New GFT research: 84% of CIOs have canceled an AI project over legacy system limits 

Today, AI-centric global digital transformation company GFT Technologies released a new study on how 945 CIOs and CTOs from organizations generating at least $500 million USD in revenue across 19 countries, see the state and future of AI. 

Researchers found that one of the biggest obstacles to enterprise AI success is legacy infrastructure, with the vast majority of executives reporting it has already forced them to abandon AI initiatives outright. 

The survey was conducted by Wakefield Research and is based on responses fielded from August 11 through 31 this year. It further revealed that AI’s promise and risks are accelerating in tandem. 

“For many organizations, legacy infrastructure is becoming a real constraint—not only on innovation, but also on security and scalability. Closing that gap between AI ambition and infrastructure readiness is critical to turning AI investment into sustainable business value,” he added. 

Key findings from the report are organized around five distinct pillars: the impacts of legacy infrastructure, AI investments outpacing realistic value, the influence of geopolitical uncertainty, skepticism over AI-driven workforce narratives, and perceived personal risk for tech leaders. 

Legacy systems and vendor risk are stalling AI plans 

When it comes to legacy infrastructure, 84% of respondents said limitations have caused their organizations to cancel an AI pilot or project, while 93% believe failing to modernize before running AI on old infrastructure will eventually trigger an enterprise-wide security crisis. 

Only 20% said their fellow C-suite executives and board members fully understand the security risks of running AI on legacy systems. 

Meanwhile, tech leaders also expressed concern that AI spending might be surpassing the value it can actually deliver, with 89% worried that global AI investment is growing faster than the business value it can realistically produce, a trend GFT says could point to an “AI bubble.” 

Geopolitical uncertainty is also reshaping vendor strategy, with recent disruptions to AI model access prompting a fundamental rethink of how enterprises source capability. Nearly all respondents, 99%, said potential government restrictions on AI access make it more important not to depend on a single AI provider, which GFT links to the suspension of access to Anthropic’s Mythos and Fable models under U.S. export controls earlier this year. 

In response, 42% are now leaning toward building AI infrastructure internally rather than buying from outside vendors. 

Many enterprises were already hedging their bets. Venture capital firm Andreessen Horowitz’s latest CIO survey found that 81% of large companies use three or more model families in testing or production, up from 68% less than a year earlier.

While tensions in the U.S. have brought these risks into sharp focus, their implications extend well beyond a single market. 

U.S. tech leaders feel the workforce pressure most

On workforce narratives, 91% of respondents believe some public companies cite AI to justify workforce changes that are primarily intended to boost their share price, and tech leaders also see growing personal exposure, with 89% concerned that a wrong workforce decision made while scaling AI could put their own job at risk.

U.S. technology leaders in particular feel the pressure most acutely. Some 93.3% of the country’s CIOs and CTOs believe companies use AI as cover for share-price-driven workforce changes, ahead of the 90.6% global figure. 

Respondents from the country were also the most concerned of any region about AI investment outpacing value, at 92.1% compared with 80.6% in Europe, the Middle East and Africa, and 47.0% said they were very or extremely concerned about their own job security, above the 42.9% global figure. 

“U.S. technology leaders are carrying more pressure than most, over whether AI is delivering real value, workforce trust, and their own personal exposure if something goes wrong,” said Rishi Chohan, CEO of GFT USA.

Per GFT, the findings point to a widening AI implementation gap, with many enterprises scaling AI faster than the foundation underneath it can support. The company recommends preparing that foundation first by modernizing legacy environments, strengthening governance, managing technology dependencies, and aligning AI adoption with broader business and workforce strategies.

Featured image: Shamin Haky via Unsplash+

Disclosure: This article mentions a client of an Espacio portfolio company.