In 2016, I agreed to put a communications satellite owned by SES on a Falcon 9 booster that SpaceX had already flown. No commercial customer had ever done such a thing before, but as chief technology officer at SES, I had overseen the work our engineers had carried out with SpaceX. Their engineers had shown us how they tested the engines, checked the electronics, and prepared the booster for a second flight. I was confident in SpaceX’s ability to deliver, and they did. In March 2017, they used the booster to launch our satellite.I would oversee six Falcon 9 launches during my time at SES. But today, as an investor in growth-stage space companies, I find myself wondering what company I would book if I still had that job. SpaceX is setting aside the Falcon 9 and turning its attention to Starship, a much larger rocket; and from my conversations with people working at SpaceX, I can see why the company wants its engineers to change their focus. Most of the remaining Falcon 9 slots up to around 2030 are already booked up. So what happens now if you need to get to space?Those companies that have managed to secure a booking with SpaceX before 2030 will be glad for their foresight. In October 2025, the team at K2 Space, one of our portfolio companies, announced that they had booked a 2027 Falcon 9 flight for three satellites, having secured their place more than a year in advance. But even operators who book that far ahead can lose their place. Suppose a supplier calls you just before launch because a different customer has found a fault in a component on which you depend. Though your engineers may need just a few days to look into things, everybody else sharing the rocket has a customer waiting too. Under SpaceX’s rideshare terms, you could put the money you have paid towards another flight and pay a rebooking fee of 5% to 10%. You could work out the fee to the dollar and still have no real idea when you might start earning money from whatever it is that you have put into orbit.Those building satellites that weigh 100 kg or less may find themselves particularly exposed after 2030. On a SpaceX Transporter flight, such a company shares the rocket with dozens of other customers and buys only the space it requires. For a shared flight around Earth that passes near both poles, SpaceX advertises $350,000 for the first 50 kg and $7,000 for each additional kg. At those rates, a founder with a satellite weighing 100 kg can pencil in $700,000 for the ride. After 2030, it is hard to see how a space company could get to orbit on anything like those terms.In March, K2 Space launched its two-tonne Gravitas satellite on Falcon 9. Because it costs less to build than traditional satellites of similar power, K2 co-founder Karan Kunjur has argued that it still makes economic sense to launch such a spacecraft at a cost of roughly $7.2 million, compared with an estimated $600,000 if Starship delivers the promised savings. But K2 is a highly unusual case. Most satellite operators will struggle to offset higher launch costs with savings of that kind.At ICEYE, another of our portfolio companies, the team is looking at an alternative way to launch satellites for its European government customers. In September, its executives agreed with Arianespace to explore using Europe’s Ariane 6 rocket for future missions. Neither company announced a launch date. For now, ICEYE continues to fly with SpaceX. In October, the team launched four radar satellites on one SpaceX flight and four more on another just 19 days later.If your usual provider postpones a launch, the question is whether anyone else can take your satellite at short notice. Say you plan to launch a satellite on Neutron and Rocket Lab postpones the flight. How long can you afford to wait? You still have to pay your engineers, and until you get the satellite into orbit, you can’t sell whatever services you provide. You could try booking a flight with Blue Origin, provided its team can offer you a place on New Glenn in time; but even after engineers fly a new rocket for the first time, they may need years to build enough of them to fly regularly. My estimate is that Rocket Lab and Relativity Space may take until 2033 or 2034 to reach that point with Neutron and Terran R.At SES, we agreed to fly our satellite on a Falcon 9 booster that SpaceX had already used. Before we made that decision, our team worked alongside the people who had tested the booster and prepared it to fly again. Today, as an investor, I’d ask the people running a satellite company three questions: Which flight have you booked, how much have you agreed to pay, and what will you do if you lose your place? If I were one of their customers, I’d want those answers before I counted on having the satellite or payload in service by a particular date.The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.This story was originally featured on Fortune.com