🚨 Tesla says it more than doubled its onsite AI training compute in Texas during the first half of 2026.
Cortex 2 supports autonomy development for both Tesla vehicles and Optimus, with more capacity planned through the rest of the year.
Tesla is also expanding battery production across several sites:
• Vehicle battery packs in Berlin
• Cathode materials and lithium refining in Texas
• LFP cells in Nevada
• More 4680 cells for Cybercab, Semi, and Model Y
AI compute and battery supply remain two of Tesla’s biggest priorities as it prepares for its next phase of growth.
$TSLA
Tesla has the industrial capacity to produce 125,000 Cybercabs per year.
Just do the math: with a $30,000 net margin per Cybercab per year, that’s $3.75 billion in annual net profit — from just one factory line.
And that’s assuming they only hit 125k units. If they eventually scale to a fleet of 1M Cybercabs (which I believe is realistic long-term), you’re suddenly looking at $30 billion+ in yearly net profit from the robotaxi business alone.
That’s the kind of cash flow that can fund massive AI and robotics acceleration while still delivering huge margins.
Tesla just broke an all-time record for Operating Cash Flows:
✅ $18.685B over the Trailing 12 Months
Spending on revenue-generating assets with long useful lives also broke records:
✅ $5.789B in Q2 2026 alone
✅ $12.923 over the Trailing 12 Months
$TSLA @elonmusk
THERE ARE TOO MANY CRYBABIES HOLDING $TSLA STOCK, HERE IS WHAT YOU SHOULD KNOW:
CYBERCAB IS IN PRODUCTION.
SEMI IS IN PRODUCTION.
ROBOTAXI IS SCALING.
LITHIUM REFINERY IS IN PRODUCTION.
FSD SUBSCRIBERS ARE GROWING.
MANUFACTURING CAPACITY IS GROWING.
SOLAR IS COMING AND BEING BUILT.
SUPERCHARGING IS GROWING.
OPTIMUS GIGA FACTORY IS BEING ACTIVELY BUILT.
BATTERY AND CELL PRODUCTION IS RAMPING.
VEHICLE PRODUCTION IS RAMPING.
THE AI6 CHIP MADE BY TESLA IS GOING TO BE THE BEST AI CHIP EVER MADE.
MEGAPACK 3 IS COMING.
LOCALIZED SUPERCHARGING DATA CENTERS IS COMING.
CAPEX SPEND ON ALL AREAS OF THE BUSINESS IS GROWING.
THE KEY HERE IS THAT THEY ARE INVESTING AS MUCH AS POSSIBLE TO GROW AS FAST AS POSSIBLE.
SURE THEY ARENT GOING AT THE PACE WE ALL WANT BUT THEY ARE QUITE LITERALLY CREATING THE FUTURE THAT CURRENTLY DOES NOT EXIST.
$TSLA Q2 2026 Earnings Call Notes (Gali Style) 🤓📝
Elon Opening remarks: Great quarter, record Q2 deliveries, Model Y best-selling car of any kind in the world (I think), In locations with FSD approval we’re seeing a very high take rate. People buying cars just for FSD
FSD is a significant demand driver. As approval ramps to other countries, I think we’ll see a similar acceleration in demand
Energy business growing “incredibly fast”, will be crucial for artificial intelligence/data centers
Investing a lot, massive CAPEX year, preparing for the future. Things were investing in will yield incredible return. Best CAPEX returns we’ve ever seen
Cybercab has started production. Optimus starting production very soon. Tesla Semi has started production. Megapack production starting soon. Ramping battery cell production. Also preparing to do a massive solar panel/cell production, all the way from Silicon refinement to deploying the cell
We’re working on the most ambitious advanced technology buildout in history
Robtoaxi goals are very ambitious. But we are being super cautious about causing any accidents or harm to anyone. Although there are 40,000 automotive deaths each year in the US, they generate almost no press. But if Tesla injures one person it’s worldwide headlines news, regulators could clamp down actives. We’re going as fast as possible to scale Robotaxi, while ensuring we don’t harm anyone, or even run over a pet.
