Tesla’s robotaxi progress shifts investor focus toward AI and compute

Tesla shares moved sharply higher today, rising nearly 5% intraday, and trading up roughly 4% by midday, after CEO Elon Musk disclosed that the company is now testing robotaxis without safety monitors inside the vehicle. The move pushed TSLA to its highest level in nearly a year and brought the stock within reach of prior all-time highs.

While Tesla is still categorized as an electric-vehicle manufacturer, the market behavior increasingly suggests that investors no longer value the company primarily on vehicle deliveries, or near-term automotive margins. At approximately $1.5 trillion in market capitalization, Tesla’s valuation reflects expectations around Full Self-Driving, robotaxis, Optimus humanoid robots, and the proprietary AI compute infrastructure that supports these platforms. Today’s stock move reinforces the view that progress in autonomy and AI execution now outweighs traditional automotive metrics in driving investor sentiment.

This shift did not emerge overnight. Over the past several quarters, delivery estimates that once dominated Tesla earnings discussions have receded in importance, replaced by growing attention to AI capability, training efficiency, and deployment milestones.

Market reaction reflects autonomy progress, not vehicle deliveries
According to the chart,Tesla’s stock responded decisively to news that the company is testing robotaxis without safety monitors, with shares rising nearly 5% intraday and pushing to their highest level in roughly a year. The sharp upward move, accompanied by heavy trading volume, indicates that investors are interpreting this development as a meaningful autonomy milestone rather than a routine update. 

Importantly, the stock reaction reflects growing confidence in Tesla’s AI and Full Self-Driving execution, as shares moved closer to prior all-time highs despite limited emphasis on vehicle delivery metrics.

Tesla launched a limited robotaxi service in Austin, Texas, in June using modified Model Y vehicles equipped with Full Self-Driving software. Early deployments were geo-fenced and included a human safety monitor seated in the front passenger seat. Musk’s statement that testing is now underway with no occupants in the car marks a clear escalation beyond those initial pilots.

This development aligns closely with Tesla’s third-quarter earnings commentary, where management emphasized improvements in Full Self-Driving reliability, neural-network training efficiency, and real-world generalization rather than quarterly delivery growth. Over recent quarters, investors have increasingly discounted short-term fluctuations in vehicle shipments and instead focused on whether Tesla is executing against its longer-term autonomy roadmap.

AI chips as enabler of Tesla’s autonomy strategy
Tesla’s ability to progress toward unsupervised robotaxi testing is inseparable from its vertically integrated AI silicon strategy. Unlike competitors that rely heavily on third-party autonomy hardware, Tesla designs its own processors to handle perception, planning, and inference at scale. This approach allows Tesla to optimize hardware and software together and advance autonomy primarily through training and code rather than sensor proliferation.

In August 2025, I presented a 26-slide analysis to institutional investors at a Merrill Lynch BofA event titled Tesla Chips by Samsung/TSMC Foundry. That presentation examined Tesla’s progression from AI4 through AI6 processors, which power Tesla vehicles as well as Optimus humanoid robots. The analysis detailed how each generation increased inference throughput, reduced latency, and improved energy efficiency, all of which are prerequisites for vision-based autonomy operating without human supervision.

Tesla’s AI5 and AI6 processors represent two distinct but complementary steps in the company’s Full Self-Driving and autonomy roadmap, differentiated primarily by process node, power efficiency, and deployment strategy.

AI5, built on TSMC’s 3nm process, is expected to enter volume production in 2026 with relatively high yields in the 90–92% range and monthly wafer capacity scaling from approximately 15,000 to 18,000 wafers, making it well suited for cost-sensitive, high-volume vehicle deployments.

-- Dr. Robert Castellano, Semiconductor Deep Dive, USA.