Stability AI: Open Generative AI Models | hokai.io
Stability AI, founded 2019, builds open-source generative AI for image, audio, and video. Creator of Stable Diffusion, valued at $1B, with $225M raised.
Stability AI is a London-based generative AI company founded in 2019, known for Stable Diffusion, the open-source image model downloaded 350 million times. The company raised $225M from Coatue, Lightspeed, Eric Schmidt, and Sean Parker, and is valued at $1B. With 191 employees and CEO Prem Akkaraju leading since June 2024, Stability now focuses on enterprise products including Brand Studio, launched April 2026.
Stability AI, founded in 2019 and headquartered in London, UK, is the company behind Stable Diffusion, the open-source text-to-image model downloaded over 350 million times. The company has raised approximately $225M, achieving a $1B valuation in June 2024. Under CEO Prem Akkaraju, it launched Brand Studio in April 2026 for enterprise creative teams.
Founded: 2019 · HQ: London, England, UK · Team: 150-200 · CEO: Prem Akkaraju · Funding: $225M total raised (seed Sept 2022: $101M led by Coatue and Lightspeed; June 2024: $80M from Greycroft, Coatue, Sound Ventures, Lightspeed, Eric Schmidt, Sean Parker; WPP Corporate Minority Mar 2025) · Valuation: $1B (June 2024)
About Stability AI
Stability AI was founded in 2019 by Emad Mostaque, a British-Bangladeshi mathematician and former hedge fund manager, and Cyrus Hodes, an AI policy entrepreneur who also co-founded The AI Initiative. The company is headquartered in London, England, UK, with additional team members in San Francisco, CA. Mostaque's founding thesis was that open-source foundation models would allow global developers to build AI-powered applications without proprietary API dependency or massive compute budgets. Investor pressure led to Mostaque's resignation as CEO in March 2024, triggering a leadership restructuring that brought in Prem Akkaraju in June 2024. Stability AI's flagship product is Stable Diffusion, a latent diffusion model for text-to-image generation first released publicly in August 2022. The model has been downloaded more than 350 million times across all versions, making it the most widely deployed open-source image generation model in existence. Stable Diffusion 3.5, released in October 2024, introduced architectural improvements and more permissive model weights suited to enterprise deployments. Additional products include Stable Audio 2.5 for enterprise audio generation, Stable Video 4D 2.0 for multi-view video synthesis, and SPAR3D for converting 2D images to editable 3D meshes in under one second. In April 2026, Stability AI launched Brand Studio, a commercial creative production platform targeting enterprise marketing teams rather than the developer community that built the open-source base. Brand Studio includes four modules: Brand Central for custom model training and campaign template storage, Producer Mode for automated workflow generation from creative briefs, Curated Model Routing for task-specific model selection, and Precision Inpainting for targeted image editing. The platform is available in free Core and paid Enterprise tiers. In May 2025, Stability and Arm co-released Stable Audio Open Small, a 341M-parameter on-device model that runs on smartphones in under 8 seconds without internet access. Stability AI raised a $101M seed round in September 2022, led by Coatue Management and Lightspeed Venture Partners, at a $1B valuation. In June 2024, the company closed an $80M round from Greycroft, Coatue, Sound Ventures, Lightspeed, and individual investors Eric Schmidt and Sean Parker, alongside a recapitalization that erased over $100M in outstanding debt and $300M in future supplier obligations. WPP, the global advertising group, participated in a Corporate Minority round on March 5, 2025. Total funding stands at approximately $225M with a $1B valuation as of June 2024. Revenue comes from the paid Stability AI API for image, audio, and video generation with per-generation pricing, enterprise Brand Studio subscriptions, and managed Stability AI Solutions deployments. CEO Prem Akkaraju stated in December 2024 that the company achieved triple-digit year-over-year revenue growth and had eliminated all outstanding debt. The company reported 2024 annual revenue of approximately $50M. Prior to the 2024 recapitalization, Stability burned more than $30M in a single quarter; debt elimination by late 2024 marked a decisive financial turnaround. Prem Akkaraju became CEO on June 25, 2024, after serving as CEO of Weta Digital, the visual effects studio behind Avatar and Lord of the Rings. Sean Parker, co-founder of Napster and Facebook's first president, is Executive Chairman and joined the board in June 2024. Filmmaker James Cameron joined the board as a director in September 2024. Founder Emad Mostaque departed in March 2024 and has since focused on decentralized AI initiatives. As of March 2026, Stability AI employs approximately 191 people across London and San Francisco, covering ML research, inference engineering, enterprise sales, and creative technology roles. Stability AI's stated mission is to build the foundation to activate humanity's potential through open-access AI models. The company publishes model weights for research and commercial reuse, which distinguishes it from closed-source rivals including Midjourney and OpenAI. Research areas include diffusion architecture, audio synthesis, 3D generation, and on-device inference optimization. The company joined the Tech Coalition in February 2026 to combat online child sexual exploitation alongside major technology platforms. In November 2025, the UK High Court ruled that Stability AI was not guilty of copyright infringement despite using Getty Images data in model training, removing a major legal overhang. In image generation, Stability AI competes with