BOOM 🆘 2027 is shaping up to be the brilliant year that Tesla's AI ambitions become real 🔥 On the Q3 earnings call, Elon and the team put dates on almost everything enthusiasts have been waiting for, and the biggest one is about the car already sitting in your garage 👀 🚗 Unsupervised FSD could start reaching customer cars around Q2 2027. Ashok Elluswamy said Cybercab and the Model Y Robotaxis already run the same software, which is coming to customers soon in the V15 release. Around Q2 2027, Tesla expects the first customer cars to join the Robotaxi fleet and drive fully autonomously within Tesla's geofences and fleet management system. It starts inside Tesla's network rather than everywhere, but this is the moment owners have waited years for. The cars they already own are now on a real path to autonomy. 🤖 Tesla expects to start delivering Optimus to outside companies around the end of 2027. Elon said the remaining AI, design, and durability challenges should be solved around the end of 2027, when deliveries to other companies begin. The first jobs will be dangerous, repetitive industrial work that runs around the clock, with home robots coming later. Before then, Elon expects to release an Optimus video in a month or two showing tasks that aren't pre-programmed. Tesla also plans to put its AI5 chip into Optimus around April 2027, with about ten times the compute of AI4. The hand is still the hardest part. Tesla is setting up an "Optimus Academy" where 10,000 to 20,000 bots will learn by practicing tasks, and it's building a second Optimus factory in Austin that can make far more robots than Fremont. CFO Vaibhav Taneja said Optimus costs are trending toward where Tesla expects Cybercab costs to land. Elon sees Optimus's usefulness as dexterity, AI software, and AI chip multiplied together, all improving fast, with robots eventually helping build future robots. He expects humanoid robots to someday outnumber humans. The video is the proof point to watch. Real, unscripted work would answer the biggest doubt about humanoid robots, and a robot that costs about as much to build as a Cybercab could work for a huge number of businesses. 🚕 Robotaxi is about to expand fast, with Dallas up next. Ashok said Cybercab launches in Dallas next month, and Vaibhav said more cities should follow before the end of the year. Elon said the Austin Cybercab fleet has grown about ten times since launch. Service hours extend to 2am soon, with 24/7 operation, highways, and more weather conditions planned by the end of this year. Cybercab's AI was trained mostly on Model 3 and Model Y data, so Tesla didn't need a big Cybercab fleet to get it ready. Vaibhav said other vehicle types will join the Robotaxi fleet over time, with the goal of covering every kind of ground transportation. Lars Moravy said Cybercab has 50% fewer parts than any Tesla before it, and the real line capacity is well above the 125,000 a year in the deck, though Tesla is building only as fast as Robotaxi needs. The first Cybercabs should reach customers by the end of 2026, starting small, but Elon said most will stay Tesla-owned and be offered as transport as a service. Elon expects Robotaxi revenue to become noticeable around Q2 2027 and a big deal by Q3 or Q4. Rides are already priced below rideshare in most markets, which Vaibhav said works because there's no driver and Tesla builds its own cars. Tesla isn't at its 30-cents-per-mile cost target yet but sees a clear path there. Production isn't the bottleneck. Ashok said software sets the pace today, with regulators likely becoming the limit in early 2027. Elon predicts Tesla will eventually handle most passenger miles in any country that lets it operate freely. 🗺️ Tesla is building its own AI-powered maps and navigation. Answering a shareholder question about speed limits, parking, and routing, Ashok said Tesla is replacing third-party maps with its own AI-based mapping and navigation. You'll be able to talk to it and ask for routes, and it will learn your preferences along with what the whole fleet learns. It starts with the beginning and end of trips in V15, then takes over navigation, parking, and destination handling over the next several months. It targets exactly the everyday issues shareholders raised, and every FSD user will feel the difference. 