A Seattle-Area Dream With a Long Memory
Long before Starlink became a name for satellites overhead, Seattle had already hosted a sweeping attempt to send the internet through space. Teledesic began there in 1994, bringing together telecommunications entrepreneur Craig McCaw and Microsoft founder Bill Gates around a plan for a satellite internet network. The proposed fleet began at 840 satellites, then narrowed to 288, while the project’s total cost was said to be about $9 billion. It was an audacious calculation for its moment: build a vast network in orbit and use it to reach people from above, rather than waiting for terrestrial infrastructure to arrive below. The account of Starlink’s Redmond origins places that earlier Seattle project squarely in the lineage of what came later.
Teledesic effectively ceased operations around 2002, leaving behind an idea whose technical and financial demands had outrun its era. Yet the idea retained a human presence in the region. Lawrence Williams came from Teledesic and joined SpaceX in 2003, carrying experience from the first Seattle-area effort into a young rocket company that would eventually pursue a constellation on an even larger scale. Williams’s move from Teledesic to SpaceX gives the story a continuity that is easy to miss when satellite networks are described as sudden breakthroughs.
That continuity matters because the central ambition had already taken shape: a large group of satellites in low Earth orbit, working together as a communications system. The people and companies changed, while the underlying proposition remained demanding. A network like this required money, manufacturing, launches, spectrum, software, and the patience to keep building through years of technical work. Teledesic showed the scale of the dream. Williams’s path showed that the dream had stayed close to the people who would help carry the next version forward.
By early 2014, Elon Musk and Greg Wyler were planning that next version together under the name WorldVu. Their concept called for a constellation of around 700 relay satellites in low Earth orbit, a number that already placed the proposal in a different category from the satellite fleets that had come before. Accounts of Starlink’s early development describe Musk and Wyler working together on the plan before their discussions broke down in June 2014. Their paths then separated, and SpaceX began pursuing its own route toward the same broad horizon: internet delivered by a coordinated fleet circling Earth. The Seattle-area dream had found its next vehicle, and the work was about to acquire a home of its own.
SpaceX Gives the Internet Its Own Factory
That collaboration soon gave way to a separate SpaceX route. In the second half of 2014, the company filed an application with the International Telecommunication Union through Norway’s communications authority, using the name STEAM to register its proposed constellation, according to the project’s published history. The filing turned an ambition into a formal claim on spectrum and orbital capacity. It also gave SpaceX a path toward building the network as its own industrial system, with the company controlling the satellites, their production and the operations that would keep them working overhead.
SpaceX brought that system into public view in January 2015, when it announced Starlink alongside the opening of a satellite-development facility in Redmond, Washington. The facility gave the project a dedicated home for research, development, manufacturing and on-orbit control, as the contemporary account of the program records. Those jobs belonged together. A satellite network of this kind depended on hardware that could be designed repeatedly, built repeatedly, sent into orbit repeatedly and then managed as a fleet once it arrived there.
The operating idea was clear in its scale. Starlink would use thousands of small, mass-produced satellites in low Earth orbit to provide satellite internet access, while also serving military, scientific and exploratory purposes, according to the same account. Low orbit placed the satellites much nearer to users than earlier communications systems based far above Earth, and mass production made the constellation a manufacturing challenge as much as a spaceflight challenge. Redmond therefore functioned as a factory for an internet system, one whose essential machinery would be built on the ground before it could reach the sky.
That commitment carried a price large enough to define the company’s next decade. In May 2018, SpaceX estimated that designing, building and deploying the constellation would cost at least $10 billion, according to Starlink’s published record. The estimate captured the wager embedded in the Redmond operation: SpaceX was funding an enormous communications network years before public service would begin, while asking one facility to unite engineering, production and fleet control under a single roof.
Redmond’s work gave the project its physical center, yet its distance from the company’s main headquarters created a second center of gravity. The satellite effort now had its own people, its own operating rhythm and a mandate large enough to test how SpaceX organized itself when the internet became part of its mission.
Why Redmond Became Starlink’s Crucible
That second center of gravity sat a long way from the company’s formal headquarters. Redmond and SpaceX’s Hawthorne, California, headquarters lie about 1,600 kilometers apart, according to an account of the project’s early years, a distance that gave the satellite effort room to develop its own working culture. A former SpaceX employee told Business Insider that the company placed Starlink in Redmond partly to keep the new venture from drawing focus away from the Falcon teams at its main facilities in Texas and California. The arrangement made practical sense: SpaceX could pursue an ambitious satellite program while its launch organization kept moving on its established mission.
Redmond still needed a clear chain of responsibility, and that chain ran through Rajeev Badyal, the vice president leading the satellite division. Around him, the operation brought together specialists whose work sat at the center of the system’s development. Mark Krebs was responsible for satellite system design, while Serena Byers arrived from headquarters in the spring of 2018 to join the satellite division. The same account records each of those roles and moves, offering a glimpse of an organization assembling around a project with unusually broad demands.
