Elon Musk Doubters Are Back on Stage: Are They Wrong Again?

The bosses of some of America's biggest companies sang the same chorus this week: Elon Musk is wrong. The head of AT&T dismissed the SpaceX phone plan as unworkable, the finance chief of T-Mobile invoked the laws of physics, and the chief of Delta was reported as refusing to work with Musk. It sounds like expert judgment. Yet the same physics line was aimed at the Tesla Semi eight years ago, and that truck is now rolling off a volume line. This piece sorts the doubts worth taking seriously from the ones headed for history.

The SpaceX plan looks elegant on paper. Internet beamed down by a rooftop Starlink dish would be handed to nearby phones by a small cell next door, turning a neighbour's dish into a miniature tower. The AT&T chief's objection has three legs: matching a macro network through small cells would cost just as much, radiating signals from someone else's home raises permission problems, and nobody with a giant fibre network is doing this already. The permission point is genuinely serious and the first knot SpaceX must untie with regulators. Axios published the full interview and its report carries every line of the cost and permission argument in detail.

AT&T's History: Inventing the Revolution, Missing the Train

AT&T's record invites scepticism about today's confidence. Its laboratories invented cellular communication, yet when the company asked consultants around 1980 how big the mobile market would be by 2000, the answer came back at 900,000 phones; the real figure passed 109 million, off by more than a hundredfold. Managers who looked at brick-sized handsets and dismissed mobile as a niche had to pay 12.6 billion dollars a decade later to buy their way back in. Today the money goes to fibre, proven by the 3-billion-dollar Corning deal announced the same day as the interview. Fibre may be superb technology, but nothing obliges us to assume the institution that missed mobile is placing the right bet now.

The T-Mobile file is stranger still, because the company is also SpaceX's partner. T-Mobile sells Starlink's direct-to-phone link as an add-on service, yet its chief said last year the threat was exaggerated and its finance chief told an investor conference in September that physics forbids it as a substitute. Walls block the signal, even car windscreens stop it, more than a billion home cells would be needed, and everyone in the know is laughing at the plan, he claimed. The company's own figure, putting satellite traffic at a tiny fraction of its total load, proves today's scale is genuinely minuscule. LightReading printed the full conference remarks and its report keeps the femtocell mockery on the record.

The SpaceX Counter-Thesis: A 600-Billion-Dollar Pool

SpaceX management looks at a different number: the carriers' revenue pool. Its president said in early August that the carriers jointly book close to 600 billion dollars a year, that SpaceX could take a sizeable share, and that its service would be better; Musk added the ground network would beat current cellular bandwidth. All three big carriers' shares fell the next day, which is why the soothing statements that followed deserve a second reading. The company's trump card is a second-generation direct-to-phone constellation promising a hundred times more bandwidth, rising from next year. The spectrum picture is now official: about 17 billion dollars in cash and stock plus 2 billion in interest support, so the 20-billion figure in the video is a rounding of the total burden. EchoStar announced the official spectrum terms and its statement lists the seventeen-billion-dollar price plus interest support.

The numbers strengthen SpaceX's hand. Second-quarter revenue reached 7.8 billion dollars, nearly double a year earlier, the loss shrank from 1 billion to 541 million, and growth rode on Starlink plus computing power rented out for artificial intelligence. Starship reached orbit last week carrying 26 satellites for the first time, Starlink counts 12 million subscribers and more than 10,000 satellites. Barron's writer Al Root openly admitted in September he had been wrong, conceding SpaceX shareholders won the February merger bet after all. TechCrunch dissected the SpaceX quarter and its piece traces the growth to Starlink and artificial-intelligence revenue.

Physics Deja Vu: The Semi Lesson

The truck world remembers the 2018 version of the same physics line. When the Tesla Semi launched with a 500-mile promise, the Daimler Trucks chief replied that the same laws of physics apply in Germany and California, offering to buy two trucks for testing if the promise held. Bill Gates wrote in 2020 that battery breakthroughs would still leave electric rigs impractical, and Musk answered that he had no clue. Late September saw the first dedicated Semi factory open in Sparks: 50,000 units of yearly capacity, 500 miles fully loaded, in-house 4680 cells, and launch customers including PepsiCo and DHL. Electrek compiled the factory capacity and range figures and its report details the switch to in-house 4680 cells.