Midjourney, OpenAI DALL-E 3, and Adobe Firefly. Its core differentiator against Midjourney is local deployment: enterprises can self-host Stable Diffusion on their own hardware without API exposure or per-generation fees. Against Adobe Firefly, Stability offers open-source fine-tuning and broader model customization. In audio, competitors include ElevenLabs and Suno AI; Stable Audio 2.5's 8-step generation pipeline reduces production time from weeks to minutes. Brand Studio's 2026 launch directly challenges Adobe in enterprise creative tooling, though Adobe holds substantially greater distribution through Creative Cloud's 30M subscribers. Stability AI is incorporated in the UK and operates under UK GDPR and the EU AI Act as a general-purpose AI provider. International data transfers rely on Standard Contractual Clauses under GDPR Article 46 and European Commission adequacy decisions. The company holds SOC 2 Type II and SOC 3 certifications, verified by an independent third-party auditor across all five trust service criteria. No ISO 27001 or HIPAA-eligible designation has been publicly confirmed as of May 2026. The November 2025 High Court ruling against Getty Images reduced the company's remaining legal risk profile significantly. Stability AI remains private with no confirmed IPO timeline. Its $1B valuation and growing enterprise revenue position it as a plausible acquisition target for Adobe, Microsoft, or Google, all of which compete directly with Stable Diffusion in generative imaging. Near-term priorities include the commercial rollout of Brand Studio, expansion of media partnerships with Electronic Arts (February 2026), Warner Music Group (November 2025), and Universal Music Group, and continued open-source model releases to maintain global developer community traction.
Mission
Build the foundation to activate humanity's potential through open-access AI models.
Products
- Stable Diffusion 3.5 (Open-source image generation model): https://stability.ai/stable-diffusion
- Stable Audio 2.5 (Enterprise audio generation API): https://stability.ai/stable-audio
- Brand Studio (Enterprise creative production platform): https://stability.ai/brand-studio
- Stability AI API (Developer API for image, audio, and video): https://platform.stability.ai
- Stable Video 4D 2.0 (Multi-view video synthesis model): https://stability.ai/stable-video
Compliance
SOC 2 Type II, SOC 3, GDPR
Links
Website · GitHub · Twitter · LinkedIn · Blog · Docs
Frequently Asked Questions
What is Stability AI and what do they build?
Stability AI is a London-based generative AI company founded in 2019, best known for creating Stable Diffusion, the open-source text-to-image model that brought AI image generation to consumer hardware and global developer communities. The company builds AI models across image, audio, video, and 3D generation, with products used by studios, marketers, game developers, and individual creators worldwide. Stable Diffusion has been downloaded over 350 million times across all versions, making it the most widely deployed open-source image model in history. Beyond images, Stability produces Stable Audio 2.5 for enterprise audio creation, Stable Video 4D 2.0 for multi-view video synthesis, and SPAR3D for converting 2D images to editable 3D meshes in under one second. In April 2026, the company launched Brand Studio, a commercial creative platform designed for enterprise marketing teams that need automated, brand-consistent visual production at scale. Stability AI products are accessible via a paid API at platform.stability.ai, through AWS Bedrock, Azure AI Foundry, and the Brand Studio web platform, with self-hosted deployment supported on consumer and enterprise GPU hardware.
Who founded Stability AI and who is the CEO?
Stability AI was founded in 2019 by Emad Mostaque, a British-Bangladeshi mathematician and former hedge fund manager, and Cyrus Hodes, an AI policy entrepreneur who also co-founded The AI Initiative and several other ventures. Mostaque's central thesis was that open-source foundation models would allow any developer globally to build AI applications without proprietary API access or massive compute requirements. Mostaque served as CEO until March 23, 2024, when he resigned under sustained pressure from investors dissatisfied with the company's financial performance and debt accumulation. Following his exit, COO Shan Shan Wong and CTO Christian Laforte served as interim co-CEOs from March to June 2024. On June 25, 2024, Prem Akkaraju, former CEO of visual effects studio Weta Digital (Avatar, Lord of the Rings), was appointed CEO. Sean Parker, co-founder of Napster and Facebook's first president, joined as Executive Chairman in June 2024, and filmmaker and technology investor James Cameron joined the board in September 2024.
How much funding has Stability AI raised?
Stability AI's first major external capital was a $101M seed round in September 2022, led by Coatue Management and Lightspeed Venture Partners, with O'Shaughnessy Ventures also participating, valuing the company at $1B. In June 2024, the company closed an $80M funding round from Greycroft, Coatue Management, Sound Ventures, Lightspeed Venture Partners, and individual investors Eric Schmidt and Sean Parker. The June 2024 round was accompanied by a recapitalization in which suppliers forgave over $100M in outstanding debt and $300M in future spending obligations, removing the financial overhang that had threatened the company's operations. On March 5, 2025, WPP, the global advertising and marketing conglomerate, participated in a Corporate Minority investment round in Stability AI, signaling enterprise creative market validation. Total disclosed funding stands at approximately $225M with a $1B valuation as of June 2024. The company remains privately held with no confirmed IPO date or acquisition discussions publicly reported as of May 2026.
