A new survey has found that infrastructure complexity, rather than technological capability, has become the primary obstacle preventing enterprises from successfully scaling artificial intelligence (AI) initiatives.
DDN’s State of AI Infrastructure Report, conducted by Vanson Bourne in collaboration with Cognizant and Google Cloud, surveyed 600 U.S. IT and business leaders. The findings paint a picture of organizations eager to deploy AI but struggling with the foundational systems required to support it.
According to the report, 65% of organizations describe their AI environments as too complex to manage effectively, and more than half of the respondents (54%) have delayed or canceled AI initiatives within the past two years.
“The AI boom has hit an infrastructure wall,” said Alex Bouzari, CEO and co-founder at DDN. “Companies are chasing models and GPUs, but the real bottleneck is the data layer underneath.”
The survey identifies infrastructure fragmentation as the root cause of this complexity. Organizations typically deploy AI workloads across disconnected systems for data processing, training compute, and serving endpoints, none of which are originally designed for generative AI’s demanding requirements.
Fragmentation forces continuous data movement between systems, requires manual resource orchestration across separate silos, and prevents the unified scaling of compute, storage, and networking resources.
With AI workloads projected to grow 110% over the next year, 76% of leaders report facing fundamental data challenges stemming from legacy infrastructure and siloed datasets.
“Enterprises are discovering that scaling AI isn’t a compute problem — it’s an integration problem,” DDN Chief Technology Officer Sven Oehme said. “If your infrastructure isn’t unified, your AI can’t learn efficiently.”
Nearly all respondents (97%) indicated that cloud infrastructure is essential to scaling AI, with more than half citing it as their fastest path to production deployment.
Cloud-based deployments allow teams to experiment with new approaches, onboard GPUs more quickly, and adopt emerging technologies faster while reducing early-stage failure rates.
“The survey responses validate the investments we’ve made over the years to develop a robust cloud infrastructure that empowers organizations to easily scale AI workloads,” said Asad Khan, senior director of product management at Google Cloud.
