These days, AI is everywhere you look — but what numbers chart this explosive growth? The Wiz Research team’s new report examines exactly those data points, detailing key findings on the use of AI services and tools in cloud environments and painting a concrete, data-based picture of just how omnipresent this technology has become in the cloud. 

Through our analysis of aggregate data related to a large sample of organizations, spanning hundreds of thousands of cloud accounts across the major public cloud providers, we provide a comprehensive overview of how generative AI and machine learning are being used in the cloud and the implications this has for organizations. 

Key Findings 

AI is taking over the cloud 

• Our research shows that AI is rapidly gaining ground in cloud environments, with over 70% of organizations now using managed AI services. At that percentage, the adoption of AI technology rivals the popularity of managed Kubernetes services, which we see in over 80% of organizations! 

• Similarly, self-hosted AI SDKs and tools can currently be found in 69% of cloud environments, with 42% of organizations also choosing to self-host AI models. 

Azure OpenAI leads the way 

• Microsoft's Azure AI Services, which includes Azure OpenAI, takes the lead among managed services offered by the major Cloud Service Providers (CSPs) in total AI deployment, with 39% of organizations making use of this service.  

• Organizations have recently more than tripled their use of Azure OpenAI. Throughout a 4-month period in 2023, the number of deployed Azure OpenAI instances rose by a remarkable 228%, highlighting its rapid adoption. 

• Similarly, the OpenAI and Azure OpenAI SDKs can be found in 53% of cloud environments, likely indicating widespread development relying on these services. 

Many organizations are experimenting with AI, but only 10% are “power users” 

• While the adoption of AI in the cloud is soaring, many organizations (32%) still appear to be in the experimentation phase with these tools, deploying fewer than 10 instances of AI services in their cloud environments. 

• A few organizations (10%) seem to have already moved beyond experimentation, deploying 50 or more instances in their environments. 

• The contrast between the high prevalence of these services among organizations and the relatively low number of deployed instances might be explained by prohibitive costs and quotas currently enforced by cloud service providers. As costs go down and other limitations are eventually removed, we expect the number of instances to grow accordingly. 

The Future of AI in the Cloud 

• Our data shows that both managed and self-hosted AI services and tools are on the rise among cloud customers, as businesses across industries use AI in an attempt to enhance efficiency, automate processes, enable new product features and gain a competitive edge. 

•  But in some ways, this rapid uptake of AI technology mirrors the early stages of cloud computing — when many organizations embraced new technology without creating proper guardrails for its use. Wiz recommends improving visibility into AI service usage and fostering a culture of security ownership across development, operations, and security teams. This practice is often associated with the Shift Left concept.

Adopting AI Securely 

We believe that 2024 will be the year that clear AI success strategies emerge, with certain products and features winning the market, while others fall away. Organizations will decide which experimentation paths will prove successful and what products and features they will pursue. To navigate the journey of implementing AI tools in the cloud, Wiz recommends the following steps: 

• Build visibility into your usage of AI services: gain insight into AI service and product usage within your organization to eliminate blind spots (sometimes referred to as "Shadow AI"). Observability will help you understand which teams are using AI in your organization and what services and tools they’re working with, so you can confirm that both experimentation and development are conducted with the most secure specifications. 

• When building multi-tenant services that use generative AI models as part of their operation, follow our guide to ensure your customers’ data stays safe as well. 

• Foster a culture of ownership: in the new cloud operating model, security teams must collaborate closely with developers, cloud engineers, data scientists, and AI practitioners to manage the evolving attack surface introduced by AI tools. Democratizing security practices across the organization is always a best practice; in the face of AI adoption, this recommendation is more important than ever. 

Click below to gain more valuable insights about the incredible rise of AI in the cloud.