Joel Kowsky / AP

In 2019, American astronauts were supposed to once again leave Earth from home turf.

At the beginning of this year, NASA officials felt nearly sure that at least one of the private companies hired to help fly astronauts into space would succeed. American crews regularly fly back and forth to the International Space Station, but they don’t get there on American vehicles, nor do they leave from Cape Canaveral, in Florida. Astronauts blast off in Russian vehicles, sitting shoulder to shoulder with cosmonauts, from the deserts of Kazakhstan. This year, a program known as Commercial Crew was supposed to change that, by launching NASA astronauts from an American space port, on rockets built and flown by an American company.

With only a few weeks left in the year, that’s not going to happen.

Commercial Crew has experienced, over the years, funding shortfalls, technical failures, safety concerns—including one fiery explosion—and other delays. The effort, as is the case with most spaceflight projects, fell behind schedule. NASA astronauts, who have already been selected and trained, will have to wait until next year to lift off from Florida.

Two companies in particular—SpaceX and Boeing—are vying to be the first to launch NASA astronauts into space again. One of them, Boeing, is scheduled to conduct a crucial test of its systems tomorrow.

Boeing’s crew capsule, named CST-100 Starliner, is scheduled to launch from Cape Canaveral tomorrow morning atop an Atlas V rocket, which is manufactured by United Launch Alliance, the company’s joint venture with Lockheed Martin, another longtime NASA contractor.

No astronaut will be on board this time, just thousands of pounds of cargo and a sensor-studded mannequin named Rosie, after the World War II–era cultural icon of women workers. The rocket will carry Starliner beyond the edge of space, where the capsule will ignite its own engines to give itself an extra push into orbit. It will circle the planet before nearing the ISS and autonomously docking with the station on Saturday morning. A week later, Starliner will detach, streak through the atmosphere, and parachute down to the ground.

A successful mission would bring NASA closer to putting its astronauts on board—and provide Boeing, which has been dealing with the flaws of its 737 Max model, with some sorely needed good press.

But spacecraft are extremely complicated creations, especially when they’re designed to carry people. That complexity is what keeps Steve Stich up at night.

“You have to get the systems put together right; you have to test them; you have to get the training right,” Stich told me in a recent interview at his office in NASA’s Johnson Space Center. “Are we doing all that correctly? And are we missing something?”

Stich is the deputy manager of theNASA program to send astronauts to the ISS. NASA used to fly its own astronauts to orbit on the space shuttle, a system that got off the ground in the early 1980s, and Stich spent more than 20 years working on the program before it ended in 2011 over a mix of safety and budgetary troubles and shifting policies.

The Commercial Crew companies face a set of technical hurdles both new and old. Their propulsion systems are more complicated than those on the space shuttles, Stich says. There’s also a bit of a knowledge gap. NASA flew a winged spaceship for 30 years. SpaceX and Boeing have returned to the days of parachute touchdowns, which were used in Apollo. “It’s almost like a generation has kind of gone by, and now you have a new generation that’s relearning parachute technology,” Stich said.

Earlier this year, Elon Musk’s SpaceX completed a mission similar to the one Boeing will attempt tomorrow. The company launched its own mannequin inside a capsule, named Crew Dragon, atop a Falcon 9 rocket. The mission went flawlessly, and NASA and SpaceX officials were ebullient. The image of the capsule splashing down in the Atlantic Ocean, trailed by a trio peppermint-colored parachutes, made the prospect of a crewed launch in 2019 feasible, almost expected. “Unless something goes wrong, I would think that we’ll be flying hopefully this year, this summer,” Musk told reporters in March. The NASA administrator, Jim Bridenstine, agreed.

Something went wrong the following month. Plumes of black smoke rose into the sky over the beaches near Cape Canaveral. It was a Saturday, and Stich was at home when he got the call. It was Ben Stahl, his Commercial Crew colleague, who’s in charge of overseeing the vehicles’ emergency abort procedures. The Crew Dragon had run a test to check its engines for an upcoming simulation. Stahl spoke in a language common in spaceflight, so technical it can obscure the drama of smoke, noise, and fire: “We had a significant anomaly.” The capsule had blown up. Stich remembers thinking, What did I miss? “How did our team miss it? How did SpaceX miss it?”

No one was hurt in the test failure, but the destruction of the capsule set SpaceX’s schedule, including its in-flight abort test, back several months. Boeing completed an abort test earlier this month; the company said it was a success, despite the fact that only two of the three parachutes were deployed. Commercial Crew officials say the issue has been fixed.

NASA officials now say SpaceX will be ready to launch astronauts in the first quarter of 2020, and Boeing is expected to do the same midyear.

NASA dictates the safety requirements that Boeing and SpaceX must meet, and agency engineers work closely with their commercial counterparts. Private companies have built hardware for NASA for years, but the Commercial Crew program marks the first time the agency has handed over the reins on development like this, from the propulsion systems to the interior design (which includes a fancy touch screen and sleek upholstery). The astronauts assigned to Boeing flights train at Johnson Space Center, in Houston, but SpaceX’s astronauts go to the company’s headquarters in Hawthorne, California.

“In some ways, day to day, it’s about the same,” Stich said. “It’s more challenging in that for [the] shuttle, I had one vehicle, one set of engines, one external tank, one solid rocket booster. Here I have two launch systems, two spacecraft, and two ground systems to keep track of.”

NASA’s last trip on the Russian Soyuz system is scheduled for April. The roundtrip cost per seat has risen over the years to as much as $86 million. It is a cost that many American lawmakers chafe at each year; the optics of asking another country to launch astronauts for you 50 years after you beat them to the moon don’t look great, either. If neither company is ready by then, NASA will have to buy more slots on Soyuz. Officials said this week they are already negotiating with Roscosmos, the Russian space agency.

The Commercial Crew program has experienced its fair share of scathing audits, by federal accountability agencies and NASA’s own officials, for its delays. The latest, from the agency’s inspector general, took aim at Boeing, saying that NASA “overpaid” Boeing by hundreds of millions of dollars for its work on Commercial Crew. The extra payments, which the report described as “unnecessary,” could have been avoided “through simple changes to the flight manifest.” Both NASA and Boeing disputed the findings.

So what’s taking so long, for both of them? This is extremely complex work, officials usually say, and that takes time. (The inspector general would add management and spending problems to the explanation.)

I asked Stich whether NASA leadership is getting impatient. This fall, when Musk unveiled a prototype for a spaceship unrelated to any NASA business, Bridenstine tweeted that he wanted “the same level of enthusiasm focused on the investments of the American taxpayer,” and that “it’s time to deliver.” (Bridenstine eventually went to visit SpaceX’s Hawthorne headquarters, where he and Musk tried to dismiss any tension.) Stich, who doesn’t have a Twitter account, ignored the kerfuffle. He doesn’t feel pressure from Washington. “Headquarters expects us to have these vehicles ready to fly when we’re ready to do it safely,” he said. “And if our launch dates move, they understand.”

It’s too early to mark calendars for a crewed flight. When I asked Stich whether that other common spaceflight fear keeps him awake—the loss of an astronaut—he told me he doesn’t think about that day very much, at least not yet. “I try to think about all the things that we need to do to make sure that doesn’t happen,” he said.

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.