AI Chips · Product page

Intel Gaudi 3: The Complete Guide

The first chip in the line to carry Intel’s own name — made by its biggest foundry rival, TSMC. Simple to expert, every number links to its source.

Stuck at any point? Ask an AI about this page →
Ask an AI about this chip:CAGPX
Intel Gaudi 3 10 sections · 2 levels 19 linked sources Checked September 2026
How to read this page. Each section starts Simple, then goes to a Deep dive: stop wherever you have what you need. The small numbers are sources: click one to open the original document. Where the maker has not published a figure -- a price, a die size, a factory address -- this page says so rather than estimate it.

1.At a glance

SimpleStart here

Gaudi 3 is Intel’s current AI accelerator, unveiled 9 April 2024 with systems reaching customers from late 2024 1 2. It comes as an OAM module and a PCIe card, with 128 GB of HBM2e memory — and unlike NVIDIA’s and AMD’s flagship accelerators, it is made by Intel’s own biggest foundry rival, TSMC, on a 5 nm process 3.

Deep diveThe technical detail

Intel’s own 2024 claims, stated as projections rather than measured results: Gaudi 3 is “projected to deliver 50% faster time-to-train on average” across specific named models (Llama 2 7B/13B and GPT-3 175B), and inference throughput is “projected to outperform the H100 by 50% on average and 40% for inference power-efficiency”, averaged across Llama 7B/70B and Falcon 180B 1; separately, Intel claims up to twice the price-performance on Llama 2 70B inference 2. These are comparisons to the H100, NVIDIA’s 2022 chip, not to its newer Blackwell or Rubin generations (see our NVIDIA guide on this site).

2.Launch and history

SimpleStart here

Gaudi 3 is the first chip in Intel’s Gaudi line to carry the “Intel” name in its own launch materials — the first two generations launched under the “Habana Gaudi” brand (see those pages on this site) 1 4.

Deep diveThe technical detail

At launch, Intel named customers and partners including NAVER, Bosch, IBM, Ola/Krutrim, NielsenIQ, Seekr, IFF, CtrlS Group, Bharti Airtel, Landing AI, Roboflow and Infosys 1. Inflection AI’s enterprise product runs on Gaudi 3, announced October 2024 5.

3.What’s inside it

SimpleStart here

Two compute dies joined in one package, with 8 matrix engines and 64 Tensor Processor Cores — up from 24 on Gaudi 2 — plus 24 ports of on-chip 200 Gigabit Ethernet, so chips link over ordinary Ethernet rather than a proprietary connection 6 3.

Deep diveThe technical detail

96 MB of on-chip memory backs the 64 tensor cores 3. Intel’s rack design holds up to 64 Gaudi 3 chips with 8.2 TB of combined memory 7.

4.Full spec table

SimpleStart here
SpecGaudi 3 (OAM)Gaudi 3 PCIe (HL-338)
Memory128 GB HBM2e 3128 GB HBM2e 3
Memory bandwidth3.7 TB/s 33.7 TB/s
AI compute1,678 TFLOPS BF16/FP8 (white paper) 3; 1,835 TFLOPS (Hot Chips slides) 6Not stated separately
Power900 W air 3; 1,200 W liquid, per Intel’s Hot Chips slides 6600 W 3
Made onTSMC 5 nm 3TSMC 5 nm

Intel publishes two peak figures: 1,678 TFLOPS in its white paper 3, and 1,835 TFLOPS in its Hot Chips 2024 slides, shown next to the 1,200 W liquid-cooled rating 6. Neither says whether the figure is dense or sparse.

Deep diveThe technical detail

China-specific cut-down versions, the HL-328 and HL-388, are limited to 450 W 8. In April 2025 Intel told Chinese customers that AI chips above set memory and connection-speed thresholds need a US export licence, which covers Gaudi 9.

5.Where it’s made

SimpleStart here

Designed by Intel’s Habana Labs team in Israel; manufactured by TSMC on a 5 nm process — Intel’s own AI accelerator, made by the foundry it competes against 3.

Deep diveThe technical detail

Intel has not published which of TSMC’s fabs makes Gaudi 3. This is a contrast with Intel’s Panther Lake and Xeon 6+ processors, which Intel makes in its own Arizona fab on its Intel 18A process (see the Intel brand guide on this site).

6.Which systems use it

SimpleStart here

Dell and Supermicro sell servers with eight Gaudi 3 chips 10 11. IBM Cloud, generally available from 31 March 2025 12, was “the first cloud service provider to make Intel Gaudi 3 AI accelerators available to customers,” per Intel’s own announcement the following month 13.

Deep diveThe technical detail

Availability expanded further at Computex 2025, when Intel unveiled new GPU configurations built around Gaudi 3 7, and again with a May 2025 push described by Intel as expanding “availability to drive AI innovation at scale” 10.

7.Official pricing

SimpleStart here

Unlike NVIDIA and AMD, Intel has published a list price for its AI accelerator. In June 2024 it said a kit of eight Gaudi 3 chips on a baseboard “will list at $125,000” — about $15,625 per chip by our arithmetic 14.

