Adobe introduced its AI Foundry service Monday, representing a major shift in how enterprises can leverage generative artificial intelligence by creating custom models trained exclusively on their own intellectual property and brand assets.

The new offering lets companies work directly with Adobe’s AI experts to build tailored versions of the company’s Firefly models, capable of generating text, images, video, audio, and 3D content that aligns with specific brand guidelines and creative standards.

Addressing Enterprise Demand for Brand-Specific AI

AI Foundry represents Adobe’s most ambitious enterprise initiative to date, moving beyond one-size-fits-all AI tools to offer what Hannah Elsakr, Adobe’s VP of Generative AI New Business, described as “a secure creative foundation that allows companies to scale on-brand content and remain competitive in the attention economy.”

Unlike Adobe’s traditional seat-based licensing model, AI Foundry operates on consumption-based pricing, mirroring how enterprises budget for AI infrastructure. The service positions itself as a co-development partnership, where Adobe teams assist with data curation, guardrail setup, and model integration into existing creative and marketing workflows.

Early adopters include major brands like Home Depot and Walt Disney Imagineering. Molly Battin, Home Depot’s CMO, called Adobe Foundry “an exciting step forward in harnessing cutting-edge technology to deepen customer engagement and deliver impactful content across our digital channels.”

How Adobe Tackles Intellectual Property Concerns

The service builds on Adobe’s Firefly models, launched in 2023 and trained entirely on licensed data with verified rights to avoid copyright disputes. Adobe guarantees that client intellectual property remains separate and never feeds back into the base models, addressing a primary concern for enterprises wary of AI-related legal risks.

“We’re going to retrain our own commercially safe Firefly models using the enterprise’s IP. We keep that IP separate. We never feed it back into the base model, and the enterprise itself owns that output,” Elsakr explained. Since Firefly’s launch, enterprises have created over 25 billion assets using Adobe’s AI technology.

Market Performance and Strategic Positioning

Adobe’s stock has faced challenges this year, declining 25% through Friday’s close, with some analysts expressing caution about the company’s AI monetization timeline. However, shares showed modest gains in early trading following Monday’s announcement.

Service Structure:

• Custom model development with Adobe AI experts
• Multi-modal content generation (text, images, video, audio, 3D)
• Consumption-based pricing model
• IP protection with isolated training
• Integration support for existing workflows

Who Benefits Most: Large enterprises with substantial brand asset libraries seeking to maintain consistent creative output at scale while protecting intellectual property rights. Companies in retail, entertainment, and consumer-facing industries will find particular value in brand-consistent AI content generation.

Technical Foundation: Built on Firefly models trained exclusively on commercially licensed content, ensuring legal compliance. Custom models remain isolated from base training data, providing enterprise-grade IP protection.

The platform marks Adobe’s recognition that generic AI tools don’t meet enterprise needs for brand consistency and legal compliance. By offering co-developed custom models with consumption pricing, Adobe addresses both the creative and business concerns holding back large-scale corporate AI adoption.