SambaNova logo

SambaNova SN40L

Type: ASICArchitecture: RDUProcess: 5nmSpec confidence: Official
Memory
64 GB
HBM
Bandwidth
1.8 TB/s
memory

Overview

The SambaNova SN40L is the fourth-generation Reconfigurable Dataflow Unit and the chip behind the SambaRack system. Rather than a fixed matrix pipeline it is a coarse-grained reconfigurable array of 1,040 Pattern Compute Units and 1,040 Pattern Memory Units on a two-dimensional mesh, delivering 638 TFLOPS of BF16. Its defining feature is a three-tier memory system: 520 MB of on-chip distributed SRAM, 64 GB of on-package HBM at 1.8 TB/s, and 1.5 TB of off-package DDR at 200 GB/s. That third tier is the point of the design, letting a single socket hold many models in DDR and switch between them by moving weights into HBM in microseconds, which is how SambaNova runs a 150-expert, trillion-parameter Composition of Experts on eight sockets. The chip is built on TSMC 5nm as a dual-die socket using CoWoS packaging. Note that SambaNova publishes no specification table on its website and no datasheet; these figures come from Table II of its own architecture paper, and the memory tiers are independently confirmed by the SambaRack totals in SambaNova's product brief. SambaNova states the die size only as under 650 mm2 and the clock only as under 2 GHz, so neither is recorded here as a precise value, and no power figure is published anywhere. Neither the paper nor the brief labels the compute figure dense or sparse, and no sparse figure is published.

Performance

Peak theoretical throughput by precision type

PrecisionPeak
FP64
No verified data available
FP32
No verified data available
TF32
No verified data available
BF16
Brain Float 16
638TFLOPS
FP16
No verified data available
FP8
No verified data available
FP6
No verified data available
FP4
No verified data available
INT8
No verified data available

Specifications

Architecture

RDU

Form Factor

No verified data available

Launch Year

No verified data available

Process Node

5nm

Memory

64 GB HBM

Bandwidth

1,800 GB/s

TDP

No verified data available

Max power

No verified data available

PCU

1,040

Spec Confidence

Official

Full Specifications

Compute Engine
PCU 1,040
Chip Design
Process Node 5nm
Chiplets 2 (Dual-die socket packaged with TSMC CoWoS)
Memory
Memory 64 GB
Memory Type HBM
Bandwidth 1.8 TB/s
Interface Width No verified data available
On-Die SRAM 520 MB
Power & Thermal
TDP No verified data available
Max power No verified data available
General
Form Factor No verified data available
Architecture RDU
Process Node 5nm
Launch Year No verified data available

Datasheet & Resources

Data Provenance

Every figure traced to a source

Data Quality

Spec confidence
Official
Clock basis
Boost
Core precisions with figures
1 of 9
Normalization
All values in TFLOPS

Compare cloud providers offering on-demand GPU instances for AI training, inference, and HPC workloads.

Browse GPU Cloud Providers

Similar GPUs

SambaNova SN50

Custom · 2026
FP32: No verified data available
Compare vs SN50

Systems Using This GPU

Pre-configured systems featuring the SambaNova SN40L

SystemGPU CountPeak PerformanceTotal Power
SambaNova SambaRack SN40L-16
SambaRack · rack ·
16x SN40L 10.2 PFLOPS 10.0 kW View System

Frequently Asked Questions

How many TFLOPS does the SambaNova SN40L have?

The verified peak figures Flopper holds for the SambaNova SN40L are 638 TFLOPS BF16. Flopper does not currently have verified FP32, FP16 and FP8 throughput figures for it.

What is the power consumption of the SambaNova SN40L?

Flopper does not currently have a verified TDP (Thermal Design Power) figure for the SambaNova SN40L.

How much memory does the SambaNova SN40L have?

The SambaNova SN40L is equipped with 64 GB of memory with 1,800 GB/s of memory bandwidth.

What architecture is the SambaNova SN40L based on?

The SambaNova SN40L is based on the RDU architecture. Flopper does not currently have a verified launch year for it.

Stay Updated on GPU Releases

Get notified when new GPUs are added or specifications are updated.

Loading verification...

No spam, unsubscribe anytime.

Back to GPUs

© 2026 Flopper.io - Compare the GPUs Powering AI