Tesla (TSLA) just delivered a rare double whammy to Nvidia (NVDA) last weekend.
CEO Elon Musk revealed that Tesla got a lot of buzz AI5 self-driving chips is almost complete, and another, the A16, is already underway.
The AI speculation part is covered, Musk said on Sunday at X Dojo 3 is starting again, pushing Tesla into large-scale AI training after previously pulling back.
Nvidia threw the first punch, though, when it rolled out “little love(an open-source autonomous vehicle AI toolkit), aiming to be at CES 2026 Default Autonomy Platform Powering a ton of brands.
Kasturi responded quickly, downplaying the risk.
Clearly, it’s a mighty interesting time for the AV industry, with a tug-of-war between two giants in Nvidia and Tesla.
For Tesla, it’s all about building a closed loop covering the AV stack.
Tesla-designed in-car compute (This includes AI5, which is “almost finished,” and AI6, which is already underway)
Tesla’s camera-first software stack
Tesla’s data flywheel is powered by its own fleet
So for Tesla, it’s all about keeping a piece of autonomy within its powerful ecosystem, as Nvidia powers everything else.
For investors, these promises aren’t new, which makes follow-through all the more important.
Elon Musk says Tesla’s AI5 chip is nearing completion as next-generation self-driving hardware advancesPhoto by Bloomberg at Getty Images” loading=”eager” height=”640″ width=”960″ class=”yf-lglytj loader”/>
Elon Musk says Tesla’s AI5 chip is nearing completion as next-generation self-driving hardware advancesPhoto by Bloomberg at Getty Images ·Photo by Bloomberg at Getty Images
Tesla is looking to tighten its grip on the hardware behind self-driving.
More Tech Stocks:
Musk announced in an X post on Saturday that the EV giant is close to completing its AI5 self-driving computer chip, and that AI6 is already in development.
According to Musk, AI5 chips manufactured by Taiwan Semiconductor Manufacturing Company will be imported. High volume production in 2027Replacing the AI4 hardware. Also, Tesla has lined up Samsung Electronics for US-based chip manufacturing.
It’s very easy to get lost in AI jargon, so it’s important to be clear about things at every step of what’s going on.
Related: BlackRock CEO Gives Clear Warning on US National Debt
A15 and A16 moves are essentially about “guessing at the edge”. That’s basically what’s going on Tesla’s fully self-driving neural nets inside the car instead of relying on a third-party compute stack.
So if Tesla runs the software on its own chips, it gains a huge competitive edge:
Tesla doesn’t need Nvidia’s in-vehicle SoC (or its full “drive” platform) for its cars.
Tesla gains control of unit costs, supply-chain leverage, and chip design.
It’s important to note, though, that Tesla has already moved away from Nvidia for this In-car countdown back to 2019So the last moves are twice as low as the switch.
Nvidia is offering automakers a full-stack solution, essentially a shortcut to full autonomy. Under it NVIDIA Drive Umbrella, it’s basically selling an integrated “brain, operating system, and toolkit.”
RELATED: Goldman Sachs Revises Microsoft Stock Price Target Ahead of Earnings
So instead of building custom chips, software, security frameworks, and whatnot, automakers can plug and play into Nvidia’s robust ecosystem and get started. A big part of its appeal is Tesla’s decades-long autonomy effort or essentially a hack for companies that don’t have billions to spend on R&D.
What Nvidia bundles together:
Drive AGX In-vehicle computers, eg Orin and Thor.
A complete software stack that includes Drive OS and DriveWorks.
Drive HyperionA reference vehicle platform that comes with validated sensors and architecture.
Security and authentication tools under popular NVIDIA Halos Umbrella, like powerful AI models little loveAt the speed of training and simulation.
Tesla is sharpening its in-car AI chips, but clearly, Nvidia still holds a significant edge in computing power.
The AI5 and AI6 are tailor-made for edge estimation, however Training frontier-scale models is a completely different challenge.
Training modern AI systems is remarkably compute-hungry.
For perspective, Meta said it trained its AI model using Lama 3.1 (405B). Over 16,000 Nvidia H100 GPUs. So if we factor in 700 watts per chip, that’s approx 11.2 MW power For GPU only. That level is where Nvidia’s economics, availability, and ecosystem dominate.
However, Tesla’s decision to relaunch the Dojo 3 looks to re-enter the training game.
At this point, though, I realize Return of the Dojo 3 Possibly indicating a hybrid future.
Tesla will continue to build on its training capabilities using the AI5 and AI6 architectures, while still banking on Nvidia, where scale and economics matter.
When we see strong evidence of large-scale training clusters running on Tesla silicon, supported by throughput and cost data, the competition really increases on the training front.
Related: Morgan Stanley Tweaks Rocket Lab Stock Price Target Post-Rally
This story was originally published by TheStreet on January 19, 2026, where it first appeared in the Technology section. Add TheStreet as a preferred source by clicking here.