Badyal’s group had to turn a large corporate ambition into decisions that could survive contact with manufacturing, spacecraft design and the pace of a launch company. That work carried a particular weight in Redmond, where the satellite operation had its own facility, its own leadership and a physical separation from the people whose rockets would eventually carry the system into orbit. The separation gave the team focus, yet it also made coordination a central part of the job. A satellite design decision, a production choice and a launch requirement all belonged to the same program, even when the people shaping them worked far apart.
By the spring of 2018, the pressure inside that structure had begun to concentrate. Byers’s move from headquarters placed another experienced SpaceX hand inside the satellite division, while Badyal and Krebs held responsibility for the leadership and technical design at the heart of the effort. Redmond had become the place where a vast plan had to become an operating machine, with leaders expected to convert intent into a system that could be built and flown. That expectation was about to force a sharp decision from the top.
Musk Clears Out the Satellite Leadership
They were expected to deliver, and in June 2018 Elon Musk made clear that the schedule would now govern the organisation. On a Monday that month, about 300 people at the Redmond operation received a mass email telling them Musk would hold an emergency all-hands meeting. When the group assembled, Musk announced that the satellite division’s entire leadership team would be fired, according to an account of the meeting and its aftermath. The decision reached the people responsible for turning the division’s work into a functioning constellation, placing the pressure that had gathered inside Redmond directly on the people who had been leading it.
Rajeev Badyal, the vice president who led the satellite division, was dismissed alongside the rest of the leadership team. Mark Krebs, who oversaw satellite system design, also left in the same sweep. Contemporary reports from CNBC and the Los Angeles Times said Musk dismissed at least seven executives, a scale that gave the meeting the force of a full operational reset rather than an ordinary leadership change, according to the later account citing those reports. Redmond still held the same responsibility after the meeting: its teams had to build a satellite network capable of moving from development into orbit. The difference lay in the time available and in the leadership now tasked with carrying the work through.
That time frame arrived immediately. The goal set after the dismissals called for the first satellite constellation to launch by mid-2019, a target recorded in the account of the reset. The deadline turned the June meeting into a hard dividing line for the programme. Every remaining decision in Redmond now led toward a first operational deployment, and every team had to work against a date close enough to shape the daily rhythm of the effort. Within a year, the plan would face its clearest test: whether the satellites could leave the factory, reach low Earth orbit and begin operating together as a network.
The Launch Cadence That Made the Plan Real
Enough work had already reached orbit to give that date a practical meaning. In February 2018, SpaceX flew two test satellites, Tintin A and Tintin B, and Musk said their communications quality was “good enough to play video games,” a compact verdict that put a demanding everyday use case behind the larger engineering programme. The tests gave the team a working reference point as the schedule tightened: a network designed for use by people on the ground had to move from prototypes to satellites that could operate together at scale.
The regulatory runway was already opening beneath that effort. The FCC approved SpaceX’s initial constellation of 4,425 satellites in March 2018, according to Starlink’s published history, giving the company authority for a fleet vastly larger than the two spacecraft that had flown the month before. That approval made repetition the central task. Redmond needed a satellite design that could be built in volume, while launches had to carry those satellites into an orbit where each new batch could join a growing system.
That system took its first operational form in May 2019. A Falcon 9 carried the first 60 operational Starlink satellites on May 23 in US Eastern time, May 24 UTC, as the launch record notes. The date landed within the mid-2019 objective set after the leadership reset, yet its importance lay in what the payload represented. Sixty satellites turned a plan governed by filings, prototypes and internal deadlines into a deployable network, with hardware already positioned for the next launches to build upon.
The chosen orbit made that sequence useful. Starlink satellites operate at around 550 kilometres above Earth, according to the same account, placing the constellation in low Earth orbit and allowing the network to grow through repeated batches rather than waiting on a single large spacecraft. Each launch could extend the fleet, and each extension made the initial deployment less a ceremonial first step than a production rhythm.
By sending 60 operational satellites up together, SpaceX had established the cadence its original ambition required: build many small satellites, launch them in groups, and let an expanding fleet take shape overhead. The test flight had shown that communication worked. The May deployment showed that the organisation could begin delivering the system itself.
From Public Beta to a Global Subscriber Base
That ability to deliver the system created the next task: turning an expanding fleet into a service people could buy and use. Paid public service began in the United States in October 2020, giving the constellation a commercial purpose measured at ground level, through homes, businesses and the terminals connecting them to the satellites overhead. Later that month, on October 26, SpaceX opened its public program under the deliberately candid name “Better Than Nothing Beta”. The offer cost $99 a month, with a $499 Starlink Kit, setting a clear price for entry into the early network.
