# L2 > NVIDIA · Ada Lovelace Canonical: https://gpuninja.com/chips/l2 Data as of: 2026-10-11T17:17:44Z Path: Home › GPU chips - Specs verified: between 11 Oct 2026, 17:15 UTC and 11 Oct 2026, 17:17 UTC ## Specifications - Vendor: NVIDIA (Verified · source docs.nvidia.com · retrieved 11 Oct 2026, 16:51 UTC · checked 11 Oct 2026, 17:15 UTC) - Architecture: Ada Lovelace (Verified · source docs.nvidia.com · retrieved 11 Oct 2026, 16:51 UTC · checked 11 Oct 2026, 17:15 UTC) - Die: AD104 (Verified · source admin.pci-ids.ucw.cz · retrieved 11 Oct 2026, 16:51 UTC · checked 11 Oct 2026, 17:17 UTC) - Process: TSMC 4N (Verified · source en.wikipedia.org · retrieved 11 Oct 2026, 16:52 UTC · checked 11 Oct 2026, 17:16 UTC) - Shader cores: Unknown - Tensor cores: Unknown - RT cores: Unknown - Memory bus: Unknown - Max VRAM: 24 GB (Verified · source docs.nvidia.com · retrieved 11 Oct 2026, 16:52 UTC · checked 11 Oct 2026, 17:15 UTC) - Reference boost clock: Unknown - Reference TDP: Unknown - Host interface: PCIe 4.0 x16 (Verified · source tomshardware.com · retrieved 11 Oct 2026, 16:50 UTC · checked 11 Oct 2026, 17:16 UTC) - Launch MSRP: No public MSRP - Launch date: Unknown - Segment: Data center (Verified · source docs.nvidia.com · retrieved 11 Oct 2026, 16:51 UTC · checked 11 Oct 2026, 17:15 UTC) - PCI IDs: 10de:27b6 (Verified · source download.nvidia.com · retrieved 11 Oct 2026, 16:49 UTC · checked 11 Oct 2026, 17:15 UTC) - Status: Active (Verified · source docs.nvidia.com · retrieved 11 Oct 2026, 16:51 UTC · checked 11 Oct 2026, 17:15 UTC) ## Performance - Memory bandwidth: 300 GB/s (Verified · source tomshardware.com · retrieved 11 Oct 2026, 16:50 UTC · checked 11 Oct 2026, 17:16 UTC) - FP32 compute: 24.1 TFLOPS (Verified · source tomshardware.com · retrieved 11 Oct 2026, 16:50 UTC · checked 11 Oct 2026, 17:16 UTC) - FP16 compute: 96.5 TFLOPS (Verified · source flopper.io · retrieved 11 Oct 2026, 17:15 UTC) - BF16 compute: 96.5 TFLOPS (Verified · source flopper.io · retrieved 11 Oct 2026, 17:15 UTC) - FP8 compute: 193 TFLOPS (Verified · source tomshardware.com · retrieved 11 Oct 2026, 17:16 UTC) - FP4 compute: Not supported - INT8 compute: 193 TOPS (Verified · source flopper.io · retrieved 11 Oct 2026, 17:15 UTC) ## Boards built on this chip Board, Chip, VRAM, TDP, Length, Slots, Price, Stock No boards recorded for this chip. ## Benchmarks No benchmark results recorded. ## Sources - https://docs.nvidia.com/grid/gpus-supported-by-vgpu.html — Vendor, Architecture, Segment, Status; last verified 11 Oct 2026, 17:15 UTC - https://admin.pci-ids.ucw.cz/read/PC/10de/27b6 — Die; last verified 11 Oct 2026, 17:17 UTC - https://en.wikipedia.org/wiki/Ada_Lovelace_(microarchitecture) — Process; last verified 11 Oct 2026, 17:16 UTC - https://docs.nvidia.com/ai-enterprise/release-8/latest/infra-software/vgpu/reference/ada-lovelace.html — Max VRAM; last verified 11 Oct 2026, 17:15 UTC - https://www.tomshardware.com/pc-components/gpus/new-nvidia-ai-gpus-designed-to-get-around-us-export-bans-come-to-china-h20-l20-and-l2-to-fill-void-left-by-restricted-models — Host interface, Memory bandwidth, FP32 compute, FP8 compute; last verified 11 Oct 2026, 17:16 UTC - https://download.nvidia.com/XFree86/Linux-x86_64/580.95.05/README/supportedchips.html — PCI IDs; last verified 11 Oct 2026, 17:15 UTC - https://flopper.io/gpu/nvidia-l2/spec-sheet — FP16 compute, BF16 compute, INT8 compute; last verified 11 Oct 2026, 17:15 UTC