Investor's Corner
Why Tesla Autopilot will ultimately prove the self-driving industry leader
Tesla took an early lead in the race to develop vehicle autonomy, and its Autopilot system remains the state of the art. However, the technology is advancing more slowly than the company predicted – Elon Musk promised a coast-to-coast driverless demo run for 2018, and we’re still waiting. Meanwhile, competitors are hard at work on their own autonomy tech – GM’s Super Cruise, is now available on the CT6 luxury sedan.
Is Tesla in danger of falling behind in the self-driving race? Trent Eady, writing in Medium, takes a detailed look at the company’s Autopilot technology, and argues that the California automaker will continue to set the pace.
Every Tesla vehicle produced since October 2016 is equipped with a hardware suite designed for Full Self-Driving, including cameras, radar, ultrasonic sensors and an upgradable onboard computer. Around 150,000 of these “Hardware 2” Teslas are currently on the road, and could theoretically be upgraded to self-driving vehicles via an over-the-air software update.
Above: In its current state, Tesla’s Autopilot requires a hands-on approach (Youtube: Tesla)
Tesla disagrees with most of the other players in the self-driving game on the subject of Lidar, a technology that calculates distances using pulses of infrared laser light. Waymo, Uber and others seem to regard lidar as a necessary component of any self-driving system. However, Tesla’s Hardware 2 sensor suite doesn’t include it, instead relying on radar and optical cameras.
Lidar’s strength is its high spatial precision – it can measure distances much more precisely than current camera technology can (Eady believes that better software could enable cameras to close the gap). Lidar’s weakness is that it functions poorly in bad weather. Heavy rain, snow or fog causes lidar’s laser pulses to refract and scatter. Radar works much better in challenging weather conditions.
-
According to Eady, the reason that Tesla eschews lidar may be the cost: “Autonomy-grade lidar is prohibitively expensive, so it’s not possible for Tesla to include it in its production cars. As far as I’m aware, no affordable autonomy-grade lidar product has yet been announced. It looks like that is still years away.”
If Elon Musk and his autonomy team are convinced that lidar isn’t necessary, why does everyone else seem so sure that it is? “Lidar has accrued an aura of magic in the popular imagination,” opines Mr. Eady. “It is easier to swallow the new and hard-to-believe idea of self-driving cars if you tell the story that they are largely enabled by a cool, futuristic laser technology…It is harder to swallow the idea that if you plug some regular ol’ cameras into a bunch of deep neural networks, somehow that makes a car capable of driving itself through complicated city streets.”
Those deep neural networks are the real reason that Eady believes Tesla will stay ahead of its competitors in the autonomy field. The flood of data that Tesla is gathering through the sensors of the 150,000 or so existing Hardware 2 vehicles “offers a scale of real-world testing and training that is new in the history of computer science.”
Competitor Waymo has a computer simulation that contains 25,000 virtual cars, and generates data from 8 million miles of simulated driving per day. Tesla’s real-world data is of course vastly more valuable than any simulation data could ever be, and the company uses it to feed deep neural networks, allowing it to continuously improve Autopilot’s capabilities.
A deep neural network is a type of computing system that’s loosely based on the way the human brain is organized (sounds like the kind of AI that Elon Musk is worried about, but we’ll have to trust that Tesla has this under control). Deep neural networks are good at modeling complex non-linear relationships. The more data that’s available to train the network, the better its performance will be.
“Deep neural networks started to gain popularity in 2012, after a deep neural network won the ImageNet Challenge, a computer vision contest focused on image classification,” Eady explains. “For the first time in 2015, a deep neural network slightly outperformed the human benchmark for the ImageNet Challenge…The fact that computers can outperform humans on even some visual tasks is exciting for anyone who wants computers to do things better than humans can. Things like driving.”
By the way, who was the human benchmark who was bested by a machine in the ImageNet Challenge? Andrej Karpathy, who is now Director of AI at Tesla.
-
===
Note: Article originally published on evannex.com by Charles Morris; Source: Medium
Elon Musk
SpaceX’s next trillion dollar bet has nothing to do with rockets, Musk tells staff
Elon Musk told SpaceX staff AI revenue will soon dwarf rockets and Starlink combined entirely.
Elon Musk told SpaceX employees this week that artificial intelligence, not rockets, will soon carry the company’s revenue. In a roughly 29 minute internal address posted on SpaceX’s X account on Tuesday, Musk said AI revenue will pass every other line of business at SpaceX “probably in September” and pull further ahead by the fourth quarter.
The numbers he gave are specific. SpaceX currently runs 1.4 gigawatts of AI compute capacity. Musk wants that at 10 gigawatts by the end of 2027, a jump he tied directly to revenue: “if we bring 10GW of AI online by the end of next year, it will be $300 billion to $500 billion a year in revenue.” He called those “big numbers,” which undersells a projection larger than what most countries produce in a year.
We made rockets reusable and are rebuilding the internet in space. The next challenge: making life multiplanetary and understanding the true nature of the universe
Watch @ElonMusk deliver a company update to @SpaceX employees pic.twitter.com/5c8rxoCQfu
— SpaceX (@SpaceX) August 11, 2026
Musk went further on where AI fits into SpaceX’s future. “Probably in four or five years, AI will be 99% of the value of SpaceX,” he told staff, adding that digital intelligence would eventually run “a trillion times” ahead of biological intelligence as computing scales. He tied that growth to the company’s founding mission, telling employees “we must win on AI, because the future is overwhelmingly AI and robots,” with the payoff meant to help fund Starship and a Mars program that increasingly runs through Terafab, the joint Tesla, SpaceX and xAI chip plant.