Early customers were joining a service still finding its shape, and their numbers quickly gave that experiment a different scale. By June 2021, SpaceX reported around 69,000 active users in 11 countries. Six months later, in December 2021, the company reported one million active subscribers. The shift from tens of thousands of users to one million showed that the service could move beyond its initial beta audience while the constellation continued to serve more places.
That growth kept gathering speed. Starlink reported two million active customers across more than 60 markets by September 2022, then passed 4.6 million customers by the end of 2023, before reporting more than nine million customers across more than 155 countries and markets in 2024. Each milestone reflected the same commercial handoff: satellites launched in batches became a recurring service, and a recurring service gave the network a growing base of paying users.
By September 30, 2026, that base had reached 13.6 million consumer subscribers. Starlink now provides broadband internet service in approximately 160 countries and territories, a reach that makes the customer story inseparable from the size of the fleet supporting it. The next measure of Starlink’s transformation sits above those subscribers, in the sheer number of satellites SpaceX has placed in orbit and the influence that scale has created.
How Scale Turned Into Dominance
The question is how many satellites SpaceX has placed in orbit. By October 2026, the answer stood at approximately 11,150 Starlink satellites in low Earth orbit, a fleet large enough to account for approximately 75% of all active maneuverable satellites around Earth, according to Starlink’s published profile. That share gives the constellation a presence measured across the orbital environment itself, rather than solely through the people and places it serves. Each satellite belongs to a system that SpaceX has kept enlarging through repeated launches, and the accumulated fleet has become one of the company’s clearest expressions of industrial scale.
The scale carries particular force because SpaceX controls the route into orbit as well as the network waiting there. In 2025, Business Insider reported that SpaceX’s share of everything sent into space was close to 80%. That launch position gave the company a powerful deployment advantage: a constellation could grow through the same organization that sent its satellites aloft. The result is visible in the gap between a satellite project and an orbital fleet. Starlink moved into a category where its own spacecraft made up three-quarters of the active maneuverable satellites in Earth orbit, while SpaceX retained an enormous share of the capacity used to place payloads there.
Those two forms of scale, the satellites already overhead and the ability to continue launching them, reshaped Starlink’s role inside SpaceX. By the end of 2025, Starlink had become the company’s largest business segment, generating $11.4 billion in revenue and $4.4 billion in operating income, according to the company figures collected in Starlink’s profile. Revenue at that level turns the constellation into a central business engine, while the operating income shows how the network’s expansion had begun to produce substantial returns.
The business outcome also clarifies why the fleet’s size matters. More than eleven thousand satellites represent a service infrastructure on an exceptional physical scale, and the company’s launch reach gives that infrastructure room to keep evolving. Starlink’s dominance therefore rests on a reinforcing arrangement: satellites generate the network’s reach, launches sustain the fleet, and the resulting service has grown into SpaceX’s largest segment. That arrangement places a private communications network at the center of a much wider set of decisions about where connectivity can reach and how it can be used.
Starlink’s Power Reaches Far Beyond Broadband
The result is a network whose importance reaches well past the subscriber count and revenue that made its scale visible. From the outset, Starlink was designed for internet access alongside military, scientific and exploratory purposes, a range of uses that gives the constellation a presence in settings where reliable communications can carry immediate consequences. The project’s original design therefore contained the seeds of its present influence: the same infrastructure that connects an individual customer can also become part of a wider operational system.
Ukraine made that influence starkly concrete. On 26 February 2022, Elon Musk wrote that “Starlink service is now active in Ukraine. More terminals en route,” placing the network inside a war already shaped by the availability of communications. The service’s activation created a continuing relationship between SpaceX, Ukrainian users and government institutions; Musk later said the service was costing $20 million a month in Ukraine, and the US Department of Defense signed a contract with SpaceX in June 2023 to pay for its use there. That arrangement showed how a commercial satellite network could become a communications resource with military significance.
It also put Musk’s personal decisions under unusual pressure. In 2023, he said he had declined a request to activate Starlink in support of a Ukrainian assault on Crimea, explaining that he feared “SpaceX would be explicitly complicit in a major act of war.” Business Insider reported the statement as one example of the authority concentrated in the company operating the service. The question extends beyond a single decision: a constellation built to carry data across vast distances can shape who communicates, where that communication reaches and when access changes.
The network’s reach also changes the view from Earth. On 3 June 2019, the International Astronomical Union warned that large constellations such as Starlink could impair astronomical observations, bringing an older scientific concern into direct contact with the rapid expansion of satellite communications. The IAU’s statement framed the tradeoff clearly: a system that places broadband and other services within reach of users on the ground also places thousands of moving objects across the sky observed by astronomers. As Starlink’s decisions continue to matter in conflict zones and in observatories, the Redmond reset reads as the beginning of a communications system with power in both places.
Comments (0)