Zeiss spent €200M on a greenfield SAP migration. Then walked away.

Here's what actually happened, and what every SAP leader should take from it.

Zeiss had been on R/3 since the early 90s. By 2020, the board knew the scale of what they were dealing with: three decades of accumulated customisations, technical debt, and process clutter baked into a system long past its shelf life. ECC mainstream support ends in 2027, so the clock was ticking.

CIO Carsten Trapp saw it clearly: going greenfield was "a lifetime opportunity to clean up, because we have messed up our R/3 system over the last 30 years." A fresh start. SAP standard processes from day one. No baggage carried forward. The approach was deliberate, structured, and had board backing. They picked Azure as the platform, ran four pilots across very different business profiles, built a global template, and planned to roll it out to 150+ entities by 2030.

On paper, it was a textbook greenfield programme.

Then, in September 2026, Zeiss quietly announced they were ditching it. After six years and an estimated €200 million spent, they are now switching to a brownfield approach: migrate the existing landscape to S/4HANA, then build on top of it.

The official line is that the plan has been "realigned to achieve faster progress." What that tells you between the lines is this: the complexity of changing business processes and technology at the same time, across a company of Zeiss's scale and diversity, proved harder to sustain than the ambition that launched it.

This is not a Zeiss failure story. It is an industry-wide reality check.

Greenfield sounds clean. The idea of starting fresh, removing customisations, running SAP standard, it is genuinely compelling. But at enterprise scale, especially when you have multiple business segments with legitimately different process needs, a uniform greenfield template runs into the real world faster than anyone wants to admit. The data alone is a major factor: you cannot escape 30 years of legacy data quality issues just by choosing a new approach. That history travels with you, whatever migration path you take.

What Zeiss did well: they ran pilots, they had governance through a design authority, they built templates. The fundamentals were right. What slowed them down was the sheer scope of simultaneous business and technology transformation across a global, multi-division company.

The lesson is not that greenfield does not work. It is that greenfield without ruthless scope control, phased commitment, and clean data readiness from the start will always cost more than the plan said it would.

The 2027 deadline is not moving. And for organisations still sitting on R/3 or ECC, watching Zeiss pivot should feel like a warning, not just a news story.