Mark Dankberg is the CEO of a company that spent more than a decade building a satellite constellation that may have arrived five years too late. On Wednesday, April 29, Falcon Heavy launched the third and final ViaSat-3 satellite into geosynchronous transfer orbit, completing the most expensive broadband constellation ever fielded from geostationary Earth orbit. The satellite, designated F3, carries a terabit-class antenna built by L3Harris—not Northrop Grumman, not the same architecture as its two predecessors, because the first satellite's antenna failed catastrophically in orbit in 2023. What Dankberg said about that failure is revealing. "We didn't have to fix anything, but it's still fragile technology. In order to avoid those anomalies, often schedule turns out to be the dimension that you trade off—and we don't want that anymore, so that's why we're going into much simpler concepts." Translation: Viasat chose to redesign the antenna, eat the delay, and launch a less capable version rather than roll the dice on another tension-deployed reflector. That is not the decision SpaceX would have made. That is not the decision most space companies make. It is the decision a company makes when it has already absorbed one multi-hundred-million-dollar loss and cannot afford another.
To understand why this constellation matters and simultaneously why it arrived at a strange moment, you need to understand the competitive landscape it was built for versus the one it is entering. ViaSat-3 was conceived in the early 2010s, when geostationary broadband was the dominant architecture for global connectivity outside of fiber-rich urban areas. LEO constellations existed mostly as academic thought experiments and SpaceX's prototype Starlink. By the time ViaSat-3 F1 launched in December 2022, Starlink had already demonstrated that the LEO model worked, and by the time F1's antenna failure killed most of that satellite's capacity in July 2023, Starlink had deployed more than 4,000 satellites and become a household name. Amazon's Project Kuiper was in development. Telesat's Lightspeed was in advanced stages. The GEO broadband market that Viasat had spent a decade and several billion dollars preparing to dominate had transformed into a market that now had a proven LEO incumbent with a two-year head start and relentless manufacturing cadence. Viasat's competitive advantage—a single satellite that could deliver terabit-scale capacity to a region—suddenly looked less like a masterstroke and more like a constraint.
The technical specifications tell part of the story. Falcon Heavy delivered F3, a 6.6-metric-ton spacecraft, to geosynchronous transfer orbit in a highly efficient configuration. Center booster B1098 performed three separate burns over five hours to place the satellite in an orbit with a perigee of about 23,000 kilometers, an apogee near GEO's 35,786-kilometer altitude, and an inclination of only three degrees. This "EP-friendly orbit"—favorable for electric propulsion—is the explicit advantage of Falcon Heavy over the Atlas 5, which Viasat used for the first two satellites. The transfer will take roughly two months, after which F3 will maneuver to its reserved orbital slot at 155.58 degrees East longitude. Once there, it will add more than one terabit per second of capacity to the Viasat network. That is real capacity. ViaSat-3 F1 and F2 are already in operation or nearing service entry. This is not a paper constellation. It is a fully realized hardware milestone in an industry where many constellations remain half-finished or abandoned.
What created the conditions for F3's launch was the resolution of the F1 antenna anomaly and the decision to change course rather than replicate the failed design. After F1's bloom failure in July 2023—the large deployable reflector did not unfold correctly, destroying the satellite's capacity—Viasat faced a choice: redesign the antenna or absorb the loss, keep the schedule, and hope F2 would not fail the same way. F2 eventually completed its deployment successfully, but only after operational constraints from eclipse season caused delays and risk. F3 took the redesign path. L3Harris built an antenna that deploys "under control the entire time," as Dankberg put it, avoiding the mechanical tension release that caused F1's failure. The trade-off was that F3 likely has less deployable area and therefore less raw antenna gain than it would have with the original Northrop Grumman design. Viasat chose reliability over performance. That is a decision that tells you something about the confidence level inside Viasat about the margin for error in a competitive environment.
Who benefits from F3's arrival and who does not is straightforward. Viasat benefits in the near term: it can now sell connectivity backed by an actual constellation instead of a partially failed one. Customers in Asia-Pacific—F3's target region—benefit from competition against Starlink, which still has capacity constraints in some markets outside North America. Falcon Heavy's reusable booster side cores, which successfully landed at Cape Canaveral, benefit from the mission profile: this was the booster's 12th flight, and the reusable recovery cost for GEO missions has become standard. Northrop Grumman, however, does not benefit—L3Harris got the F3 antenna contract because Northrop Grumman's design failed. LEO constellation operators benefit immensely from the fact that Viasat spent a decade and untold capital preparing to field a terabit-class GEO satellite network just as LEO was proving cheaper, more global, and more flexible for enterprise and consumer use. Viasat's own investors may benefit in time, but only if the company can drive down the cost of Viasat-3 airtime to competitive levels before Starlink and Kuiper mature further. That is the actual question.
The honest read is this: ViaSat-3 is a legitimate technical achievement that proves GEO broadband is not going anywhere—it still has genuine advantages for high-capacity, regionally concentrated use cases like commercial aviation and maritime services. But the market Viasat-3 was designed to serve no longer exists in the form it did when this constellation was conceived. Starlink has already demonstrated that LEO broadband can work globally, that consumers will buy it, and that the manufacturing and launch economics make it cheaper per gigabit than GEO. Viasat's window to monetize this constellation before every major broadband operator has a LEO backbone is roughly 18 to 24 months. The company is not oblivious to this—Dankberg acknowledged in his comments that "the rise of Starlink and other LEO networks since ViaSat-3 was conceived has transformed the competitive landscape." But acknowledgment is not strategy. Viasat will need to price Viasat-3 capacity at a significant discount to GEO incumbents while simultaneously maintaining margins, or find enterprise and mobility markets where GEO's higher throughput and lower latency variability justify premium pricing. The constellation is real. The problem is real too.
Three specific deliverables over the next six months will determine whether Viasat can execute on this window. First: ViaSat-3 F2 must complete its final payload deployments and enter full service by early summer 2026, not slip further. F2 is currently in in-orbit testing and has completed its primary reflector deployment, but eclipse season delays have created schedule pressure. Second: Viasat must announce concrete pricing and contract wins for F3 capacity in the Asia-Pacific region before late 2026. Without signed commercial commitments, the constellation's ability to generate return on capital becomes speculative. Third: watch whether Viasat initiates any development work on a ViaSat-4 constellation or pivots entirely to LEO as a competitive hedge. If the company commits to another GEO build, it will signal genuine confidence in GEO's market durability. If not, it will suggest Viasat sees GEO as a legacy asset that must be monetized, not a platform for long-term growth.