Tenstorrent AI accelerator cards
Ranked 8 of 8 AI accelerators by recommenders
Developers recommend Tenstorrent cards as an accessible underdog alternative to Nvidia and AMD, and praise their predictable execution design. Critics say the software stack is immature and power efficiency is poor.
Developers recommend Tenstorrent cards as an accessible underdog alternative to Nvidia and AMD, and praise their predictable execution design. Critics say the software stack is immature and power efficiency is poor.
10 people gave an opinion
- recommend it
- 7
- criticize it
- 3
- approval share
- 70%
Recommended in 5 of 13 months
- November 2025: not recommended
- December 2025: recommended
- January 2026: not recommended
- February 2026: recommended
- March 2026: not recommended
- April 2026: not recommended
- May 2026: recommended
- June 2026: recommended
- July 2026: not recommended
- August 2026: recommended
- September 2026: not recommended
- October 2026: not recommended
- November 2026: not recommended (month in progress)
Why it doesn’t carry the badge
- 7 people recommend itNeeds 10 or more, not met.
- 2.3 recommenders for every criticNeeds 2 to 1 or better, met.
- Recommended in 5 of 13 monthsNeeds 4 or more, met.
The same three rules apply to every product, and the badge can’t be bought. How the badge works
What people say about it
The reasons people give, with the number of comments that give each one. Open a reason to read what they said.
Why people recommend it
Each mark is one comment
Accessible underdog for hobbyists
“If you insist on an hobbyist accessible underdog, you'd go with Tenstorrent, not AMD.”
Comment from 18 Feb 2026 in “BarraCUDA Open-source CUDA compiler targeting AMD GPUs” “Tenstorrent is awesome, can't wait to see if I can afford any of their hardware in 5ish years.”
Comment from 20 May 2026 in “RISC-V and Floating Point” Strong performance for its class
“And Tenstorrent greatly exceeds it, but how will you actually get one of those cards?”
Comment from 29 Aug 2026 in “GLM-5.3 is now open-weight” “We'll eventually get similar performance locally with tenstorrent hardware.”
Comment from 9 Jun 2026 in “Claude Fable 5”
What people warn about
Each mark is one comment
Weak outside matrix operations
“TT hardware is only good for matmuls and some elementwise operations, and plain sucks for anything else.”
Comment from 27 Nov 2025 in “TPUs vs. GPUs and why Google is positioned to win AI race in the long term” “Tenstorrent is selling chips that are pretty weak”
Comment from 28 Nov 2025 in “Vsora Jotunn-8 5nm European inference chip”
Other AI accelerators people recommend
The four most recommended other AI accelerators, with the same counts for each, so you can compare like for like. The tick on each bar marks the 2 to 1 line the badge needs.
| Product | Recommend | Criticize | Approval | Months | Owned | Badge |
|---|---|---|---|---|---|---|
| Google TPU | 73 | 15 | 83% | 11 | — | Approved |
| Taalas HC1 | 59 | 15 | 80% | 8 | — | Approved |
| Taalas hardwired LLM chip | 45 | 8 | 85% | 8 | — | Approved |
| Cerebras wafer-scale chip | 29 | 8 | 78% | 10 | — | Approved |
| Tenstorrent AI accelerator cardsThis page | 7 | 3 | 70% | 5 | — | — |
What people said
All 10 comments that mention it, newest first. Each links to the original comment.
“And Tenstorrent greatly exceeds it, but how will you actually get one of those cards?”
“Groq and Tenstorrent are in a similar boat, ASICs don't really threaten CUDA”
“The software stack is pretty immature, definitely very DIY.”
“We'll eventually get similar performance locally with tenstorrent hardware.”
“Designing for predictable execution flow is one of the advantages of Tenstorrent hardware.”
“Tenstorrent is awesome, can't wait to see if I can afford any of their hardware in 5ish years.”
“If you insist on an hobbyist accessible underdog, you'd go with Tenstorrent, not AMD.”
“The chips don't strike me as particularly power efficient (especially at idle which matters for developers)”
“Tenstorrent is selling chips that are pretty weak”
“TT hardware is only good for matmuls and some elementwise operations, and plain sucks for anything else.”
Counts come from 1,795,640 Hacker News comments, November 2025 to November 2026. How we count. Report a mistake.