Elon Musk launches TERAFAB: The $25B Tesla-SpaceXAI chip factory that will rewire the AI industry
-
None of this is entirely new territory. SpaceX told investors much the same story during its first earnings call as a public company on August 4, where Musk moved the company’s $1 trillion revenue target up a year to 2030 and said Starlink could someday carry a majority of the world’s internet. What the all hands video adds is a hard deadline and a specific power figure Musk had not given publicly before, along with a franker pitch to his own workforce that AI, not launch cadence, is now the thing SpaceX is betting its future on.
The AI revenue itself is not coming from SpaceX training its own models. It is largely Starlink acting as the network layer for xAI’s workloads, plus SpaceX renting out compute capacity directly, the same approach behind the roughly $16 billion the company spent on AI infrastructure in a single quarter.
Musk closed the video with a pitch aimed at recruiting and retention rather than investors, telling employees that anyone who helps SpaceX win the AI race will eventually get the chance to go to the moon or Mars themselves. Whether SpaceX can turn 1.4 gigawatts into 10 in seventeen months is the more immediate question, and one that will show up in quarterly numbers well before anyone leaves Earth.
In a new note to investors on Tuesday, Morgan Stanley analyst Andrew Percoco said that Tesla has one big financial question to answer for investors regarding its Robotaxi rollout, Full Self-Driving software, and Optimus.
Percoco said in the note that, for the most part, investors are still very positive about the direction the company is headed. However, there are some things the firm would like to see, and they have to do with financials.
Tesla (TSLA) Q2 2026 earnings results: miss on EPS, beat on revenue
Tesla bulls are more than convinced that the company’s Full Self-Driving software is proof it can develop physical AI. Financially, however, there are still some questions, especially on elevated spending, which CEO Elon Musk said would occur as the company works to roll out Robotaxi faster and continue developing its Optimus robot.
The latter two are where Tesla will have to prove progress to investors, as Percoco writes that both projects “will require clearer evidence that Robotaxi is scaling and more tangible Optimus proof points to support the ROI on elevated capex.”
Percoco said the second quarter earnings call did not change his long-term thesis of where Tesla is positioned in the AI race, which is out in front. However, there are concerns that weaker gross margins and higher R&D spend will stress financials, and that has “sharpened our (and investors’) focus on measurable progress across Robotaxi and Optimus.”
Additionally, Robotaxi still needs to be proven with more operation in existing cities while maintaining safety but improving how many rides it gives in any given time, he said. For Optimus, Percoco wrote that he is “still looking for evidence beyond commentary around SOP.”
-
Morgan Stanley put Percoco in charge of covering Tesla after long-time analyst Adam Jonas transitioned to the automotive side.
Currently, Morgan Stanley has a $415 price target on Tesla and a ‘Hold’ rating on the stock. It is trading at around $330 at the time of publication, which was 2:30 P.M. on the East Coast.
SpaceX’s massive investment in AI will make it a big winner, Argus Research said after the company’s successful earnings call last week.
The firm also upgraded shares to a Buy from Hold and set a $160 price target.
SpaceX (NASDAQ: SPCX) is currently recovering from its heavy AI infrastructure investments, as it spent nearly $16 billion in Q2 alone. The company did this primarily by monetizing high-demand GPU compute capacity at a much faster pace than traditional data center economics would suggest.
Company CFO Bret Johnsen said that SpaceX would be able to pay back anything on new deployments within a year.
There are plenty of ways the company can do this:
Leasing excess compute capacity through contracts
SpaceX has already built Colossus and Colossus II, largely for its own model training. However, much of that capacity is already rented out to third parties. It already has major deals with Anthropic, Google, and Reflection AI. These partnerships are adding billions per month to SpaceX’s spreadsheet.
-
High utilization driven by industry-wide scarcity
The demand for advanced AI training and inference capacity continues to exceed what is available for use. SpaceX can fill new racks quickly after they come online, so the capital deployed converts into revenue with minimal idle time.
Additionally, management and outside observers have described the new compute capital as behaving more like a cost-of-goods-sold than traditional multi-year capex, especially because of this rapid monetization pattern.
Capacity has already scaled from ~0.4 GW a year to 1.4 GW annually by the end of Q2. There are targets of more than 2 GW by year-end.
High incremental margins on the rental business once capacity is online
GPU cloud providers often operate at strong gross margins. SpaceX can monetize capacity that was already partially built or can be added efficiently. This means that incremental EBITDA margins on the rental revenue are usually high. This accelerates cash recovery relative to the gross capital outlay.
Parallel monetization of its own AI software and applications
Beyond pure infrastructure rental, SpaceX also generates revenue from Grok through subscriptions and usage, from X through ads, data, and other related services, enterprise APIs, and the planned integration of the Cursor coding tools acquisition.
These application layers ride on the same compute infrastructure and provide additional high-margin streams that could offset build-out costs. AI-segment revenue overall rose sharply to about $2.6 billion in Q2, according to Motley Fool. This was driven primarily by the infrastructure contracts, but the software side is also partially responsible.
Efficient, large-scale deployment and vertical integration advantages
SpaceX has emphasized the rapid construction of power and cooling infrastructure and favorable cost-per-megawatt economics relative to industry benchmarks in some disclosures.
-
Combined with its ability to scale capacity aggressively and the fact that many contracts start generating revenue within months of capacity coming online, the effective payback compresses dramatically compared with more conventional multi-year data-center projects.
SpaceX’s dominant near-term recovery path will turn the AI clusters into a hyperscale-style compute rental business for other leading AI companies while still using a portion for internal models.
SpaceX’s next trillion dollar bet has nothing to do with rockets, Musk tells staff
Tesla has one big financial question to answer for investors: Morgan Stanley
