Fortanix® Inc., a global leader in data and AI security and a pioneer of Confidential Computing, today released the results of a new survey examining the current state of enterprise AI deployment and how regulated enterprises are balancing the security and compliance challenges of deploying AI. Among the findings: 93% of AI leaders understand that traditional encryption does not protect data while it is being processed by AI, and recognize security and compliance concerns, even as organizations send sensitive data to AI services they are concerned may expose proprietary information.
Fortanix released a survey of 300 AI leaders across banking, financial services, healthcare, insurance and government.
93% of AI leaders understand traditional encryption does not protect data while being processed by AI.
61% of organizations still send sensitive data to public cloud AI services.
59% have delayed an AI project in the past year due to data security or privacy concerns.
64% of organizations have deployed closed-weight frontier AI models on-premises.
95% say sovereign AI is important or critically important to their AI strategy.
61% of organizations still send sensitive data to public cloud AI services, even though 93% say they're concerned about exposing proprietary prompts, models or intellectual property to AI infrastructure providers.
That tension is having real consequences: 59% of organizations have delayed an AI project in the past year due to data security or privacy concerns, and 41% have completely blocked a deployment over security. Regulatory compliance now ranks as the single biggest obstacle to deploying AI on sensitive data, cited by 46% of respondents, ahead of data exposure and exfiltration risk.
The survey found that many organizations are moving forward despite the above concerns: 61% of organizations still send sensitive data to public cloud AI services, even though 93% say they're concerned about exposing proprietary prompts, models or intellectual property to AI infrastructure providers. Nearly 40% described themselves as “very concerned.”
“It's clear that AI leaders in regulated industries have become educated about the fact that their data is at risk when a model starts processing it,” said Anand Kashyap, CEO and co-founder of Fortanix. “What they need is proof that the environments their models and data are processed in are secure and untampered. Confidential Computing gives organizations the proof and cryptographic assurance that their data and models stay protected while AI is running.”
Conducted by Atomik Research among 300 AI leaders across banking, financial services, healthcare, insurance and government, the survey suggests that enterprises are increasingly looking to change where and how AI is deployed rather than simply accept the risks of public cloud AI. 64% of organizations have deployed closed-weight frontier AI models on-premises, while 97% are at least somewhat likely to adopt leading AI models if they're securely available on-premises.
Sovereignty is also becoming a strategic consideration: 95% say sovereign AI is important or critically important to their AI strategy. AI leaders expect hybrid environments to be their primary inference location over the next 24 months, ahead of hyperscaler cloud, fully on-premises deployment and neoclouds or AI factories combined.
Confidence in how AI inference is secured remains limited. Just 44% of AI leaders describe themselves as “very confident” that data is protected during inference, while 52% are only “somewhat confident.” When asked what would increase their confidence in deploying AI on sensitive data, advanced infrastructure security like Confidential Computing came out on top, cited by 38% of respondents, more than any other single answer.
93% say they're concerned about exposing proprietary prompts, models or intellectual property to AI infrastructure providers.
61% of organizations send sensitive data to public cloud AI services.
64% of organizations have deployed closed-weight frontier AI models on-premises.
97% are at least somewhat likely to adopt leading AI models if they're securely available on-premises.
95% say sovereign AI is important or critically important to their AI strategy.
AI leaders expect hybrid environments to be their primary inference location over the next 24 months, ahead of hyperscaler cloud, fully on-premises deployment and neoclouds or AI factories combined.
About Fortanix
Fortanix is the global leader in data and AI security and a pioneer of Confidential Computing, delivering a unified platform to protect sensitive data, AI models, and applications across on-premises and multi-cloud environments—at rest, in transit, and in use. Built on hardware-enforced security, Fortanix enables workloads to run in tamper-proof, isolated enclaves, protecting against data leakage, model extraction, and unauthorized access, even from privileged insiders.
