Tech • AI • Robotics • Game
SaaS vs AI: the beginning of the end for Adobe!
8/10TL;DR
An open-source developer’s rapid, AI-assisted cloning of major Adobe apps has intensified fears that software interfaces are becoming easier to replicate, while the real competitive moat shifts toward data, workflow integration and measurable business outcomes.
KEY POINTS
An AI-built Adobe challenger gained immediate traction
On 7 October, Atlanta developer Brandon Thomas released Artcraft, an open-source suite designed to mimic Photoshop, Illustrator, Premiere, Lightroom, Acrobat, After Effects and InDesign. Built in Rust with help from Claude Opus 5, the tools are free, run locally and avoid subscription fees. Thomas said the suite is only at about 50% parity, but the Photoshop-like component, Photocraft, still drew about 43,000 GitHub stars within days.
Frustration with subscription fees helped spark the project
The trigger was dissatisfaction with Adobe’s pricing and cancellation practices. That issue has already had legal consequences: in March 2026, Adobe reached a $150 million agreement with US authorities over contested termination fees. The case reinforced a broader perception that expensive seat-based software is vulnerable when users can increasingly rebuild what they need.
The cloning method relied on behavior, not source code
The project used a clean room approach, reproducing how the software behaves without copying proprietary code. That is increasingly feasible because products like Photoshop are exhaustively documented through official manuals, tutorials and years of public analysis. With modern coding agents able to inspect screens, click through interfaces and iterate overnight, recreating familiar software patterns has become far cheaper and faster than before.
Markets are worried even as Adobe keeps growing
The anxiety is not reflected in Adobe’s operating results. The company recently reported quarterly revenue above $6 billion, nearing $7 billion, up 13%, with roughly 1 billion monthly active users. Yet its stock was said to be down about 37% for the year at one point in 2026, showing that investors are pricing in future pressure on margins and pricing power rather than current demand.
The biggest threat is not a clone, but AI becoming the interface
The core concern is that users may stop opening specialized apps at all. Instead of manually editing an image, they may describe the desired output to Claude, ChatGPT or similar systems and refine results conversationally. In that scenario, traditional software loses direct contact with the user, and the interface layer shifts to general-purpose AI assistants.
Seat-based SaaS models look increasingly exposed
The discussion highlighted a growing divide between software sold by seat and software sold by usage. Companies charging per user may struggle if AI reduces time spent in the app and compresses the number of active seats. Consumption-based businesses such as Snowflake and Datadog were cited as better aligned with this transition, while design and productivity vendors built around subscriptions per employee face sharper questions.
Image software is shifting from tools to outcomes
Photoroom illustrates that change. Rather than selling access to editing features, it increasingly positions itself around output and performance, especially in e-commerce imagery. The model is moving from selling seats to selling tokens, then further toward selling business results such as higher conversion or sales uplift, reflecting a broader shift from software as a workspace to software as a revenue engine.
Specialized data remains a defensible moat
General AI models can generate content, but they still struggle with domain-specific constraints such as product fidelity in commerce images, where the generated visual must accurately match the item sold. That gives vertical players an advantage if they own proprietary evaluation data, confidence scoring and workflows tied to measurable use cases. The competitive edge comes less from generic generation and more from trusted performance in narrow but valuable tasks.
Context and system-of-record control are becoming decisive
As AI assistants connect to email, calendars, drives, messaging tools and databases, they accumulate the context needed to automate work across an entire business. That could reduce the role of stand-alone tools such as CRM platforms for some teams, while increasing the value of software that can plug deeply into these AI environments. Vendors with reliable infrastructure, broad integrations and machine-readable documentation may still benefit if AI agents consistently recommend and use their services.
CONCLUSION
AI is making software interfaces easier to imitate, but it is not eliminating software value altogether. The winners are likely to be companies that control critical data, integrate into AI-driven workflows and charge for outcomes rather than for seats.
Ask a question