Have you forgot when Stability AI was the rebel hero of generative AI? More recently, in 2022, its open-source Stable Diffusion allowed anyone to create amazing AI art without the control of corporations in between. Several years later and the company is awash with law suits, top management resigning/sacked and financial bankruptcy. What happened to cause a company worth one-billion dollars in the past to lose footing so dramatically? How about we all take a layer off? There is more to this failure than the failure of one company. It serves as a warning sign to all the AI industry.
The Rise: How Stable Diffusion Shook the AI World
Stable Diffusion was an out breath of fresh air when it got released. It was local, that is, it ran on your own GPU, as opposed to DALL-E or MidJourney was not censored and was entirely open-source. Six months and 10 million downloads later the artists, the developers, even the meme creators were flocking to it. Stability AI was officially anti-OpenAI, an advocate of the openness, in a time of walled-garden AI.
However, things were not so good early on. Its charismatic but divisive founder Emad Mostaque claimed ambitious things, like partnerships with Amazon, governments, even Hollywood. Many never saw the light of the day. At the same time, rival models such as MidJourney slowly developed sustainable subscription plans where Stability hit the compute expenses on its expense (with no apparent idea about a revenue plan).
Leadership Thyofony: The Emad Mostaque Issue
The style of leadership by Mostaque was a lightning rod. These ex-employees talk of a highly visionary but erratic CEO who focused more on hype than doing. In 2023 the main researchers and executives exited on mismanagement. According to one former engineer who talked to The Information: “I was working on the future, but the company operated much like a startup without brakes.”
And after that the trouble over money. By the end of 2023, Stability was frantically searching to secure emergency cash at a drastic discount. Why? It is reported that they had millions of dollars in unpaid AWS bills where cloud services were threatening its access. The problem with Stability was that it lacked a safety net that other new AI startups would have enjoyed via the deep pockets of Microsoft in the case of OpenAI (or due to the ability to make a lot of money via subscriptions in the case of MidJourney).
Legal Nightmares: Lawsuits That Choked Growth
The legal back fire was in the open-source of stability. The decision to sue Getty Images on scraping 12 million copyright images without permission due to lack of permission could be instrumental in the copyright law on AI. Next were class-action lawsuits filed by artists, who paralleled font producers OpenAI and Meta.
The irony is: in its pursuit of openness Stability became an easy rival. As opposed to closed models (where training data is concealed), Stable Diffusion was transparent and therefore lawyers had an obvious paper trail to play with.
The Fatal Flaw: No Path to Profit
Speaking of business models, or the absence of it. MidJourney had it cracked with over 200M in revenue annually through subscriptions. Stability? It depended on nebulous enterprise transactions and local affability. Plans were leaked including an “AI app store” which did not become a reality.
Real-life parable: Imagine Stability to be this kickass indie band who would not agree to be signed by a label. Very good as far as credibility is concerned, not good at paying rent. In the while, there were big backers in competitors:
- OpenAI Microsoft ($13B investment)
- Anthropic (B $4B commitment) 18
- Stability AI -???.
By 2024, even an open-source contender such as Mistral was moving faster than Stability with more well-financed strategies.
- Expert Take: Could Stability Have Survived?
I interviewed Sarah Chen (pseudonym), a VC that did not invest in Stability:
“The technology was game-changer but the economics was a hue-and-cry. No moats no monetization, just hype. In AI, you have to be deep pocketed or brutal in your business. Stability lacked the two.“
Final Thoughts: A Wake-Up Call for AI Startups
The failure of Stability AI is not only about bad luck it is a master tutorial on how not to manage an AI business. Idealism of the source? Noble. Ignoring revenue? Suicide. The lesson? Even the most innovative of technology can come undone without discipline, leadership, and a strategy of how to earn profits.
Where do Stable Diffusion go next?
Acquisition? (Perhaps through a cloud giant with a desire of an open model)
Slow death? Lawsuits and debts may just be stacking up (If).
Community rescue? Without Linux (AI, however)
Your turn: Did Stability deserve its fate or could the company have been rescued by superior leadership? Leave your take in the comment section below.