Adobe in the AI era: confidence, questions, and your next move
Investors are unsure if Adobe can keep its edge as AI makes content creation easier and cheaper. The stock is down about 25% this year. At its Los Angeles conference, Adobe showed new features and a clear shift: let creators use AI models from Google and OpenAI inside Adobe tools, including Photoshop.
Here's the signal: Adobe wants to be the hub, even if the best model for your job isn't theirs.
Why investors are skeptical
New AI tools sit outside Adobe's ecosystem. OpenAI's Sora and Google's Veo are grabbing attention for video and image generation. Consumer platforms like Canva keep growing. That combo puts pricing and product pressure on Adobe.
Citigroup's Tyler Radke flagged "structural AI-driven competitive and pricing pressure." Bloomberg's estimates point to slower growth in Digital Media. Still, some on Wall Street see progress: Evercore ISI's Kirk Materne called Adobe's updates a step toward addressing the "existential risk" question. KeyBanc's Jackson Ader noted strong customer reactions to the expanded model choices.
Adobe's response
Adobe is going model-agnostic. You can pick Firefly or external models depending on the job. Firefly has been positioned as safer for commercial use, with guardrails for copyright and content. It's already been used to produce more than 29 billion assets.
Financially, Adobe says AI brings in about $250 million directly. But when it counts pricing, retention, and upsells influenced by AI, it pegs "AI-influenced revenue" at roughly $5 billion. CTO Ely Greenfield says many teams keep Firefly for paid client work and use other models for brainstorming or experiments.
What this means for creatives
The stack is shifting from "one tool for everything" to "one workspace that routes to the right model." That's good news if you care about speed and output quality. It also means your edge comes from how you package your process, not which logo sits on the toolbar.
Practical playbook
- Use Firefly for client-facing, rights-sensitive deliverables. Keep a paper trail: prompts, settings, and model version.
- Ideate with non-Adobe models when you need range or motion. For video tests, try Sora; for concept visuals, compare at least two models before you commit.
- Build a prompt library tied to outcomes: ad concepting, storyboard frames, product comps, mood passes. Tag by model and style.
- Stress-test models against your niche (fashion, food, UI kits, product viz). Create a 10-shot benchmark and score outputs monthly.
- Quote for value, not hours. Price packages by deliverable and revision caps. AI shortens production; your taste and direction are the product.
- Protect the handoff. Keep editable files, layer structure, and seed settings for continuity across campaigns.
- Watch license terms. Model sources and training data still matter for brand work. If in doubt, default to Firefly or clear it with legal.
The bigger picture
Adobe's CEO says the suite remains essential for pro results. That's plausible-and still, the center of gravity is shifting to whatever model does the job best today. If Adobe stays the switching layer, it wins. If creators move full-time to external tools, pressure grows.
Your best hedge: get fluent across models, keep your workflow modular, and sell outcomes clients can't get from a prompt alone.
Helpful links
Your membership also unlocks:
