Broadcom custom accelerators
Co-designed XPUs behind most hyperscaler silicon.
Full Broadcom custom accelerators record →Head to head
Broadcom custom accelerators vs NVIDIA Blackwell
Broadcom custom accelerators (Broadcom) and NVIDIA Blackwell (NVIDIA) both sit in Chips & Compute. NVIDIA Blackwell shipped more tracked changes in the last 30 days (9 vs 0). Every value below comes from the latest crawl of the vendors' own pages.
| Attribute | Broadcom custom accelerators Broadcom | NVIDIA Blackwell NVIDIA |
|---|---|---|
| Segment | Chips & Compute | Chips & Compute |
| Version | 2025 | Blackwell |
| Entry cost | Enterprise supply agreements | Pricing is not publicly listed (enterprise hardware architecture/system) |
| Pricing tiers | — | — |
| Model stack | powers Google TPU | Blackwell · Blackwell B200 · Grace CPU |
| Context window | — | — |
| Public API | ||
| Routes models | ||
| Changes / 30d | 0 | 9 |
| Origin | Global | Global |
| Capabilities | Custom XPU design, Tomahawk networking, Hyperscaler programs | 72-GPU NVLink domain, FP4 inference, HBM3E 192GB, CUDA ecosystem, Second-generation Transformer Engine, Fifth-generation NVLink, FP4 AI Support, Tensor Cores with microscaling formats |
NVIDIA Blackwell
NVL72 rack-scale coherent GPU domains, the default substrate for frontier training and inference.
Full NVIDIA Blackwell record →