Still, the mockers earned one point: the calendar. The Semi entered volume output in 2026 instead of 2019, seven years late. A rival who lost no sleep over the truck in 2018 slept soundly for years, and cheaply. The physics objection lost while the schedule objection won, and that split is the finest lens for reading today's phone debate. The question is speed, not possibility.

The Delta file smells more personal than technical. The airline picked Amazon Leo over Starlink for onboard internet; Starlink reportedly wanted passengers connected with no login portal while Delta insisted on its own welcome screen. Then came leaked remarks from internal events, carried by an aviation gossip account: the chief executive supposedly wants nothing to do with Musk and believes the reputation stories are true. Even the leaker wrote he could not verify the quotes, and the airline stayed silent, so caution is mandatory. OneMileAtATime collected the claims about the Delta chief and its post stresses that none of the quotes are confirmed.

The Sky-High Preference Battle

The product scoreboard makes Delta's call look hard. United says Starlink flies on more than 560 of its aircraft with the whole fleet due by the end of 2027, while Delta's Amazon installs start no earlier than 2028 and cover only 500 jets. So a Delta passenger will wait years for speeds a United passenger enjoys today. Ryanair fought a similar battle in January: its chief priced antenna drag at 250 million dollars a year in extra fuel, Musk called him misinformed, and the row escalated into mutual insults. With Air France, Qatar, Emirates and British Airways all choosing Starlink, Delta walking alone suggests the decision weighs the man more than the machine.

The history of doubt begins with rockets. In 2014 a Russian official joked amid sanctions that Americans could reach the station by trampoline; when SpaceX flew two NASA astronauts on Crew Dragon on 30 May 2020, Musk fired back that the trampoline was working. Neil Armstrong told Congress in 2010 he lacked confidence in commercial rocketeers, Arianespace claimed in 2015 it would beat SpaceX in the open market, and NASA chief Duffy last October called Starship behind schedule while reopening the Moon contract to rivals. Last week's orbital flight closed that file too, and Rogozin was among the few who took the joke gracefully.

Tesla's charge sheet is just as long. Former GM vice chairman Lutz said in 2017 the company was going out of business and would never reach 2019, Morgan Stanley cut its worst case to 10 dollars a share in 2019. The freshest came last August from a former Waymo chief: a car with a minder inside is obviously no robotaxi. Within a year Tesla opened Cybercab rides to the public in Austin on 4 September, with no wheel and no pedals; Texas filings show 45 vehicles, and the American road-safety agency opened an inquiry. Reuters broke the safety-review angle of the Cybercab debut and its dispatch cites the Texas filing with forty-five vehicles.

When Dates Slip: An Honest Accounting

Now the honest accounting, because Musk's calendar truly cannot be trusted. The coast-to-coast driverless trip promised in 2016 finished on the last day of 2025 with zero interventions, eight years late yet real. The million robotaxis pledged for 2020 never appeared; Austin has a genuine service today, but not a million cars. The Roadster, unveiled in 2017 for 2020, now targets a demo on 15 October this year, and the company that called the Moon a distraction in January spoke in February of a self-growing lunar base within a decade. Any investor who ignores this pattern courts risk knowingly. Balanced voices deserve a hearing too: the Bernstein analyst's everyone relax, towers are going nowhere note captures today's reality. Starlink mobile can fill a genuine rural and disaster gap yet stands far from replacing terrestrial networks. Indoor coverage gaps, users hating carrier switches, and Cybercab cars leaning on remote operators show the technology's raw edges. These objections should calibrate expectations, not bury the product.

The investor rule that follows fits in two lines. Doubters are usually right about when and wrong about whether: the Semi was seven years late and real, the robotaxi was years late and real, Starship reached orbit late and real. So when a big name says it cannot be done, ask what that person has riding on it. An executive defending a giant business speaks with more certainty than the facts deserve, while the critic of the schedule is probably right and plans should assume so.

The coming tests are already dated. Late-October earnings calls will show whether carriers still wave Starlink away or start explaining how they will compete, which would signal something changed inside. The Starship ship-catch attempt targets late October, the Roadster demo lands on 15 October, and V2 satellites rise next year toward full phone service at the end of 2027. The chip analyst who called the Terafab factory harder than Mars missions and Sam Altman, who dismissed orbital data centres for this decade, herald the next wave of doubt. Smart people will scoff again; the question is who misses what this time.