NVIDIA Groq 3 LPX
Rack-scale inference chip built on the LP30 die, from Nvidia's Groq acquisition
N/A
Datacenter
0.5GB
SRAM
150
GB/s bandwidth
4nm
process
Groq 3 LPX is the first chip to ship from Nvidia’s $20 billion acquisition of Groq, announced in production at Hot Chips 2026. Each LP30 die carries around 500MB of on-chip SRAM with 150 TB/s of bandwidth and handles only token generation, not training or prefill. Nvidia pairs the chips into liquid-cooled racks of 256 LPUs as a companion to its Vera Rubin platform, aimed at the low-latency demands of agentic AI workloads.
// similar hardware in Datacenter
NVIDIA B200
NVIDIA Blackwell flagship datacenter GPU for frontier model training
NVIDIA · Datacenter · N/A
NVIDIA H100 SXM
Hopper-architecture datacenter GPU for large-scale AI training
NVIDIA · Datacenter · ~$30000
NVIDIA H200
High-memory Hopper refresh for large-model inference
NVIDIA · Datacenter · ~$32000
AMD Instinct MI300X
AMD flagship datacenter accelerator for large-model AI training
AMD · Datacenter · ~$15000
AMD Instinct MI325X
High-memory CDNA 3 refresh for large-model inference
AMD · Datacenter · ~$16000
NVIDIA A100 80GB
Ampere-generation datacenter GPU, still widely deployed
NVIDIA · Datacenter · ~$15000