Adobe Recreated with AI: Is the $1 Trillion Software Industry Facing a Revolution?

AI coding could turn today's legacy software problems into tomorrow's market opportunities.

A recent software experiment may offer an early glimpse of a much larger industry shift. In October 2026, developer Brandon Thomas drew attention for releasing open-source, Rust-based applications designed to reproduce many workflows of Adobe tools, including Photoshop, Illustrator and Premiere Pro. He said he used Claude-assisted coding. These are early projects, not proven replacements for Adobe. But the significance goes beyond their current quality.

For me, the key question is not whether one person can replicate parts of Adobe. It is what happens when thousands of teams can redesign established software faster, at lower cost, and around the needs of today's users.

The real opportunity: fixing what customers already dislike

Across the software industry, many products carry weaknesses that have accumulated over years or decades:

• Legacy architecture. Old codebases can be difficult and expensive to maintain.

• High prices. Licence fees, subscriptions and customization costs keep rising for some buyers.

• Poor user experience. Complex menus and too many screens slow down routine work.

• On-premises design. Some systems were built for in-house data centres, while customers want cloud options.

• Slow change. Upgrades, integrations and new features can take months or years.

These deficiencies create a business opening. A new entrant does not need to copy every button in a twenty-year-old application. It can serve the same business use case with modern architecture, clearer workflows, fewer features, and a better price.

Why I expect a major change by 2029

I believe the next one to two years will bring another meaningful jump in AI coding tools such as Claude and competing agentic systems. More of the development lifecycle could become AI-assisted: requirements analysis, architecture, coding, tests, documentation, integration and deployment. Skilled engineers will still need to direct and validate that work.

This is not pure speculation. Google's 2025 DORA study, based on nearly 5,000 technology professionals, found that about 90% used AI at work and more than 80% reported productivity improvements. It also warned that AI can increase software-delivery instability when testing and engineering discipline are weak.

That distinction matters. Writing code quickly is not the same as operating secure, reliable software. Yet better coding agents, combined with disciplined engineering, could substantially lower the cost of launching useful products.

A global wave of new competitors

India and China have large pools of software talent, engineering businesses and technology entrepreneurs. I expect founders and smaller teams there to explore AI-built alternatives to expensive commercial products. The same opportunity exists in Malaysia and many other markets.

Some teams may begin with open source. That can attract contributors, speed testing and lower adoption barriers. Revenue may come later from managed hosting, support, customization, integration or enterprise security. Open source is not a legal shortcut, however: independently written code must still respect intellectual property, contracts and applicable laws.

What changes in the software economy?

First, established vendors may face stronger pricing pressure in selected categories. Customers could have more credible alternatives to costly subscriptions. Second, software projects that were once uneconomic for smaller businesses may become affordable. Third, outsourcing firms may need to move from selling development effort to delivering tested business outcomes.

Jobs will change too. Routine coding effort may decline per project, while demand may grow for architecture, domain expertise, cybersecurity, testing, data engineering, product design and AI-agent supervision. This is a shift in the mix of skills, not proof that software engineers will disappear.

The most valuable startups may not be replicas. They may remove the need for familiar applications altogether. Instead of navigating ten screens, a user could state a goal. AI agents could prepare the work, handle routine steps and request approval at the right points.

Banking offers a useful illustration

Consider banking CRM, reporting, onboarding and operations workflows. Many institutions still run fragmented applications with repetitive data entry and costly integration layers. An AI-native product could combine better workflows, API-first connectivity, cloud deployment and role-specific assistants.

Core ledgers are different. Moving balances, transactions and controls is high risk. Regulation, resilience, security and migration history matter as much as coding speed. I expect disruption to start at the edges of enterprise systems, then move inward as new suppliers prove themselves.

My three-year scenario: 2027 to 2029

• 2027 — First wave. More focused, open-source and low-cost alternatives will emerge. Many will remain immature.

• 2028 — Better products. Stronger AI agents may help small teams deliver more complete, cloud-ready systems with improved quality.

• 2029 — Market pressure. In selected software segments, enterprise buyers may gain real leverage over price, functionality and deployment choices.

This is a scenario, not a forecast of guaranteed adoption. Established firms can use the same AI tools to modernize their products. Enterprises will still weigh migration costs, support, compliance and operational risk.

The bigger change is not cheaper code

My expectation is that 2026–2029 could mark an important turning point for software engineering and the software economy. Legacy architecture, high costs, difficult user interfaces and the shift to cloud create the demand. Better AI coding tools create the supply.

The companies that succeed may not be those that reproduce old software most faithfully. They may be those that solve the same problems more simply, securely and affordably.

The question for software leaders is simple:

If customers can get the same outcome through a simpler, cloud-ready and lower-cost product, how long can legacy complexity remain a competitive advantage?

Sources and further reading

[1] Creative Bloq, 8 Oct 2026: Adobe alternatives built with AI.

[2] ArtCraft / PhotoCraft project (source code and description).

[3] Google Cloud, 2025 DORA: AI-assisted software development.