AIToday
Top Companies' AI MovesTop Companies AIPublished: Nov 8, 2026, 06:31 JST

Tesla's SI6 chip targets 80X more compute than Gen 4

Send on LINE
Tesla's SI6 chip targets 80X more compute than Gen 4

3 Key Points

  1. What happened

    Tesla unveiled a chip roadmap under 'SI' (Super Intelligence) branding, saying its SI6 inference accelerator will bring 80X more compute performance than the existing 4th-generation chip.

  2. Why it matters

    The roadmap signals Tesla intends to keep designing its own AI silicon rather than rely on outside suppliers, with SI6 pitched as a major jump over its current inference hardware.

WHO IT HITSThe roadmap matters most to AI hardware engineers and supply-chain planners tracking Tesla's custom silicon, and to teams weighing whether Tesla's in-house accelerators could eventually compete with merchant AI chips. Tesla's stated 80X figure is a company projection, not an independently measured result.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

Tesla's disclosure puts concrete direction behind a chip program it has kept largely internal. The roadmap now spans two tracks: inference chips, where SI6 is pitched at 80X the compute of the 4th-generation part, and training chips, where D3 abandons an SRAM-centric architecture for a multi-chiplet design with stacked HBM. That shift suggests Tesla is rethinking how its training hardware is built, not just how fast it runs.

The naming change is the other signal. Tesla is applying 'super intelligence' branding across its accelerators, and Musk said SpaceXAI would be redubbed SpaceXSI for the same reason — indicating the terminology is being coordinated across his companies rather than used in one product line.

What the roadmap does not include is timing. Tesla disclosed no dates for SI6 or D3, and the 80X figure is a company projection against its own 4th-generation chip rather than a benchmark against other vendors' hardware.

FAQ
How much faster is Tesla's next chip?
Tesla expects its 6th-generation inference accelerator, SI6, to bring 80X more compute performance than the existing 4th-generation chip.
What changes in Tesla's training chip?
The D3 training accelerator will stop using an SRAM-centric architecture and instead adopt a multi-chiplet design with stacked HBM.
Why is Tesla using 'SI' branding?
Tesla has adopted the term 'super intelligence' instead of artificial intelligence for its accelerators, renaming them with SI monikers. Musk said SpaceXAI would be redubbed SpaceXSI for the same reason.
Top Companies AIRead Original Article
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleArista revenue jumps 58.5% to $3.66 billion, a record