Deep diveThe technical detail

Rental prices are inconsistent across providers. Denvr Dataworks was “pre-selling now” with no published per-chip rate as of our check 15; IBM Cloud offers a quote tool only 16; Intel’s own Tiber AI Cloud pricing page lists no Gaudi hourly rate 17. India’s IndiaAI Mission compute portal lists one Gaudi 3 card (128 GB) at ₹153.00 an hour and a full 8-card server at ₹1,224, before GST — the same rate the portal lists for one NVIDIA H100 SXM card 18.

8.Real-world performance

SimpleStart here

Intel’s own 2024 claims, stated as projections: on average 50% faster time-to-train and 50% better inference throughput than NVIDIA’s H100, 40% better inference power efficiency (each figure averaged across specific named models, not a general claim), and up to twice the price-performance on Llama 2 70B inference 1 2.

Deep diveThe technical detail

These are Intel’s own comparisons, against NVIDIA’s H100 — a 2022 chip, not NVIDIA’s newest — and Intel’s own release describes them as “projected” figures, not measured benchmark results. The 50%/50%/40% figures are averages across named models only (Llama 2 7B/13B and GPT-3 175B for training; Llama 7B/70B and Falcon 180B for inference), not a blanket claim across all workloads 1. We found no equivalent Gaudi 3 comparison against NVIDIA’s newer Blackwell (B200/B300) or Rubin chips in the sources checked for this page (see the NVIDIA guide on this site for those chips’ own figures).

9.What came before, what came next

SimpleStart here
Came before: Gaudi 2, covered on its own page on this site. Came after: nothing has shipped yet — Intel dropped Falcon Shores as a product in January 2025, keeping it as an internal test chip only, and its next inference GPU, Crescent Island, had not launched as of September 2026.
Deep diveThe technical detail

Intel said in January 2025 that Falcon Shores, the chip once expected to follow Gaudi 3, would be used only as an internal test chip “without bringing it to market,” shifting focus to Jaguar Shores, a rack-scale system, instead 19. Unlike Gaudi 3 itself — which went from an April 2024 unveiling to shipping systems within months — no successor accelerator had reached general availability as of this page’s last check, making Gaudi 3 Intel’s current accelerator for longer than any earlier Gaudi generation held that position.

10.Hidden in plain sight

SimpleStart here

Gaudi 3 costs the same per hour as NVIDIA’s H100 SXM on India’s IndiaAI Mission compute portal: ₹153 for either chip, before GST 18.

Deep diveThe technical detail

Intel published a list price for its AI accelerator — something neither NVIDIA nor AMD does for their flagship chips — yet built that chip using TSMC, the foundry business Intel itself competes against for outside customers 14 3. And despite Gaudi 3 losing its planned direct successor when Falcon Shores was shelved, Intel has kept selling it: the IBM Cloud general-availability date (March 2025) and the “expands availability” push (May 2025) both came after the January 2025 Falcon Shores announcement, not before it 19 12 10.

11.Sources

19 sources, checked September 2026. Where NVIDIA or another maker has not published a figure, this page says so rather than estimate it.

  1. Intel Investor Relations: Intel unleashes enterprise AI with Gaudi 3 (9 Apr 2024)Official
  2. Intel Investor Relations: Intel launches Xeon 6 and Gaudi 3 (24 Sep 2024)Official
  3. Intel: Intel Gaudi 3 AI Accelerator white paperOfficial
  4. Intel: Habana Gaudi2 launch fact sheet (10 May 2022)Official
  5. Intel Newsroom: Inflection AI and Intel launch enterprise AI system (7 Oct 2024)Official
  6. Intel at Hot Chips 2024: Intel Gaudi 3 (Roman Kaplan)Conference
  7. Intel Investor Relations: Computex 2025: Intel unveils new GPUs for AI (May 2025)Official
  8. The Register (press): Intel’s China-specific Gaudi 3 parts (12 Apr 2024)Press
  9. TrendForce (press): Intel reportedly faces new US licence rules for Gaudi in China (17 Apr 2025)Press
  10. Intel Newsroom: Intel Gaudi 3 expands availability (May 2025)Official
  11. Supermicro: SYS-822GA-NGR3 (8x Gaudi 3) product pageOfficial
  12. IBM Newsroom: Intel and IBM announce availability of Gaudi 3 on IBM CloudOfficial
  13. Intel Newsroom: IBM Cloud is First Service Provider to Deploy Intel Gaudi 3 (22 Apr 2025)Official
  14. Intel Investor Relations: Intel accelerates AI everywhere at Computex 2024 (4 Jun 2024)Official
  15. Denvr Dataworks: PricingOfficial
  16. IBM: Intel Gaudi 3 AI accelerators on IBM CloudOfficial
  17. Intel Tiber AI Cloud: PricingOfficial
  18. IndiaAI Mission (Government of India): IndiaAI compute price calculatorGovernment
  19. Intel: Q4 2024 earnings call comments (30 Jan 2025)Official

Ask an AI about this page

Opens your assistant with this page as the source, and a question rather than a summary. It will ask what you are building before it answers.

ChatGPTClaudeGeminiPerplexityGrok

Nothing is sent from here. The link carries only this page’s title and address.