A bold claim, but not yet a delivered reality
Elon Musk’s latest Starlink message is sweeping: SpaceX’s satellite network, he argues, could eventually carry a majority of the world’s internet traffic, at least in the countries where it is allowed to operate . The key point is that this is still a forecast, not a present-day measurement of global traffic. Current reporting describes Musk’s claim as a forward-looking ambition tied to Starlink’s next generation of satellites, not evidence that Starlink already provides most of the world’s internet .
That distinction matters. Starlink is already one of the most visible satellite broadband systems in the world, but “most of the world’s internet” is a much larger claim than serving remote households, ships, aircraft, emergency responders and rural users. The phrase implies a structural shift away from today’s internet backbone, which is still dominated by fiber, terrestrial mobile networks, subsea cables, data centers and national telecom operators. Musk’s argument is that Starlink can become a central layer in that system as artificial intelligence, robotics, vehicles and machine connectivity create far more bandwidth demand than traditional consumer browsing ever did .
The freshest development is that Musk has attached more specific engineering numbers to the vision. In a September 20 post captured by the Elon Musk Archive, he said next-generation Starlink satellites would provide about 10 terabits per second of connectivity in both directions, with a path to more than 100 terabits per second, and would carry a SpaceX-designed Nvidia Vera Rubin NVL72 computer . That post gives the latest version of the claim a more concrete shape: Musk is no longer talking only about more satellites, but about satellites that are also high-power, high-throughput computing platforms .
The Gen3 bet behind the headline
The technological basis of the claim is Starlink Gen3. Recent reporting says SpaceX is seeking authority to operate up to 100,000 non-geostationary Gen3 satellites, a number that would move Starlink into an entirely different scale of infrastructure . The application reportedly covers two very-low-Earth-orbit altitude bands, one around 323 to 327.5 kilometers and the other around 473 to 477.5 kilometers, with inclinations ranging from 26 degrees to 96.9 degrees .
The low altitude is not a footnote. Lower orbits can reduce latency and help failed satellites deorbit more quickly, but they also increase the complexity of traffic management, spectrum coordination and long-term orbital safety . A 100,000-satellite system would not merely expand Starlink; it would test whether regulators, launch systems and orbital operations can support a constellation of unprecedented commercial scale .
Musk’s September 20 technical statement also raises the bar on what “next generation” means. A target of roughly 10 Tbps per satellite, with a possible path to 100-plus Tbps, would be a dramatic leap from the capacity profile usually associated with consumer satellite broadband . The same post said each satellite would be a 250-kilowatt system, a power class that suggests thermal, electrical and operational challenges beyond those of a simpler relay spacecraft .
Still, the numbers should be treated as targets. One recent analysis cautioned that Musk’s 10 Tbps figure is a forward-looking disclosure and should not be confused with a capacity already validated in orbit . That warning is important because Starlink’s public story often moves faster than its regulatory and launch reality. A claimed design target can influence investor expectations, market perception and carrier negotiations long before the satellites prove the performance at scale.
Regulation is the immediate bottleneck
The most current regulatory detail is straightforward: SpaceX’s Gen3 filing has entered the FCC review process, but it has not been approved . “Accepted for filing” means the application is complete enough to be reviewed; it does not mean SpaceX has permission to deploy the full 100,000-satellite system . That is the central caveat in assessing Musk’s statement.
This regulatory posture changes the way the claim should be read. Musk can say Starlink has a path to carrying a majority of internet traffic, but that path runs through spectrum rules, orbital-debris analysis, national licensing, international coordination and likely opposition from competitors or public-interest groups . Even if the engineering works, Starlink’s global reach depends on where governments allow SpaceX to operate, a limitation Musk himself has acknowledged in the majority-internet claim .
The country-by-country issue is especially important. Starlink can be technically global without being legally universal. Some governments may restrict satellite broadband for security, sovereignty, telecom-policy or competition reasons. Others may welcome it for rural coverage, maritime connectivity, disaster response or aviation. Musk’s formulation — “where we’re allowed to operate” — is therefore not a throwaway line; it is the legal boundary around the entire ambition .
Why Musk links Starlink to AI and robotics
The story is not only about household broadband. Musk’s case for Starlink’s scale depends on a future in which machines generate far more traffic than human users. Current reporting says he tied the demand side to AI, humanoid robotics, vehicle robotics and a massive number of connected robots, arguing that this traffic could require far more bandwidth than the consumer internet of the past .
That is why the Nvidia Vera Rubin NVL72 reference stands out. Musk’s post suggests future Starlink satellites may not simply move data between users and ground stations; they may also carry powerful onboard computing resources . If SpaceX can combine connectivity, edge computing and global coverage, Starlink could become more than an access network. It could become a distributed space-based layer for data transport, machine coordination and possibly AI workloads.
But this is also where the claim becomes hardest to evaluate. The current sources confirm Musk’s technical target and the existence of the regulatory filing, but they do not prove that a space-based network can economically displace terrestrial infrastructure for most global traffic . Fiber remains extremely efficient for dense routes. Mobile networks are already embedded in national spectrum systems. Data centers are built around power, cooling and land-based interconnection. Starlink’s advantage is coverage and resilience; its challenge is matching terrestrial economics at massive scale.
Investors hear a revenue story
Musk’s majority-internet claim is also a business narrative. Reporting on the SpaceX call described Starlink as a major growth engine, with the promise of higher-capacity satellites supporting a potentially much larger revenue base . Musk’s logic is that if V3 or Gen3 satellites deliver far more bandwidth, Starlink can cut the price per bit and still expand total revenue .
That is a powerful story for investors, but it depends on execution. SpaceX must build the satellites, launch them quickly enough, win approvals, maintain reliability, manage congestion and convince customers that Starlink is not only available but preferable. The plan also depends heavily on Starship-class launch economics, because a 100,000-satellite system would be difficult to deploy and replenish through the existing launch cadence alone .
The current reporting also points to early operational steps. TeslaNorth reported that Starship Flight 14 was slated to deploy the first 26 production Starlink V3 satellites, each adding about 1 Tbps of constellation capacity . If that deployment path succeeds, it would give SpaceX a practical bridge from today’s Starlink system toward the much larger Gen3 vision. If it slips, the timetable behind Musk’s claim becomes less credible.
The bottom line
The current state of the story is clear: Musk is making a huge claim about Starlink’s eventual role in global internet traffic, and the newest reporting gives that claim a more specific technical and regulatory context. The promise rests on Gen3 satellites that Musk says could offer about 10 Tbps each, a proposed 100,000-satellite filing that is under review but not approved, and a business thesis that AI-era traffic will require new global infrastructure .
That does not mean Starlink is already providing most of the world’s internet in 2026. It means Musk is arguing, in September 2026, that Starlink can become the dominant carrier of future internet traffic if SpaceX can turn the Gen3 plan into working, licensed, economically competitive infrastructure. The claim is ambitious, measurable and still unproven. Its next test will not be another slogan; it will be approvals, launches, real capacity and sustained service at global scale.