SpaceX and Qualcomm, the US chip designer, announced a multi-year agreement on 4 November under which Starlink Mobile will use Qualcomm’s Dragonwing Cellular Infrastructure platforms to build a nationwide 5G network in the United States. A day later, on Deutsche Telekom’s 5 November earnings call, the German operator’s chief executive Tim Höttges said the effort would need at least €200 billion — roughly $220 billion — to seriously challenge America’s established carriers, according to Yahoo Finance.
Key facts
- Qualcomm called the agreement its “biggest 5G deal yet,” per TrendForce.
- Nationwide rollout is planned to begin in 2027, with tens of thousands of sites envisaged.
- Höttges said SpaceX has acquired 76 megahertz of spectrum, against roughly 400 megahertz held by T-Mobile US, Deutsche Telekom’s American subsidiary.
- As of October 2026, SpaceX had launched more than 11,500 Starlink satellites, with over 11,156 operating in low Earth orbit. TrendForce’s reach figures are estimates rather than disclosed totals: it puts coverage at 150–160 countries and regions and users at 9–10 million.
Why it matters
TrendForce frames the deal as a move toward “space-ground integration” — pairing the Starlink constellation with a ground network — rather than a satellite company filling coverage gaps for someone else. In its view that could challenge AT&T, Verizon and T-Mobile, the three incumbent US wireless carriers, whose ground-only model has meant heavy spending in crowded cities and thin returns in remote areas.
The speed problem driving the buildout
Scale in orbit has not translated into mobile broadband, TrendForce says. It reports that Starlink’s Direct to Cell service — which connects ordinary phones to satellites — has hit hard technical limits: only 5–10 MHz of cellular spectrum from partner carriers, limited handset transmit power and antenna gain, and satellite payload throughput of just 10–20 Mbps. Together, TrendForce says, those constraints hold real-world per-user speeds to around 1–4 Mbps.
In practical terms that is early-mobile-data territory: enough for messaging, voice and maps, but roughly an order of magnitude slower than a typical 4G cell or a fixed home broadband line — and a different product from Starlink’s dish-based home service, which uses a dedicated terminal rather than a phone.
That split also shapes what buyers pay and where they can sign up: Direct to Cell reaches handsets only where a partner carrier resells it, while the dish service is sold directly by Starlink with its own hardware and monthly price. SpaceX has not said what Starlink Mobile will charge, or which markets beyond the US it will cover.
When SpaceX picked Qualcomm’s Dragonwing platforms, TrendForce says, it bought hardware that lets it handle heavy broadband traffic in densely populated areas, where satellite capacity cannot stretch.
Pointing the signal
Expanding ground capacity is only half of it, TrendForce writes: traditional satellite signals spread over wide areas and weaken before reaching receivers, making them prone to distortion. Starlink V3 satellites reportedly carry 64-transmit, 64-receive Massive MIMO antenna arrays, while the Dragonwing platform brings 3D beamforming to Starlink Mobile’s terrestrial base stations, helping aim radio signals more precisely at individual devices.
What the incumbents say it costs
Four figures are now in circulation, and they measure different things:
- At least €200 billion (about $220 billion) — Höttges, via Yahoo Finance. The total he estimates is needed to seriously challenge established US carriers, covering spectrum (he said expanding SpaceX’s position would take “triple digit billions”) plus terrestrial infrastructure. Yahoo Finance stresses this is his estimate, not a disclosed SpaceX spending commitment.
- $20 billion to $30 billion — Höttges. Tower rental alone, benchmarked against T-Mobile’s 85,000 towers, before SpaceX installs antennas and other equipment.
- A further $120 billion to $130 billion — telecom specialist William Webb, cited by Yahoo Finance. Additional investment he estimates would be needed to rival the incumbents.
- Up to $500 billion — rival operators. The high end of the range, as rival operators have put the bill as high as $500 billion.
Höttges paired the numbers with a challenge: “If he has the money, go for it,” he said, before welcoming Elon Musk’s ambitions with a pointed “Welcome to the party.” He also argued that incumbents had refined their networks over two decades. Yahoo Finance’s account of a Reuters report — not the Reuters wire itself, which this article has not seen — has him adding that he expects SpaceX to enter the market regardless of the obstacles: “We have to assume that it is coming.”
Tens of thousands of sites, or more than 80,000?
The two site-count figures come from different places and count different things. The “tens of thousands of base stations” figure traces to the companies themselves, in TrendForce’s account of the Qualcomm agreement, and refers to equipment in the announced plan. The “at least 80,000 American cell sites” figure comes from earlier third-party reports cited by Yahoo Finance about SpaceX’s leasing discussions, including with the US tower operator Crown Castle, and counts physical locations it is said to be pursuing. SpaceX has confirmed neither as a committed total, so the lower, company-sourced number is the more conservative reading and the 80,000-plus figure is better treated as a reported ambition.
The partner that is also a rival
Deutsche Telekom sits on both sides of this. Yahoo Finance notes T-Mobile was Starlink’s first direct-to-cell partner, but says its parent company wants competing satellite providers rather than dependence on a single operator. Höttges also argued that satellites cannot fully replace terrestrial networks, particularly in cities and inside buildings.
Background: the money behind it
The push comes as SpaceX’s business model shifts. TrendForce reports that in the third quarter of 2026 Starlink contributed 36% of SpaceX’s revenue but nearly 85% of its operating income, and reads Starlink’s contribution to SpaceX’s quarterly results as the company moving beyond a capital-intensive orbital buildout into a cash-generating phase, with both average revenue per user and subscriber numbers expected to surge once the ground network arrives. For Qualcomm, TrendForce says, the agreement opens another growth front as the smartphone market matures, giving it a foothold in “space-ground integrated” telecom infrastructure and a path toward its Open RAN and 6G ambitions.
The first measurable test is 2027: whether the Dragonwing order becomes working cell sites on leased towers that year, and how much spectrum SpaceX adds to its 76 megahertz, will show which of these cost estimates was closest.
Frequently Asked Questions
Which Qualcomm units will Starlink Mobile’s cell sites actually run on?
TrendForce names two: the QRU100 radio unit and the QDU100 distributed unit. It describes them as base-station-class hardware forming the radio and baseband backbone of a conventional 5G cell site, and says they are large, power-hungry and heat-intensive — not parts that could go inside a smartphone.
Do you need a special phone to use Starlink Direct to Cell?
No. According to TrendForce, Direct to Cell satellites emulate standard 4G and 5G cell towers, so existing smartphones that support the standard bands are natively compatible. Users would need no special equipment.
Is Deutsche Telekom’s own business under pressure from this?
Not in its latest results. Yahoo Finance reports Deutsche Telekom raised its 2026 outlook on 5 November after strong quarterly figures and announced plans for a €2 billion buyback in 2027. Höttges did warn, though, that investors already expect tougher competition and pressure on wireless operators’ long-term profitability.
![Using Your Own Router with Starlink: Bypass Mode [2026]](https://starlink-news.com/wp-content/uploads/2026/09/using-your-own-router-with-starlink-bypass-mode-20-f6c09f95-120x86.png)
