Amazon builds its 1,000th Leo satellite while the count in orbit holds at 396

Vodafone and Vodacom have been beta customers for Amazon Leo across Europe and Africa since April ^[20]. Ars Technica counts airlines, shipping companies and governments among the businesses watching what Amazon sells next ^[28].

What Amazon is pitching this month is the plant. It took an empty 170,000 square foot building in Kirkland in December 2022 and had it producing satellites 15 months later ^[11]. It builds its own propulsion, and it replaced decades-old space testing with consumer-electronics methods that cut checks from weeks to hours, TNW reported ^[12]. Ars, which interviewed Badyal alongside production director Paul Palcisco and business vice president Chris Weber, describes a line that can build a handful of satellites every day ^[4]. Both outlets rank Amazon second among satellite manufacturers, behind SpaceX [c19, c22].

Customer coverage depends on launches. Amazon has 604 satellites built that are not in orbit ^[21]. The first generation needs 3,232 ^[13], so 2,836 have yet to fly ^[23]. The 396 up now are about a quarter of the 1,616 the FCC wanted by 30 July ^[24]. Amazon blamed rockets for that shortfall when it asked in January for a two-year extension, citing a shortage of launch vehicles ^[16].

The FCC's restoration test, as TNW reports it, credits the half of the program Amazon has already scaled. For the hardware condition, Amazon can point to 1,000 finished satellites [c1, c2, c14]. The manifest condition rests on bookings, and the next rocket on Amazon's list, Vulcan, is only now coming back from a grounding [c10, c16]. TNW's assessment is that "a thousand finished satellites and two Vulcans standing vertical is that argument, made in public" ^[27]. Badyal told Ars an Arianespace flight would follow the Vulcan return, with at least one more Vulcan this year ^[8].

For a team weighing Leo, two axes sort the satellites behind any offer. One is whether a satellite is in orbit or still on the ground. The other is whether it launched before the 30 July deadline or after it. The 396 flying now were all up by the July count ^[10]. Everything the Vulcan and Ariane flights add will be in orbit but post-deadline, required to avoid interfering with Starlink until October 2027 at the earliest ^[25]. The 604 on the ground serve no customer until a rocket flies ^[21].

The forcing question for any sales call is how many satellites behind the offer will be in orbit on the contract start date, and on which booked flight the rest arrive. I'd keep Leo at pilot scale until the in-orbit count moves, and tie any larger commitment to named flights on the manifest. The cost of waiting is staying with Starlink, which Ars notes has dominated the market for the last half decade, with OneWeb offering only limited service ^[19].