A widening gap is emerging in engineering AI adoption, with just 9% of organisations achieving mature, scaled deployments while 80% remain stuck in pilot programmes, according to SimScale’s 2026 State of Engineering Report.
The survey of 350 engineering leaders in large enterprises highlights a growing phenomenon dubbed “pilot purgatory”, where experimentation surges but enterprise-wide implementation stalls. While the number of companies running AI pilots rose sharply—up 62% year-on-year—the proportion transitioning to scaled deployments increased by only 29%, underscoring a critical execution bottleneck.
Despite this lag, momentum is building. About 36% of design and simulation teams are already integrating AI or agentic AI into workflows, signalling what SimScale describes as a tipping point for the industry. Organisations failing to advance beyond pilots risk falling behind faster-moving competitors.
“Moving from AI pilots to scaled deployment is still one of the most critical challenges for engineering organisations right now,” said David Heiny, co-founder and CEO of SimScale. “The fastest moving teams are transitioning in as little as three months, but many are lagging behind, creating another significant gap.”
The average transition from pilot to full-scale deployment stands at eight months, though more than half of respondents report timelines of up to a year or longer. In contrast, leading organisations are compressing this cycle to between three and six months, gaining a clear competitive edge.
Three primary barriers to scaling AI were identified: data preparation and availability (74%), governance and compliance concerns (48%), and software interoperability challenges (42%).
However, SimScale suggests these obstacles may be overstated. Many AI-driven engineering applications can deliver value with less data than anticipated, while governance concerns are easing—87% of organisations now permit AI-driven pass-or-fail decisions at design gates.
Early adopters are already demonstrating measurable gains. Clean technology firm Convion has deployed Physics AI models to enable real-time design exploration, while RLE International has developed an end-to-end AI workflow capable of predicting vehicle aerodynamics in seconds.
The performance improvements are significant. Tasks that traditionally required up to 17 hours can now be completed in six, while simulation iterations have accelerated from weekly cycles to multiple runs per day. Design exploration has also expanded dramatically, with teams testing hundreds of variants instead of just a handful.
Confidence in AI’s value is near universal, with 99% of respondents expecting meaningful business impact within the next 12 months.
“The shift is from human-paced to machine-paced iteration,” Heiny added. “If companies are not embedding AI into engineering workflows now, they risk being left behind.”
