Product1 publisher3 min readPublished
The petition leans on Viasat's own coexistence filing, which SpaceX reads as showing the satellite could interfere about a third of the time in two Ka-band slices the FCC already fenced off in July.
The Product Desk · Product desk

Compiled by The Product DeskSomething wrong?How this is made
The person with a September start date on a Viasat contract is now waiting on a compliance notification from a regulator, a step no procurement plan accounts for. The FCC granted ViaSat-3 F2 US market access in late July and wrote the catch into the grant itself: Viasat must refrain from commencing, or otherwise cease, operations in the 18.8-19.3 GHz and 28.6-29.1 GHz bands until compliance is notified [5][6]. Coordination stopped being good practice there and became a precondition for switching on part of the payload.
That is the lever SpaceX is pulling. Its petition asks for a block or strict operational restrictions [1], and David Goldman's filing, which accuses Viasat of operating in the NGSO Ka-bands "without making any meaningful attempt to coordinate," says Viasat's own coexistence analysis, the study submitted to show F2 could share those bands, indicates the satellite could cause interference approximately one-third of the time when downlinking in them [3][4].
The two contested slices are 500 MHz each, 1 GHz in total [9]. A third of a 24-hour day is about eight hours [10]. That is the order of magnitude SpaceX's proposed remedy implies: a fixed-capacity geostationary service going quiet on those frequencies for a third of the day, sold to customers who, per DataCenterDynamics, are less sympathetic than SpaceX to the idea of Viasat standing down whenever a Starlink satellite is in the way [16].
The geometry explains why each side thinks the other should move. F2 sits at 35,786 km, while Starlink's shells run 340 to 570 km with more than 11,000 satellites in them [11][12], putting the geostationary satellite roughly 63 to 105 times further out [13]. Directional antennas and angular separation can keep beams apart in principle [15], and the LEO operators' position is that the burden falls on the satellite doing the illuminating [14]. Steering continuously around 11,000 moving objects is a different job from demonstrating once, in a filing, that it can be done.
Buyers of satellite capacity notice whether the link held at the hour they needed it, and escalate when it did not. The capacity sheet they buy against tells them throughput and coverage. Neither document names the frequency band a beam uses or the license condition attached to it, and that is where the risk currently sits.
Two questions follow, worth checking against your own contract. Does it specify capacity and coverage only, or does it name frequencies and the regulatory conditions on them? And if those two bands stay dark past September, is the notice period shorter or longer than the time needed to get another operator's terminal installed and commissioned? Named bands plus fast switching means a service credit tied to the FCC condition can be asked for and meant. Capacity-only plus slow switching means a licensing exposure nobody quoted you for is being carried anyway.
Viasat's 2023 loss of F1 to a reflector deployment malfunction was a hardware risk, and the insurance market repriced around it [17]. The exposure on F2 lives in a docket instead, which makes the compliance notification, rather than the launch photograph, the document worth asking Viasat for in September.</body_markdown> </invoke>
Ranked by verification strength, evidence, and original report placement.
SpaceX submitted a petition to the FCC seeking either a blockage or strict operational restrictions on Viasat's ViaSat-3 F2 satellite, claiming its radio frequency allocation and orbital configuration could create electromagnetic interference with Starlink.
The 25-page petition was written by SpaceX VP of satellite policy David Goldman, who wrote that "Viasat continues to operate in the NGSO Ka-bands without making any meaningful attempt to coordinate" and that Viasat has failed and even refused to coordinate with SpaceX for years.
SpaceX's petition claims that Viasat's own analysis, submitted to attest that F2 could coexist with Starlink, revealed the F2 was capable of causing interference approximately one-third of the time when using the 18.8-19.3 GHz and 28.6-29.1 GHz bands for downlinking.
The FCC cleared F2 for the US market in late July, on the condition that Viasat reach an agreement with the LEO operators alleging radio interference in the same spectrum bands.
The FCC grant specifies that Viasat "must refrain from commencing, or otherwise cease, operations in the United States in the 18.8-19.3 GHz and 28.6-29.1 GHz frequency bands pursuant to this grant of market access until compliance is notified."
Publishers with included, body-backed reporting in this cluster.
1 article · September 7, 2026
Follow any of these and your For You feed starts watching them — no settings page required.
product
Amazon Leo seeks FCC approval for new ground station on vacant plot in Twin Valley, Minnesota1 publisher
invest
SpaceX commits $100B to a Louisiana marsh, and only 10,950 flights a year makes it cheap3 publishers
product
FCC kills the gigabit goal, and 100/20 becomes the planning floor3 publishers
product
Starlink turns on Vietnam with an ownership waiver, a 600,000-subscriber ceiling and a 2027 rideshare1 publisher
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Primary filings, one voice
The procedural spine can be checked against paper: Goldman's petition is quoted with its page count, and the FCC's grant language is reproduced word for word, band edges included. The number that will decide the argument is weaker. Interference a third of the time is SpaceX's reading of a Viasat analysis DatacenterDynamics never shows, and neither Viasat nor the commission is quoted. The closing passage on Carr's likely leanings rests on nothing filed.
On orbit, two slices switched off
A six-tonne satellite already holds US market access and has a September service date, so this is not a paper dispute. What it cannot yet do is transmit in the gigahertz of Ka-band at issue, because the July grant keeps those bands shut until compliance is notified. The installed base on the other side of the argument is real too: more than 11,000 Starlink satellites are the population whose exposure the petition is built on.
One party's arithmetic travelling as fact
The reporting stays closer to the documents than most petition coverage, and it is careful to note F2 was already cleared in July. The overstatement is narrower: the one-third figure moves through the piece as a property of the satellite when it is SpaceX's interpretation of Viasat's numbers, Goldman's line about harm to millions of consumers goes unweighed, and the guesswork about which way the commission will lean adds temperature the evidence in the story does not supply.
Rival asking the regulator for silence
SpaceX gains commercially if a competitor's satellite stays dark in two Ka-band slices weeks before its service date, and Amazon Leo's parallel objections come from the same competitive position against the same legacy operator. With Viasat not answering the phone, the only party on the record is the one with something to win, and the petition it wrote is the story's main document.
Reliable on process, thin on outcome
Verbatim quotation from the petition and the grant makes the sequence of events dependable. Everything past that, how the FCC will handle the petition, whether continuous avoidance is workable for F2's customers, what the commission's politics imply, rests on one trade outlet's reading with no response from Viasat and no second account to compare.