Altimeter Capital’s Freda Duan said broader access to Nvidia’s full self-driving stack could change how autonomy is adopted across the industry.
- Nvidia is offering the full self-driving “brain” for the first time, with real-world performance yet to be proven at scale, according to Altimeter’s Freda Duan.
- Duan highlighted differences in training spend and data disclosures between Nvidia and Tesla.
- Duan said on-road deployments will be the key signal for evaluating Nvidia’s approach.
Freda Duan, a partner at Altimeter Capital, said Nvidia’s move to offer a full, end-to-end autonomous driving stack has introduced new uncertainty into how the market values Tesla’s self-driving story, potentially leaving the EV maker trading with a higher, sentiment-driven discount rate.
Nvidia’s Shift From Chips To A Full AV Stack
Duan described Nvidia’s latest push on X as a potential “Android moment” for self-driving if it succeeds, referring to the possibility that a broadly available, full autonomy “brain” could reshape how autonomous technology is adopted across the industry.
She said Nvidia has been involved in autonomous driving for years with mixed results, but that this marks the first time the company is offering the full decision-making system, rather than only the picks-and-shovels infrastructure stack. Duan added that whether Nvidia’s model performs at production-grade levels across real-world edge cases remains an open question until it is tested on the road at scale.
Her comments come as Nvidia broadens its autonomous driving push alongside new AI hardware. At the Consumer Electronics Show in Las Vegas, Nvidia CEO Jensen Huang said the company’s next-generation Vera Rubin chips are in full production and highlighted Alpamayo, a suite of autonomous driving models, simulation tools and datasets designed to help vehicles reason through complex road scenarios.
Nvidia said Alpamayo and the data used to train it will be released more broadly so automakers can evaluate performance.
Training Spend And Long-Term Economics
Duan said Tesla’s self-driving training spend is estimated at $3 billion - $4 billion in fiscal 2024, with roughly $5 billion per year likely required to sustain its edge. She said that if Nvidia’s approach is successful, Tesla’s terminal market share will likely go down.
She noted that many vehicles currently on the road are equipped with only one or two cameras, adding that hardware readiness remains a factor. She said these constraints are expected to change over time and added that Chinese automakers do not face the same limitations.
Duan also said broader access to autonomy platforms could have implications for ride-hailing-style networks such as Uber and Lyft.
Tesla’s Chip Roadmap
Duan’s remarks come alongside Tesla’s ongoing work on custom AI hardware. Tesla’s vehicles currently use the AI4 processor, with AI5 close to tape-out and AI6 already in development, as part of a plan to introduce new chip designs annually.
Tesla CEO Elon Musk has also said that the company may need to build a “gigantic” semiconductor fabrication plant to meet future demand from its autonomous driving and robotics programs, citing potential output limitations from existing manufacturing partners.
What The Market Is Watching Next
Duan said the real answer to the debate around Nvidia’s approach will come from on-road performance rather than model papers or theoretical discussions. She said the first meaningful signal will be the Mercedes-Benz CLA deployments expected in 2026.
Until then, Duan said most conclusions remain educated guesswork, and added that self-driving remains frontier technology, with no clear consensus on long-term market penetration or sustainable market share for any single player, and said new information that materially shifts expectations is likely to amplify volatility.
How Did Stocktwits Users React?
On Stocktwits, retail sentiment on Tesla was ‘extremely bearish’, while sentiment on Nvidia was ‘bullish’, with message volume described as ‘high’ for both stocks.
While Tesla’s stock has risen 5% over the past 12 months, Nvidia’s stock has jumped 30% over the same period.
For updates and corrections, email newsroom[at]stocktwits[dot]com.<