🌍 Elon expects FSD (Supervised) approval across the EU by around Q1 2027. Elon said EU-wide approval could come within a few months, and he'd be surprised if it takes past Q1 2027, though the timeline isn't Tesla's to control. Meanwhile, Tesla is offering benefits on certain cars in Europe to grow its fleet before approval, and it's working with regulators in China, where FSD already runs on some cars. FSD has over 1.7 million paid customers worldwide, with half on subscriptions for the first time. Every new car in North America includes a free month, yet over 75% of buyers still signed up for paid FSD at delivery. Tesla called FSD its biggest demand driver, with showroom visitors now asking for FSD test drives, and it plans to share more FSD metrics, including revenue. An analyst estimated FSD is nearing $1 billion in annual recurring revenue and could at least double each car's lifetime profit, and Vaibhav said his numbers were correct. Vaibhav added that competitor CEOs who took FSD drives in the US were blown away. Tesla has also published a Robotaxi safety report, and Ashok doesn't expect anyone to match unsupervised FSD for several years. Paying on day one with a free month already included is about as strong a vote of confidence as a product can get. 🧠 Tesla's Terafab aims to make its first test chip around mid-2027. The Austin research fab will start commissioning equipment in the first half of 2027, with a first test chip around May or June and a goal of useful chips at volume around Q3 2028. Most of the first equipment is already ordered. Elon expects tens of billions in spending through early 2028 and likely over $100 billion long term, with the cost split between Tesla and SpaceX still being worked out. He called it potentially the most valuable building in the world and key to scaling Optimus by easing bottlenecks in memory, packaging, and logic chips. Owning its chip supply would remove one of Tesla's biggest limits on building robots and cars, and the SpaceX cost split hints at shared infrastructure between the two companies. 🏭 Tesla is aiming for 5 million vehicles a year by around 2030. Elon said demand is exceeding production by a wide margin, and Tesla has a good shot at about 5 million vehicles a year by 2030. That's more than 2.5 times Q3's annual pace of about 1.95 million. Factories hit record rates, Tesla applied for permits to expand Giga Berlin, and a new Semi factory has started in Reno. Q3 also marked the first time Tesla topped 480,000 deliveries two quarters in a row. Based on preliminary data, Model Y was likely Europe's best-selling EV and South Korea's best-selling passenger car. And Elon called the Roadster unveiling arguably the coolest product demo ever. 🔋 Megapacks are almost sold out for 2027. Elon said Megapacks are almost sold out for 2027, production is ramping as fast as possible, and margins should rise meaningfully. Vaibhav said data center demand is extreme and expects better pricing ahead. Q3 energy margin jumped to 33.1% from 20.4% in Q2, mainly from tariff refunds and a warranty adjustment with a cell supplier, and the refund money is going into US production capacity. On the possible "Megapod" product, Elon said it's still early, with more clarity around Q1 or Q2 2027. Selling out a year ahead gives Tesla real pricing power, and data center demand isn't slowing down. 💵 Tesla lowered its 2026 capex guidance and lined up $30 billion in borrowing capacity. Vaibhav said 2026 capex should land around $20 billion, down from roughly $25 billion previously, funded mainly by operating cash. Tesla also secured up to $30 billion in borrowing capacity but doesn't plan to draw on it this year. Opex will keep rising until new products are commercialized, and capex will grow in the coming years with Robotaxi vehicles, Optimus capacity, solar manufacturing, AI compute, and the chip fab. The team declined to discuss a possible SpaceX merger. Spending less than planned while lining up a big credit line tells me Tesla is gearing up for a much bigger investment wave on its own terms. 🚀 Tesla has never looked better positioned. For years, the question about Tesla's AI ambitions was when. This call answered with dates: unsupervised FSD on customer cars, Robotaxi revenue taking off, Optimus heading to customers, and Tesla's first homemade chip. Add sold-out Megapacks, demand outrunning production, and a path to 5 million cars a year, and Tesla's foundation is stronger than ever. As Elon put it, after years of grind, Tesla is now headed for a lot of fun and significant growth.