TECH NEWS – Dojo3 is back on track, and Elon Musk claims the AI5 chip could beat Nvidia’s Blackwell on price-to-performance.
Musk has been loudly marketing Tesla’s chipmaking ambitions for quite a while, especially after his multi-trillion-dollar pay package was approved. He previously floated plans around a TeraFab, arguing that Samsung and TSMC won’t be able to satisfy Tesla’s future chip demand. Now, in posts on Twitter, the CEO not only suggested the company is planning its in-house silicon four generations ahead, but also signaled that the Dojo3 supercomputer effort is back in motion. Musk also offered a glimpse of what the AI5 silicon is supposed to deliver.
Just a few months ago, Tesla said it was winding down the Dojo supercomputer project, given the automaker’s reliance on hardware from companies like Nvidia. The decision reportedly followed the departure of key figures tied to Dojo, which ultimately pushed the company to shelve the effort. Musk’s latest announcement, however, points to Dojo3 being back on the table – a sign that Tesla urgently needs compute, and that’s exactly why it’s doubling down on its aggressive in-house chip roadmap.
Now that the AI5 chip design is in good shape, Tesla will restart work on Dojo3.
If you’re interested in working on what will be the highest volume chips in the world, send a note to [email protected] with 3 bullet points on the toughest technical problems you’ve solved.
— Elon Musk (@elonmusk) January 18, 2026
Solving AI5 was existential to Tesla, which is why I had to focus both the teams on that chip and I’ve personally spent every Saturday for several months working on it.
This will be a very capable chip. Roughly Hopper class as single SoC and Blackwell as dual, but it costs…
— Elon Musk (@elonmusk) January 19, 2026
Details on Dojo3 remain scarce, but based on Musk’s posts, it’s reasonable to expect an AI5-based cluster – especially given Tesla’s stated intent to standardize its silicon across vehicles and Optimus. The CEO also framed AI5 as a “highest volume” chip, hinting at far broader chip-and-cluster configurations than prior efforts. He added that the custom silicon lineup is meant to keep evolving, claiming Tesla plans to expand all the way to AI9 on a nine-month product cadence – notably similar to how Nvidia operates. Musk is going all-in here because Tesla wants to exploit a cost-structure advantage once FSD becomes mainstream, and controlling the chip supply chain would let the company tune silicon to its own needs, creating leverage against competitors.
On performance, Musk said Tesla is targeting roughly Hopper-class capability in a single-chip setup, with a dual-chip configuration that can rival Blackwell. He also claimed AI5 will be “pocket change,” reinforcing the idea that Tesla wants a cost-effective silicon ecosystem at scale. Musk is famously optimistic, but chip work is an art that takes decades to master, given the sheer number of constraints that have to be balanced.
Could Tesla become the next major chipmaker? Musk’s comments point in that direction, but execution will be everything: Tesla would need to cover every stage of the pipeline, from design through validation and long-term silicon stability.
Source: WCCFTech