What products does Stability AI make?
Stability AI's flagship offering is Stable Diffusion 3.5, an open-source latent diffusion model for text-to-image generation available for self-hosting, via API, and through AWS Bedrock and Azure AI Foundry for enterprise deployments. Its open weights enable developers to fine-tune the model on custom datasets, a capability unavailable in closed tools like Midjourney and Adobe Firefly. Stable Audio 2.5 is an enterprise audio generation model that creates stereo tracks up to three minutes from text prompts in only eight computational steps, dramatically reducing production time versus earlier 50-step approaches. Stable Video 4D 2.0 handles multi-view video synthesis for 3D content creation in gaming and film pipelines. Brand Studio, launched April 8, 2026, is a commercial creative platform for enterprise marketing teams with modules for custom brand model training, automated brief-to-production workflows, task-specific model routing, and precision image inpainting. Stable Audio Open Small, co-developed with Arm in May 2025, is a 341M-parameter on-device model that generates stereo audio on smartphones in under 8 seconds without internet access.
Where is Stability AI headquartered and how big is the team?
Stability AI is headquartered in London, England, UK, where the company was originally incorporated and where its primary executive and research leadership is based. The company also maintains a team presence in San Francisco, CA, focused on US enterprise sales and business development. As of March 2026, Stability AI employs approximately 191 people across both offices, according to data from Tracxn, PitchBook, and GetLatka; PitchBook reports 184 and GetLatka reports between 186 and 197, with Tracxn's March 2026 figure of 191 being most current. This headcount reflects a contraction from the company's 2022-2023 peak, when rapid growth preceded financial difficulties and leadership restructuring. The June 2024 overhaul under CEO Prem Akkaraju and Executive Chairman Sean Parker stabilized headcount, and selective hiring has since focused on ML inference engineering, enterprise sales, and creative technology. Current open roles include Global Director of Sales, Solutions Engineer, and ML Inference Lead positions.
What is Stability AI's mission or research focus?
Stability AI's stated mission is to build the foundation to activate humanity's potential through open-access AI models, a positioning that differentiates it from closed-source competitors including Midjourney, OpenAI, and Adobe. The company publishes model weights for research and commercial reuse, enabling fine-tuning, local deployment, and downstream product development without access fees or usage-based pricing from Stability AI itself. Active research areas include diffusion model architecture improvements, enterprise audio synthesis, 3D generation from 2D inputs, and on-device inference optimization as demonstrated through the Stable Audio Open Small collaboration with Arm. The company joined the Tech Coalition in February 2026, a global alliance of technology companies combating online child sexual exploitation. In November 2025, the UK High Court ruled in Stability AI's favor against Getty Images in a copyright infringement case, a significant precedent for the generative AI industry. Stability AI has been recognized by Fortune as one of the 50 AI Innovators and by TIME as one of the Most Influential Companies, with Stable Audio named to TIME's Best Inventions list.
Is Stability AI compliant with SOC 2, GDPR, HIPAA?
Stability AI has achieved SOC 2 Type II and SOC 3 certifications, both verified by an independent third-party auditor across all five trust service criteria: security, availability, processing integrity, confidentiality, and privacy. These certifications are described by the company as a milestone placing it among a small number of AI providers to reach this security standard for the generative AI sector. For GDPR compliance, Stability AI relies on European Commission adequacy decisions under Article 45 GDPR and Standard Contractual Clauses under Article 46 GDPR for international data transfers, as explicitly stated in the privacy policy at stability.ai/privacypolicy. The Stable Assistant Privacy Policy also acknowledges UK GDPR and Swiss Federal Act on Data Protection obligations and discloses legal bases for processing including user consent. No ISO 27001 certification or HIPAA-eligible designation has been publicly announced by Stability AI as of May 2026; enterprises with HIPAA requirements should confirm compliance status directly with Stability AI's enterprise team. A dedicated trust center portal is referenced in company communications as the repository for compliance documentation.
Who are Stability AI's main competitors?
Stability AI's primary competitors in image generation are Midjourney, OpenAI DALL-E 3, and Adobe Firefly. Against Midjourney, Stability AI's key advantage is local self-hosted deployment: developers and enterprises can run Stable Diffusion on their own hardware without API exposure or per-image fees, which is critical for privacy-sensitive or on-premise workloads Midjourney cannot support. Against OpenAI DALL-E 3, Stability offers open model weights and fine-tuning capability; DALL-E 3 is only available as a managed API with no weight access. Against Adobe Firefly, which is embedded in Creative Cloud for 30M subscribers, Stability AI competes on customization and open-source flexibility but faces a significant distribution disadvantage. In audio generation, ElevenLabs and Suno AI are key rivals; Stable Audio 2.5's 8-step generation pipeline and enterprise pricing model differentiate it for media production workflows. The April 2026 launch of Brand Studio positions Stability AI as a direct competitor to Adobe in enterprise creative tooling, though Adobe's $12B+ annual creative software revenue and deep enterprise relationships create a formidable gap to close.