Elon Musk’s pushed back hard against critics who’ve questioned Tesla’s massive spending on artificial intelligence. Short-seller Jim Chanos had argued that Tesla’s Full Self-Driving system hasn’t progressed beyond Level 2 autonomy since development started in 2014. Musk countered that training data wasn’t actually the problem holding back FSD progress—at least not until lately.
The criticism pointed out that competitors like Mercedes-Benz achieved Level 3 regulatory approval while Tesla stalled at supervised autonomy. GM Super Cruise and Ford BlueCruise reached Level 2 classification faster than Tesla despite shorter development timelines. Critics questioned why a decade-long project hadn’t moved beyond driver-assist capabilities, especially without regulatory milestones to show progress.
Musk explained that Tesla’s vehicle fleet already provided enough real-world data for current autonomy levels. He said the company’s framework was designed so that artificial intelligence processing happened directly inside vehicles rather than in expensive data centers. This approach meant Tesla didn’t need massive spending on centralized computer infrastructure. Customer vehicles collected data while driving, which reduced reliance on generated data for training. Baker emphasized that reduced capital expenditure does not imply lack of seriousness in AI development. The distributed inference model allows customers to effectively pay for inference hardware through their vehicle purchases.
However, Musk acknowledged the situation has changed for Tesla’s robotics programs. The company’s new Cortex 2 infrastructure, revealed during responses to short-seller criticism, specifically targets training for Optimus humanoid robots. This project represents a strategic shift away from FSD focus toward robot development, with AI-related investments expected to push broader market capital expenditure significantly higher. Musk projects that Optimus robots could represent 80 percent of Tesla’s future value. Current FSD users can access the system through monthly subscription options for ongoing access.
Meanwhile, Tesla’s R&D costs jumped 57 percent year-over-year while operating margins declined. The company received shareholder approval for a $1 trillion compensation package despite these financial pressures. Musk announced plans for unsupervised FSD delivery “a few months away” heading into 2026, alongside “text and drive” capabilities expected within one to two months.
This pattern reflects Tesla’s broader strategic evolution. Integration with Musk’s artificial intelligence company xAI creates what he describes as recursive benefits between data, software, and hardware. While critics remain skeptical about FSD timelines and AI spending priorities, Tesla continues pushing toward its robotics vision alongside autonomous driving development.