The massive GPU clusters needed to train Stability AI’s popular text-to-image generation model Stable Diffusion are apparently also at least partially responsible for former CEO Emad Mostaque’s downfall – because he couldn’t find a way to pay for them.
According to an extensive exposé citing company documents and dozens of persons familiar with the matter, it’s indicated that the British model builder’s extreme infrastructure costs drained its coffers, leaving the biz with just $4 million in reserve by last October.
Stability rented its infrastructure from Amazon Web Services, Google Cloud Platform, and GPU-centric cloud operator CoreWeave, at a reported cost of around $99 million a year. That’s on top of the $54 million in wages and operating expenses required to keep the AI upstart afloat.
What’s more, it appears that a sizable portion of the cloudy resources Stability AI paid for were being given away to anyone outside the startup interested in experimenting with Stability’s models. One external researcher cited in the report estimated that a now-cancelled project was provided with at least $2.5 million worth of compute over the span of four months.
Stability AI’s infrastructure spending was not matched by revenue or fresh funding. The startup was projected to make just $11 million in sales for the 2023 calendar year.
Its financials were apparently so bad that it allegedly underpaid its July 2023 bills to AWS by $1 million and had no intention of paying its August bill for $7 million. Google Cloud and CoreWeave were also not paid in full, with debts to the pair reaching $1.6 million as of October, it’s reported.
It’s not clear whether those bills were ultimately paid, but it’s reported that the company – once valued at a billion dollars – weighed delaying tax payments to the UK government rather than skimping on its American payroll and risking legal penalties.
The failing was pinned on Mostaque’s inability to devise and execute a viable business plan. The company also failed to land deals with clients including Canva, NightCafe, Tome, and the Singaporean government, which contemplated a custom model, the report asserts.
Stability’s financial predicament spiraled, eroding trust among investors, making it difficult for the generative AI darling to raise additional capital, it is claimed. According to the report, Mostaque hoped to bring in a $95 million lifeline at the end of last year, but only managed to bring in $50 million from Intel. Only $20 million of that sum was disbursed, a significant shortfall given that the processor titan has a vested interest in Stability, with the AI biz slated to be a key customer for a supercomputer powered by 4,000 of its Gaudi2 accelerators.
Intel’s smaller than expected investment wasn’t the only fundraising effort to fall short. In July 2023 Mostaque reportedly pushed a fundraising plan that sought to bring in $500 million in cash plus $750 million in computing facilities from the likes of Nvidia, Google, and Intel.
These plans quickly fell apart, with Forbes pointing to an alleged meeting between Mostaque and Nvidia CEO Jensen Huang that ended disastrously. Mostaque, for his part, has denied any such meeting ever happened.
By late 2023, venture capital outfit Lightspeed had reportedly raised alarm bells expressing its shock after discovering the poor state of Stability’s cash flow, which apparently had not previously been disclosed. The VC firm also urged Stability’s board to sell the floundering business. But, despite reports the AI developer was shopping around for a buyer, a sale never happened.
By December, the startup had pivoted to a subscription model for commercial use of Stable Diffusion with prices starting at $20 per month.
However, behind the scenes, Stability was said to be weighing another strategy for propping up revenues: Reselling its compute resources as a managed service.
The plan was apparently to resell Stability’s GPU debt, which could have brought in $139 million in 2024 — assuming doing so didn’t violate its contract with AWS. Similar plans were said to have been drawn up to resell its GPU capacity at CoreWeave to VC firm Andreessen Horowitz.
Making matters worse, Stability was also having a hard time retaining staff, it is claimed. In November, Ed Newton-Rex, who was heading up development of a text-to-audio model, resigned in protest over arguments made by the biz that training on copyrighted material constituted fair use. Last month, Stability waved goodbye to several key researchers behind the development of its hallmark model Stable Diffusion.
The situation came to a head late last month when Mostaque revealed on social media he had resigned stating that “the concentration of power in AI is bad for us all,” and that in order to address it he had “decided to step down to fix this at Stability and elsewhere.”
Today, Stability AI is under new management with COO Shan Shan Wong and CTO Christian Laforte currently serving as interim co-CEOs until the position can be filled on a more permanent basis.
Despite the change in leadership, Stability’s future remains uncertain. Even if the biz manages to turn its financials around, it’s still facing down multiple copyright infringement cases brought by Getty and other artists, who allege their works were used without permission to train its signature model. ®
Source: Stability AI reportedly ran out of cash to pay its bills for rented cloudy GPUs