We’re providing live updates on X (Robotaxi account) as we scale to new cities. Expanding in Florida and Texas. Will continue to scale very rapidly. More than 10% per week in terms of miles driven
Optimus. Will be biggest product ever. But very complex to solve. One of the hardest things to do is make an autonomous humanoid robot that you simply ask it do to a task. It can watch a video and do it, without any programming. Many challenges with dexterity, and surviving lots of wear and tear. Hardest production ramp we’ve ever done at Tesla. Everything is new, and needs new supply chains. Optimus will follow normal S curve of manufacturing, but initial portion of S curve will be very flat and long. You’ve seen lot of robot demos online, but they are preprogrammed or controlled. Optimus will be the first robot to do these tasks by itself. It will have human hand level dexterity. Human hand is an incredible thing. Optimus will have human, then superhuman level dexterity.
Terafab. Expect to announce a location soon. And will provide more details then. Will leave that info for the launch announcement. It will be big, and deserves its own spotlight, not squeezed into an earnings call. Have to solve memory logic and packaging to scale. Have placed equipment orders for development fab in Austin. Dev fab is pretty cool. Will have a very fast iterative cycle to try out new chip designs. Building will be first of its kind on earth. Will try high risk and high payoff bets on AI chips.
Digital Optimus. Basically driving a computer screen like you drive a car. For the car its pixels in, controls out. Digital Optimus will be the same. Will be a self driving computer essentially. Digital Optimus is important for physical Optimus. Because physical Optimus will need to know how to work computers. This is in partnership with SpaceX’s Grok. Grok will tell Optimus what to do, provides it a series of tasks, then optimus does them.
Megapod. Building out design with AI computers. Digital Optimus in a box. We have 7GW of power at superchargers and growing. We can place Megapods at many of these and have distributed power for AI. Could be quite a big deal.
So much happening, hard to squeeze into an earnings call. It’s going to be one of the best years ever for Tesla. Next year will be even better.
CFO Remarks:
Q2 continued Q1 trends of resurgence in demand. Exited Q2 with largest automotive backlog since 2023. Production growth limited by supply chain, electrical components.
Sales data is showing, FSD is one of the main reasons people are buying cars. 55% of deliveries in NA had FSD subscriptions at the time of vehicle delivery.
Nearly 1.5M paid subscriptions (about 45% subscriptions, 55% lifetime purchases).
Automotive margins (excluding credits) declined. Came from warranty changes and tariff relief. Excluding those, margins would be almost flat.
Energy business is lumpy. Tied to customer timelines. It was the 2nd largest quarter for energy. Energy gross margins declined from 29% to 20%. Because of warranty true up ($240M), also tariffs benefits from Q1 did not repeat. ASPs for industrial storage going down because of competition. Long term the energy business will stay at 20-25% gross margins.
Service margins hit 14.1%, all time high. Operating expenses ramped, because of R&D with Semi, Optimus, Cybercab and other AI initiatives. Also costs related bringing AI compute online. We are in a big investment cycle. R&D will continue to grow in 2026 and beyond.
Net income was positively impacted by $1B in SpaceX equity gains.
Free cash flow was negative for the quarter, as guided. Because of CAPEX doubling sequentially. It will increase further in the 2nd half of 2026. CAPEX for this year will be above $25B. It will grow for the next few years as we keep investing in Robotaxi, Optimus, Terafab, Solar, etc …
SAY Q&A
Ashok answering …
Robotaxi questions were many, so providing a comprehensive view.
Robotaxi going very well. Impeccable safety record. Over 380,000 miles unsupervised. 0 notable incidents. Want to emphasize 0 incidents so far. Everyone said we needed LIDAR, but we are showing that is not true. This record is a huge validation of Tesla’s AI push.
Scaling. We started just one year ago in Austin. Had safety monitors back then. We’ve grown at compounding rate week over week with unsupervised miles.
Robotaxi fleet is already running versions of V15 FSD software. We planned 7 major improvement tracks in FSDv15, we’re making progress on all of them. Car is riduclously safe and capable. All tech improvements are working, and expect the payoff to be huge. And we will sustain the same growth rate through the rest of this year and beyond.
Cybercab. It’s phenomal product. Anyone who rides in it falls in love with the experience. Exponential growth will continue for foreseeable future. Cybercab will be awesome.
Data for Optimus training. (I got distracted, Elon replied to one of my tweets while listening lol)
AI strategy for Optimus is the same thing as FSD. Pixels in, controls out. Optimus will be easy. Can ask it to do anything, and it will just do it. V15 software works on Optimus as well.
