Tesla began taking Cybercab orders by invitation on November 13, 2026, at $29,990 with a required $299/month Autonomy Subscription, December 2026 estimated delivery in Austin, and operation limited to Tesla's existing Robotaxi service areas in Texas and Florida.
MKBHD's bet terms were that Tesla would deliver a Cybercab to a regular customer for $30,000 or less before the end of 2026. As of November 13, 2026, orders are invite-only with no public order page, and no customer deliveries have happened yet, so his hair is safe for now. This is a personal account of building an educational robotics project with Fable 5.2; no factual claim to verify. General commentary about the humanoid robotics race; no specific factual claim to check.
JPMorgan reportedly forecast US humanoid robot production of 300,000 to 400,000 units by 2030 and about 1.5 million by 2035, according to Reuters. A link share to a guide; no factual claim to verify.
The FAA's nationwide ground stop ended Friday evening after about five hours, though travel disruptions were expected to continue through the weekend. A deals roundup; no factual claim to check.
The FAA grounded all US flights on Friday, citing a nationwide communications failure among air traffic control systems. A Black Friday deals article; no factual claim to check.
MMarques Brownlee just posted the Cybercab order page is live, and the fine print is doing a lot of work.
CInvite only, three hundred dollars a month, and it only drives in a few Texas cities.
MSo it's a thirty thousand dollar car you can't actually buy, that needs a subscription to move.
CAnd he's got a bet with the internet that his hair survives if Tesla misses the deadline.
MHis hair is safe. No deliveries yet, and invite-only is not 'publicly available.'
CBut December delivery estimates? That's sooner than I expected.
MSure, for the people Tesla picks. This is a beta test with a price tag.
CEvery new car starts somewhere. The Model 3 was a mess at launch too.
MThe Model 3 didn't need a monthly fee to drive itself.
CNo, but it also didn't drive itself at all. This is a different animal.
MA two-seat animal with no steering wheel, stuck in a geofence.
CYou're acting like geofencing is a failure. It's how you do this safely.
MI'm not saying it's unsafe. I'm saying it's not the revolution we were promised.
CIt's the first step. You don't go from zero to everywhere overnight.
MBut Tesla's been promising this for years. At some point, 'first step' becomes an excuse.
CSo what's the alternative? Wait until it's perfect and never ship?
MNo, but don't sell it as a product when it's still an experiment.
CThey're not selling it to everyone. It's invite only. That's the point.
MThen don't put a price on it. Don't call it an order page.
CYou have to start the process somewhere. People want to know what it costs.
MFine. But Brownlee's bet was for Tesla to deliver a Cybercab to a regular customer for thirty thousand dollars or less before the end of 2026. This is cutting it close.
CDecember is still 2026. But as of November 13, 2026, orders are invite-only and no customer deliveries have happened yet, so he hasn't lost the bet.
MHe said 'publicly available.' Invite only doesn't count.
CWe'll see. I wouldn't bet against Tesla on deadlines. They miss them, but they do ship.
MSpeaking of shipping, JPMorgan just put out a wild number on humanoid robots.
CHundreds of thousands of humanoid robots in the US by 2030.
MThat's a lot of robots. And even more by 2035.
CReuters has the report. The industry's moving from demos to real jobs.
MI'll believe it when I see one doing something useful outside a warehouse.
CThey're already in warehouses. Amazon's been testing them for years.
MTesting is not the same as four hundred thousand units.
CNo, but the curve is steep. Once they work, they scale fast.
MOr they hit a wall. Remember when self-driving was two years away? That was a decade ago.
CThat's different. Robots don't have to deal with traffic and pedestrians.
MThey have to deal with stairs, doors, and not falling over.
CBoston Dynamics has been doing backflips for years. The hardware's there.
MAgreed. The software that makes them useful? That's the hard part.
CWhich is exactly why JPMorgan's number is interesting. They're betting on the software catching up.
MOr they're just putting out a big number to get headlines.
CJPMorgan doesn't need headlines. They need to be right for their clients.
MFair. But four hundred thousand robots in four years? That's a factory every week.
CChina's already building them fast. Unitree, UBTECH, a whole ecosystem.
MAnd that's the real story. If the US doesn't move, China owns this market.
CWhich is why Tesla, Figure, Apptronik are all pushing hard.
MIt's a race. And the prize is the next industrial revolution.
CMeanwhile, the FAA ground stop is over. Five hours of chaos, and now the cleanup.
MWired says planes are flying again, but the effects will last all weekend.
CA nationwide communications failure in air traffic control. That's not a small glitch.
MNo, and it's the second time this year. Someone needs to answer for this.
CThe FAA's been underfunded for years. This is what that looks like.
MUnderfunded or mismanaged? There's a difference.
CBoth, probably. But you can't run a modern air system on old tech.
MAnd you can't ground every flight in the country without a backup plan.
CThe backup plan was to stop everything. That's the safe call.
MSafe, but expensive. Airlines are going to lose millions.
CCheaper than a mid-air collision.
MTrue. But this keeps happening. That's the part that bothers me.
CIt's a wake-up call. The whole system needs an overhaul.
MAnd it won't happen until someone in Congress feels the pain.
CThey fly commercial too. They felt it today.
MThen maybe something actually changes.