AI chip · Amazon
AWS Inferentia 2
Second-generation AWS custom inference accelerator; the Inf2 EC2 instance family. Sits alongside Trainium in AWS's LLM stack.
Market position
AWS's inference-side silicon for LLM serving on Inf2 EC2 instances. Positioned against NVIDIA L4/L40S on cost-per-token; used inside Bedrock for smaller-model inference. Trainium 2 has replaced it as the general-purpose choice at hyperscale; Inferentia 2 still ships for cost-optimised inference workloads.
DEPLOY editorial. What the vendor PDF cannot tell you.
Physical-AI cross-link
Derived from claims already on record (vendor first-party TDP, throughput, price, and the reference-design cluster registry). The arithmetic is shown per row so it can be audited. No estimated inputs.
- Perf per watt
- 1.52 FP8 TFLOPS/W380 TFLOPS ÷ 250 W = 1.52
What fits in 32 GB
Weights-only footprint for public open-weight LLMs. Pure arithmetic: params × bytes/param. Excludes KV cache and activation memory; add ~10-30% headroom for real serving. A model that does not fit at FP16 may still fit at INT8 or INT4 with quality trade-offs. Not a benchmark.
| Model | Params | FP16 | INT8 | INT4 |
|---|---|---|---|---|
| Llama 3.1 8B dense | 8 B | 16 GB ✓ | 8 GB ✓ | 4 GB ✓ |
| Llama 3.1 70B dense | 70 B | 140 GB ✗ | 70 GB ✗ | 35 GB ✗ |
| Llama 3.1 405B dense | 405 B | 810 GB ✗ | 405 GB ✗ | 203 GB ✗ |
| Llama 3.3 70B dense | 70 B | 140 GB ✗ | 70 GB ✗ | 35 GB ✗ |
| DeepSeek V3 MoE (671B total, 37B active per token) | 671 B | 1342 GB ✗ | 671 GB ✗ | 336 GB ✗ |
| DeepSeek R1 MoE (671B total, 37B active per token) | 671 B | 1342 GB ✗ | 671 GB ✗ | 336 GB ✗ |
| Qwen 2.5 7B dense | 7 B | 14 GB ✓ | 7 GB ✓ | 4 GB ✓ |
| Qwen 2.5 72B dense | 72 B | 144 GB ✗ | 72 GB ✗ | 36 GB ✗ |
| Mixtral 8x7B MoE (46.7B total, 12.9B active per token) | 46.7 B | 93 GB ✗ | 47 GB ✗ | 23 GB ✓ |
| Mixtral 8x22B MoE (141B total, 39B active per token) | 141 B | 282 GB ✗ | 141 GB ✗ | 71 GB ✗ |
| Gemma 2 27B dense | 27 B | 54 GB ✗ | 27 GB ✓ | 14 GB ✓ |
| Command R+ dense | 104 B | 208 GB ✗ | 104 GB ✗ | 52 GB ✗ |
| Kimi K2 MoE (1T total, 32B active per token) | 1000 B | 2000 GB ✗ | 1000 GB ✗ | 500 GB ✗ |
Math: FP16 = params × 2 bytes; INT8 = params × 1 byte; INT4 = params × 0.5 bytes. MoE models sum every expert (full weights on disk), not the per-token active subset.
Common questions
Answer-first, sourced. Every claim below traces to a specific field on this page or a cited datasheet. FAQPage schema is emitted so LLM crawlers can lift these verbatim.
How much does AWS Inferentia 2 cost?
AWS Inferentia 2 launch pricing is not on record in the DEPLOY registry. Vendor list prices are typically published in the vendor's own briefing rather than the datasheet.
How much memory does AWS Inferentia 2 have?
32 GB of HBM3 at 820 GB/s. Vendor datasheet, first-party.
How much power does one AWS Inferentia 2 draw?
250 W TDP (thermal design power) per chip. System-level draw is higher: see the reference-design section for rack-level kW.
Which open-weight LLMs fit on one AWS Inferentia 2?
Of the 13 public open-weight LLMs DEPLOY tracks: 2 fit at FP16, 3 at INT8, 4 at INT4 (weights only, excludes KV cache). The full table is above with per-model math. Sparsity, offloading and multi-GPU serving change the picture; this row is single-chip weights-only.
Key facts
- Class
- Compute SoC (AI accelerator)
- Designer
- Amazon
- Safety-critical?
- No (data-center inference / training)
- Record as of
- 2026-08-25
Specifications (9 fields, click to expand)
Vendor datasheet figures (first-party). Dense throughput first; sparse (2:4) in parentheses where NVIDIA quotes it. The exhaustive spec sheet lives on the datasheet URL below the table: this row set covers what buyers actually compare on.
- Process node
- TSMC N7
- TDP
- 250 W
- Memory
- 32 GB HBM3
- Memory bandwidth
- 820 GB/s
- BF16 (dense)
- 190 TFLOPS
- FP8 (dense)
- 380 TFLOPS
- Form factor
- Inf2 EC2 instance
- Announced
- 2022-12-01
- Released
- 2023-04-01
Source: vendor datasheet
Compare with
Sources
Adoption rows appear as we verify chip-integration claims per model. Every row is a public claim the maker or a tier-1 source has stated; verification tier (verified / reported / inferred) is shown per row. See every chip on record for the full catalog or the Amazon record.