a listed price is quoted on request. Based in the USA, with worldwide shipping.
NVIDIA Hopper-architecture data center accelerator in a dual-slot, passively cooled PCIe card for AI training/inference and HPC servers. Requires ser…
NVIDIA Hopper H200 NVL data center GPU, passive dual-slot PCIe card for LLM inference and HPC in NVIDIA-certified servers. GPU only (no power adapter…
NVIDIA Blackwell data center GPU in SXM module form factor with 180 GB HBM3e, for HGX B200 baseboards. Sold in multi-unit quantities only.
NVIDIA Blackwell professional workstation GPU with 96 GB ECC GDDR7 for AI development, rendering and simulation; dual-slot, double flow-through coole…
AMD CDNA 3 data center accelerator with 192 GB HBM3 for large-model AI training and inference.
NVIDIA GeForce RTX 5090 graphics card with 32 GB GDDR7 (512-bit) and MSI VENTUS 3X triple-fan cooler, factory overclocked.
NVIDIA Ampere-architecture data center accelerator, dual-slot passive PCIe card with ECC, for AI and HPC servers.
NVIDIA Ada Lovelace universal data center GPU for generative AI inference, training and graphics workloads; dual-slot passive PCIe card.
A high-frequency 5th Gen EPYC (Turin) processor that boosts up to 5 GHz, popular as the host CPU in GPU servers where fast single-thread response kee…
A 64-core Xeon 6 performance-core processor from the 6700P line, built for mainstream two-socket servers on the FCLGA4710 platform. It supports eight…
Samsung's 12-high HBM3E stack, sold under the Shinebolt name, used alongside AI accelerators on 2.5D packages. HBM is supplied to GPU and ASIC makers…
SK hynix's 12-layer HBM3E, one of the main memory stacks used with current NVIDIA and AMD AI accelerators. It is sold directly to chip and package ma…
A PCIe version of NVIDIA's Hopper data center GPU built for large language model inference. Each card carries 94 GB of HBM3 and can be bridged in pai…
The SXM5 module form of NVIDIA H100 for HGX H100 baseboards, where eight GPUs are linked through NVSwitch. It is the higher-power H100 variant, aimed…
Blackwell Ultra GPU module for NVIDIA HGX B300 systems, positioned above B200 for large-scale AI reasoning and training. It is sold as a module for H…
The SXM module version of NVIDIA H200, used on HGX H200 four- and eight-GPU baseboards. It pairs the Hopper GPU with 141 GB of HBM3e, which helps lon…
NVIDIA · AMD · Intel · Texas Instruments · STMicroelectronics · Infineon Technologies · Microchip Technology · Analog Devices · Samsung · NXP Semiconductors · Micron Technology · Espressif Systems · All manufacturers
From individual components to advanced AI accelerators and enterprise data-center hardware.
STMicroelectronics VN7140ASTR Automotive High-Side Switch
$1.06
Individual ICs and modules →
Texas Instruments TPS54331DR 3A Step-Down Converter
$1.13
Regulators, converters and power devices →
Microchip AVR128DA48-I/PT 8-bit AVR DA MCU
$1.45
MCUs for embedded and IoT designs →
Intel Core Ultra 5 245K Desktop Processor
$150.49
Desktop and embedded CPUs →
ASUS Prime GeForce RTX 5080 16GB OC
$1,119.99
Graphics and workstation cards →
AMD EPYC 9554 64-Core Server Processor
$1,929.89
Server CPUs, adapters and data-center parts →
NVIDIA B300 (Blackwell Ultra) SXM GPU Module
$43,400.00
Data-center GPUs and AI accelerators →
Example listings from our catalog with their current verified price. Prices and availability may change without notice.
GPUs, CPUs, memory, networking and the supporting power hardware for AI training and inference clusters. Quote a single node or a full deployment.
AI GPUs & accelerators
H100, H200, B200, MI300X and other data-center accelerators.
17 listed
Server CPUs
AMD EPYC and Intel Xeon processors for AI and compute nodes.
11 listed
HBM memory
High-bandwidth memory used on AI accelerators.
3 listed
DDR5 server memory
DDR5 RDIMMs and DRAM for servers.
7 listed
Networking
InfiniBand and Ethernet adapters, DPUs and switch silicon.
13 listed
Data-center hardware
Servers, boards and platform components.
Quote on request
Power & cooling components
Power stages, controllers and protection for AI servers.
Quote on request
Need 100, 1,000 or 10,000+ units?
Send us the part numbers, quantities and delivery destination. We reply with unit pricing for your volume, lead times and shipping options.
We source parts beyond our listed catalog.
Tell us the manufacturer, part number and quantity you need. Our sourcing team will check availability.
As a semiconductor components supplier based in the USA, we offer electronic components wholesale with worldwide shipping. We quote business buyers around the world, including:
We ship worldwide. International shipping, duties and taxes are quoted before your order is confirmed. Send your part numbers, quantities and delivery country with a quote request.
Prices are shown in US dollars with the date they were last checked. Quantity pricing is shown only where real price breaks exist.
We source through our supplier network and do not hold inventory. Each listing shows the current supplier availability, lead time or backorder status.
Items without a verified price, large quantities and parts we don't list are all quoted on request. Request a quote.
