SpaceX Chief Executive Officer Elon Musk said his company’s first complex test flight of the Starship spacecraft lived up to its expectations, despite a series of engine failures and other malfunctions that eventually led to the rocket losing control. Speaking in a follower-only chat on Twitter on April 29, the social media company he bought last year, Musk predicted that the company would be ready to attempt another launch in about two months, with a better chance of reaching space.

“The result was what I expected and maybe a little more than I expected.” He said these expectations are that the vehicle will leave the runway and obtain “important” data during flight, including maximum dynamic pressure, or max-Q. “Overall I think it was a really great flight.”

Musk outlined the flight, noting that the problems started immediately during takeoff, as 3 of the Raptor’s 33 engines in the vehicle’s super-heavy booster weren’t firing or shutting down during launch. “These engines didn’t blow up, but the system didn’t think they were healthy enough to get them to full power.”

He added that the minimum engine required for take-off was 30 employees, which made the vehicle appear to be tilted as it passed through the field. In T+27 seconds, the 19 Raptor engines lost communication and at the same time “some kind of energy event” broke part of the heat shield around this engine and three other engines. According to him, at that moment “fires” could be seen from behind the rocket.

At T+62 seconds, additional damage was done to the heat shield around the other Raptor Engine 30, but that engine continued to run. At T+85 seconds he said “really went into the fan”, cutting off communication with the other engine. “From this point on, we lose control of the rocket’s thrust vector,” he said, meaning the rocket could no longer steer.

It’s unclear what caused the engine failure, but Musk said it wasn’t caused by a “rock tornado” made up of concrete cushion debris from the engines pushing during launch. “Surprisingly, we see no evidence that the rock tornado caused significant damage to engines or heat shields,” he said. “It might be, but we haven’t seen evidence of that yet.”

SpaceX did not attempt to separate the upper part of the Starship from the Super Heavy as it crashed in the later stages of flight. Musk said it took much longer than expected for the explosives to shatter the vehicle’s tanks, taking about 40 seconds, and time controllers activated the flight termination system.

He predicted that with the next vehicle and refurbished runway likely to be ready in six to eight weeks, retraining the flight termination system would be a highly anticipated item for the next launch. “Hopefully in a few months we’ll be ready to fly again.”

Musk downplayed the damage done to the site itself, including concrete rubble scattered over nearly 400 acres scattered around the site and a smoke that dropped sandy material more than 10 kilometers away. “The debris was mostly sand and rock,” he said, “but we don’t want to do that again.”

Field changes include placing a water-jacketed “steel sandwich” underneath the launch mount. “You actually have a huge, heavy-duty steel showerhead that faces upwards,” he said, with this drainage system that softens dust and debris.

SpaceX was working on this device prior to launch, but it was not ready in time. “If we had waited to dig a hole, we wouldn’t have flown,” Musk said. Data from statistical fire tests in February, where 31 of Raptor’s 33 engines were running at 50 percent of rated thrust, caused “fairly moderate erosion” of the reinforced concrete Fondag used at the site. “We thought it would be good for a run.”

SpaceX also said it will also replace damaged tanks at the site’s tank farm, which are currently scheduled to be replaced with vacuum jacketed versions. He added that the launch tower itself did not suffer “significant” damage.

A super-heavy booster called Booster 9 will be used at the next launch, but the company hasn’t decided which of the Starship’s upper stages will fly, he said. “The engines in the next Booster 9 are much newer and more consistent, with significant reliability improvements,” he said, adding improved protection. “I think with the Booster 9 we will see a much more reliable engine condition.”

He expressed optimism that the second run would be successful, at least because of the rank split. “Our goal in the next flight is to make a landing and hopefully a successful landing and get into orbit.” “I think we have a good chance of getting into orbit on the next flight.”

Later in the nearly hour-long conversation, he said the next launch gave him “more than a 50% chance of going into orbit.” But this launch will repeat the flight profile of the original flight, a “near-orbit” orbit that would land the Starship on the Hawaiian coast 90 minutes after launch, one full orbit outside.

According to Musk, SpaceX will attempt to launch four to five Starships this year. “I’d be surprised if we left without orbiting this year,” he said, giving the company “more than 80%” a chance to do so, and within 12 months it had risen to almost 100%.

He said the company will spend about $2 billion this year on Starship, which the company believes can sustain without external financing.

“Again, excitement is guaranteed,” Musk said of the next launch. “Not success.”