Elon Musk is angry with the lack of progress SpaceX has made in developing the Raptor engines that power its Starship rocket.

He described a dire situation the day after Thanksgiving in a companywide email, a copy of which was obtained by CNBC.

“The Raptor production crisis is much worse than it seemed a few weeks ago,” Musk wrote.

“We face genuine risk of bankruptcy if we cannot achieve a Starship flight rate of at least once every two weeks next year,” Musk added later.

Starship is the massive, next-generation rocket SpaceX is developing to launch cargo and people on missions to the moon and Mars. The company is testing prototypes at a facility in southern Texas and has flown multiple short test flights. But to move to orbital launches, the rocket prototypes will need as many as 39 Raptor engines each – necessitating a sharp ramp in engine production.

Musk’s email to SpaceX employees provides more context to the significance of the departure of former Vice President of Propulsion Will Heltsley earlier this month. Heltsley had been taken off Raptor development before he left, CNBC reported, with Musk noting in his email that the company’s leadership has been digging into the program’s problems since then – and discovering the circumstances “to be far more severe” than Musk previously thought.

SpaceX did not immediately respond to CNBC’s request for comment.

The SpaceX founder and CEO’s email was first reported by Space Explored, a subset of technology blog 9to5Mac.

Musk wrote in the email that he planned to take the long Thanksgiving holiday off. But, after discovering the Raptor situation, Musk said he would personally work on the engine production line through Friday night and into the weekend.

“We need all hands on deck to recover from what is, quite frankly, a disaster,” Musk wrote.

The billionaire founder has repeatedly described production as the most difficult part of creating SpaceX’s mammoth rocket. The company has steadily built up its Starship production and testing facility in Boca Chica, Texas, with multiple prototypes in work simultaneously.

The company’s next major step in developing Starship is launching to orbit.

Musk on Nov. 17 said that SpaceX will “hopefully launch” the first orbital Starship flight in January or February, pending regulatory approval by the FAA as well as technical readiness.

SpaceX wants Starship to be fully reusable, with both the rocket and its booster capable of landing after a launch to be recovered for future flights. SpaceX’s Falcon 9 rockets are partially reusable. The company can regularly land and re-launch the boosters but not the upper portion, or stage, of the rocket.

Musk earlier this month that he wasn’t sure if Starship would successfully reach orbit on the first try, but emphasized that he is “confident” that the rocket will get to space in 2022. He also noted at the time that Starship development “is at least 90% internally funded thus far,” with the company not assuming “any international collaboration” or external funding.

SpaceX has raised billions in funding over the past several years, both to fund Starship and its satellite internet project Starlink, with the company’s valuation recently hitting $100 billion.

But, while SpaceX has launched about 1,700 Starlink satellites to orbit so far, Musk said the first version of the satellite “is financially weak.” The company has been steadily growing Starlink’s user base, with about 140,000 users paying for service at $99 a month.

Earlier this year SpaceX outlined improvements for the second version of the satellite, with Musk saying in his email that “V2 is strong” but can only be launched effectively by its Starship rockets.

To date SpaceX has launched Starlink satellites with its Falcon 9 rockets, but Musk outlined that those rockets do not have the mass or volume needed to effectively deploy the second generation satellites. That means the success of the Raptor engine program is also critical to the long-term financial stability of SpaceX’s Starlink service, which Musk has talked about spinning-off in an IPO.

Notably, SpaceX is currently ramping up production of its Starlink antennas “to several million units per year,” Musk said in the email, but those will be “useless otherwise” if Raptor does not succeed.