A new study on AI by GFT Technologies concludes that outdated infrastructure is one of the headaches for CIOs, as well as one of the biggest obstacles to the success of AI in companies. The vast majority of technology leaders say this situation has already forced their organizations to abandon AI initiatives.
The GFT survey, carried out by Wakefield Research among 945 CIOs and CTOs from 19 countries – between August 11 and 31, 2026 – in companies with at least $500 million in annual revenue (€433.2 million annually) shows that the potential of AI and the risks associated with it are increasing in parallel. Key topics addressed in the study include: how legacy systems are undermining AI initiatives, growing skepticism about the return on investment in AI, geopolitical and regulatory uncertainty that is leading companies to rethink their supplier strategy, growing distrust over how companies explain AI-related workforce decisions, and increasing personal risk for technology leaders.
“AI adoption is advancing rapidly, but our research shows that companies are increasingly recognizing the importance of the foundation on which it rests. For CIOs in many organizations, legacy infrastructure is becoming a real constraint: not only for innovation, but also for security and scalability. Bridging that gap between AI ambition and infrastructure readiness is critical to turning AI investment into sustainable business value,” says Marco Santos, Global CEO of GFT Technologies.
Main findings of the report
Legacy infrastructure is holding back AI
CIOs surveyed believe that outdated systems hinder AI progress and create serious security risks. 84% say that the limitations of their legacy systems have caused their organization to cancel an AI pilot or project. Likewise, 93% believe that failing to modernize before running AI on outdated infrastructure will end up triggering a security crisis that affects the entire organization.
“Companies are discovering that a future AI strategy cannot be built on an infrastructure of the past. Modernizing legacy systems is no longer just a matter of technological efficiency: it is a condition to be able to scale AI safely, comply with regulatory requirements and convert investment into real business results. In our relationship with companies in Spain we see that this need is gaining weight in their strategic decisions,” says Manuel Lavín, executive director of GFT Europe and CEO of GFT Spain.
Investment in AI is growing faster than the value it can generate
Technology leaders or CIOs are reflecting concerns about how spending on AI is outstripping the value it can actually bring, which could indicate an “AI bubble.” 89% are concerned that global investment in AI may be growing faster than the business value it can realistically generate.
Geopolitical uncertainty is reconfiguring supplier strategy
Recent disruptions to access to AI models are leading companies to fundamentally rethink how they access the AI capabilities they need. Although recent tensions in the US have brought these risks to the forefront, their implications go far beyond a single market.
99% say potential government restrictions on access to AI increase the importance of not relying on a single AI vendor, following the US government’s dispute with Anthropic over access to its Mythos and Fable models earlier this year.
In this context, 42% of respondents are now inclined to build AI infrastructure internally rather than acquiring it from external providers.
Skepticism about the discourse around the impact of AI on workforces
Technology leaders question the idea that AI alone explains recent workforce changes. 91% believe that some listed companies cite AI to justify changes to their workforce whose main objective would be to boost their share price.
Personal risk for technology leaders
As AI scales, company CIOs and CTOs are increasingly concerned that the consequences of potential failure may fall on them personally. 89% are concerned that making the wrong workforce decision when scaling AI could put their own job at risk. Only 20% say other members of their organization’s senior management and board of directors fully understand the security risks of running AI on legacy systems.
What the data shows
GFT analysis points to a growing gap in AI implementation. Although companies continue to accelerate AI adoption, many are scaling AI faster than their underlying infrastructure can support. Legacy systems lead to repeated project failures, reliance on a single vendor leaves AI roadmaps exposed to forces outside any company’s control, and technology leaders increasingly bear the personal and professional cost when things don’t go as planned.
Despite all this, adoption continues to accelerate. Companies that are getting real value from AI treat it as the result of a deliberate bottom-up investment, not an added layer on top of an infrastructure, governance, and workforce strategy designed for another era.
The report’s findings point to a clear priority for companies looking to close the AI deployment gap: building a solid foundation first is what allows the value and speed achieved later to be sustainable and scalable. That means modernizing legacy environments, strengthening governance, managing technology dependencies, and aligning AI adoption with broader business and workforce strategies.
Outdated technology compromises security and slows AI progress
“The pressure to show results in AI is real, but moving faster than the infrastructure itself only transfers risk. Building the foundation first is what makes the subsequent value and speed sustainable and scalable,” adds Santos.
Summary of findings from the GFT study
- The fear of an AI bubble is growing. Nearly 90% of respondents said they are concerned that global investment in AI may be growing faster than the business value it can actually deliver.
- Legacy systems expose companies to significant security risks and are forcing AI projects to be canceled. A vast majority of respondents (93%) believe that implementing AI into their organization’s legacy systems without modernizing them first will ultimately lead to an enterprise-wide security crisis. The majority (84%) said they have had to cancel at least one AI pilot or project due to the limitations of legacy systems.
- Geopolitical and regulatory uncertainty is holding back investment in AI. Respondents said geopolitical events have led them to limit the areas in which they deploy AI (50%), reduce planned investment in AI (34%), and cancel AI projects (28%). More than half (54%) said uncertainty over AI regulation has made them more cautious about implementing it.
- AI is playing a much more complex role in workforce changes than the headlines suggest. More than 90% of respondents believe that some publicly traded companies use AI to justify workforce changes whose primary goal is to boost their share price. More than half (51%) said they have hired employees specifically to review or correct AI-generated work, and 26% said they have reinstated employees they had previously fired.
- Technology leaders face new pressures in their jobs. Nearly 90% of respondents said they were concerned that as their organization expands its use of AI, making the wrong personnel decision could put their own job at risk. Only 20% said their organization’s other senior managers and board members fully understand the security risks of using AI in legacy systems.
