Can Tesla stock really reach 1,500 dollars by late 2027? The number is floated on Randy Kirk's show, and guest Larry Goldberg offers only this: it is not impossible. Goldberg, a serial entrepreneur turned venture investor, builds the case piece by piece: if unsupervised driving, robotaxi economics, Semi and Optimus all mature in the same year, a repricing becomes unavoidable. This article stress-tests that claim from start to finish.
Goldberg's first distinction matters: 2026 is the year the strategy completes , not necessarily the year the stock flies. Tesla, he argues, has spent two years paying for its transformation from a car company into an embodied-AI company, with heavy spending, choppy earnings and shaken confidence. The slide toward 130 dollars already priced all of that pain, and the recovery is under way. So a 2027 repricing would simply be the receipt for a transformation finished in 2026.
The Model 2 dinner and the silent pivot
At the center sits a dinner that was never announced. At a senior executive's home, Elon Musk is pressed to commit to a small cheap car and reluctantly agrees at the table. Goldberg's point: Musk never changed strategy there; the Model 2 was shelved and the company was quietly redrawn. The dinner scene in the Isaacson biography is accurate, but the real decision came after: no capital returns to shareholders, every penny buried into compute and new platforms.
The bill for that choice was visible early. Goldberg recalls sharing a chart of how much compute the company would need to buy; chips could not be built in-house, Nvidia filled the gap, the schedule slipped but the direction held. Now a 30-to-50-billion-dollar capital need looms and the market is not convinced yet. The bull case says the pressure is temporary: once the Cybercab joins the fleet, heavy trucks roll and embodied intelligence becomes visible to everyone, the market will re-rate the stock.
Unsupervised driving and the regulatory bar
The first pillar is unsupervised full self-driving . Goldberg uses the supervised system almost daily and says he has not intervened for safety in a year and a half, touching the wheel only for personal preferences. His forecast is blunt: the highway safety agency NHTSA will set an unsupervised standard before late 2027, and Tesla will declare the system then. Today's geofence is, in his reading, not a safety need but a cautious administrative containment that comes off during 2027.
Skeptics ask the fair question: a single crash could ignite an insurance and liability firestorm. Goldberg concedes crashes will happen but argues attributing fault will be harder than with human drivers, with Tesla's camera footage favoring the company. The new automatic collision evasion strengthens the case: the system no longer only brakes, it steers away from impact. Scale is asymmetric: TeslAnt's data compilation puts Tesla's unsupervised robotaxi miles just past 1 million while Waymo holds over 200 million, which means regulatory pressure will soon come from more than Tesla alone.
Robotaxi: from the Austin launch to a dollar a mile
The second pillar is already running in the field. According to BusinessInsider's launch account, Tesla opened the steering-wheel-free, pedal-free Cybercab to the public in Austin on 3 September 2026: butterfly doors, two seats, hailed from the Robotaxi app. Texas records showed 45 Cybercabs initially while more than 150 vehicles were counted at the factory lot. TeslAnt's fleet registry logs the 67th Cybercab on the street, with the Model Y fleet at 150-160 against roughly 300 Waymos. On manufacturing, the company leans on its Unboxed process, claimed to halve the line.
The real break is the price point. Goldberg expects under a dollar a mile in the Cybercab next year, with the 70-cent zone making families question the second car and commuters abandon vehicles that sit parked all day. Roughly 70 percent of all trips carry a single occupant, so the two-seat math works. The fight with Uber will be decided city by city: enough cars, enough availability and a disorienting price, and riders never go back. He notes Tesla already roams under two dollars a mile with a handful of cars, so the target is not far.
That is where mission collides with Wall Street. Cheap fares grow the market but defer profits, while the market wants earnings growth now. Goldberg's answer is blunt: Tesla never played to convince investors; it plays to finish the mission as fast as possible. Against the price-earnings ratio near 300 , he offers two historical notes: as a pure car company the stock still printed 450-500 dollars, and it once ran about 20x in eighteen months. If the stories start landing in 2027, confidence expansion and multiple expansion arrive together.
Semi, energy and the compute ceiling
The third pillar is trucks and energy. Electrek reported on 25 September that Tesla started Semi high-volume production at its dedicated Sparks, Nevada plant: a 1.7-million-square-foot site designed for 50,000 trucks a year, first vehicle off the line on 29 April, formal inauguration on 24 September. TeslAnt's factory file has analysts expecting 5,000-15,000 deliveries in 2026, so nameplate capacity must not be confused with first-year output. A 500-mile variant, Megacharger support and the adjacent 4680 cell line resolve the supply knot that held the program back for years. Goldberg's accounting thesis is simple: every business line that hums covers the overhead inflated for robotaxi and robots, and 50,000 Semis mean serious operating leverage . The fourth front is invisible yet decisive: compute . Goldberg ties the delay of full self-driving version 15 directly to capacity; the new model needs ten times the parameters and everything on hand is consumed by Optimus and driving. That also answers why Tesla opened no AI-rental revenue stream: there is no spare capacity. The key part is the AI5 chip , a game changer the day it enters the data center. On the xAI comparison he objects structurally: building general-purpose compute halls is a different business from feeding your own fleet, and Tesla is simultaneously raising three factories in Austin plus a giant Texas solar plant. Autonomous trucks sit 2-3 years out, an announced factory project is a 2028-2029 story, and Terafab is a 2030 story.
The last pillar is Optimus , where expectations are deliberately kept low: no mass customer deliveries in 2027. What comes instead is proof-of-concept trials, 5,000-10,000 robots across five to ten customers, plus case studies that go well. The first 20,000 units go to internal training, factories and SpaceX, with Berlin workers reportedly wearing teleoperation suits to train the robots. Yahoo reported from Davos that Musk points public sales to late 2027, while Teslarati reported Optimus 3 in its final stage, high-volume output targeted for 2027 and a second Texas facility on the way. Plant managers will multi-source rather than bet on one vendor, and true scale is measured in tens of thousands. The warning is explicit: scaling embodied intelligence is nothing like scaling language models.
Key moments
AI commentary
"As I see it, this conversation is valuable not for the price target but for its testable claims. An unsupervised-driving standard, the end of the geofence, customer pilots: all are calendar milestones anyone can track. The 1,500 dollars reads best as a checklist, not a forecast."
AI assessment
The strongest objection is the price itself: with a P/E near 300 , part of the robotaxi expectation is already written in, and the company's calendar record is poor; Semi ran seven years late and self-driving dates kept slipping. What was promised for 2026 being retold for 2027 gives listeners a fair sense of deja vu. Bulls deflect this with the strategy-versus-price distinction, but that distinction only convinces if miles and deliveries do the talking.
Gaps remain in the file: the quadrupling-subscriptions claim, audited cost per mile, real Semi delivery counts and how far energy profits cover robotaxi spending are not yet on paper with numbers. As the senators' letter to NHTSA shows, the regulator's crash-versus-distance data framework is weak too, making independent verification of safety claims hard. On Optimus, pilot announcements must not be confused with revenue; years stand between proof of concept and serial sales.
The speaker's incentives deserve a note: a long-positioned venture investor naturally carries the bull narrative, and the show's audience is already converted. The practical takeaway for readers is the calendar: geofence removal by mid-2027, a modelable robotaxi P&L, customer case studies and the AI5 ramp. The 1,500 dollars is not a prophecy; it is a scenario that holds only if all four conditions land in the same year.
Sources
6 links; no other published story cites them. Stories sharing a link do not confirm each other; a source's origin is not inferred from how often it is cited.
tesla · robotaxi · fsd · optimus · tesla semi · cybercab · tsla