ANALYST QUESTIONS (worried they might be boring, uh oh)
Big pause, still on mute … lol classic.
Are suppliers helping to ramp Optimus? Elon answering: Suppliers have been great. Making big investments in support of Optimus and Robotaxi. Samsung and TSMC in particular are building Fabs. Putting in $10Bs to build AI compute for Optimus and Robotaxi. Panasonic investing Billions in increasing battery cell production. Also like to thank Micron for memory allocation. They had to make tough decisions there, and gave Tesla significant allocation at good terms.
Lars coming in… for Robotaxi regulation. Getting support from NHTSA. State to state regulators not ideal, especially New Jersey. Tesla going to let the performance speak for itself (safety data is my guess).
Elon answering … Robotaxi will be vertically integrated, like other business. Won’t have demand challenges because economics will be so good. Demand will be bigger than supply. Really just going through the march of 9s of reliability. It’s the only thing constraining growth in Robotaxi.
Analyst asking about Tesla / SpaceX combining … (uh oh) lol.
Elon answering … Many collabs between Tesla and Space. Especially Terafab. Going to be a gigantic project. Can talk about combining companies on earnings calls, needs an appropriate process.
Other guy chimes in .. .We are benefiting from our partnership with SpaceX. Deepend our partnership with an investment in SpaceX. Grok in the car. Grok helping with digital Optimus. Starlink integrated into Cyebrcab. Soon all of vehicles. For Robotaxis you need cell coverage everywhere. Starlink will allow people to watch 4K live sports in the car.
Robotaxi question … only dozens of cars in each market, why not scale Austin instead of other cities?
Ashok answering … Want to prove the stack works in a lot of different cities without much effort. Growth rate has been exponential … early part of exponential. Vehicles are in continuous operation, so even with a few vehicles you can get a lot of miles out of them. Want to operate a very efficient fleet, optimizing for lots of miles per vehicle in the fleet. Before go high in terms of deployment have to get the infrastructure built out.
Elon chimes in… Cybercab has a new vehicle chassee. This means we need to put lots of testing into it to validate.
Lars chiming in … Regulations different city by city and state by state. Expanding city by city is to meet those requirements one at a time.
Analyst question.. Anything at federal level needed to unlock Cybercab? Lars says no, not really.
Elon .. Very serious shortage of truckers in USA right now. Autonomous Semi is very important to address that. Focusing self driving efforts on high volume vehicles for now. Taking back seat for 6 months for Cybercab. But working next year and the scale up to high production of the Tesla Semi.
You could use starlink terminals in Tesla cars as cell towers / connectivity towers for providing relays on the ground to cells phones or anyone who needs Wifi.
Optimus has lots of specialized electronics , and circuits boards, custom designed for Tesla. For Optimus V4 built in Austin, will be much more vertically integrated. Going for 10M units year, vs 1M units year for Fremont. Will be doing PCB work in house for the V4 Optimus.
AI5 Chip will initially go into Optimus. Tesla chip team is doing great work, making great progress on AI5. Very excited about Tesla AI6 chip. Will be best AI computing chip in the world.
Question.. Pace and path of CAPEX.
Elon … We should be spending CAPEX as fast as we can, without it being “too wasteful”. Not aiming for some extreme high efficiency capital spend because that slows things down. It’s a balance between capital efficiency versus time. Overall pretty happy with how things are going. So much construction going on in so many regions. No company is investing like we are … maybe Henry Ford with the Model T .. but probably fastest industrial scale up since WW2 for America.
CFO .. all of spend on CAPEX is on productive assets. We’re all in on scaling factories.
Analyst … question on Energy storage. Is it peak shaving demand? Is it a quality of power for datacenters??
Elon answers … Not just peak shaving, but grid balancing. Balancing power from wind generation and solar generation, our batteries are awesome for this. Solar + batteries is how most energy will be produced in the future. That’s how Starlink works. Earth is like a giant satellite. Megapack will be useful for getting more energy from the grid for AI. Best way to scale energy in the US is batteries. We have 1.2-3 Terawatt capacity in us, but only using 0.5. We could double the energy of the US with just batteries. We see demand for Megapack being very very high in the future.
Call ends .. techno music resumes