DigitalOcean's latest Currents report, based on a survey of over 1,100 developers, CTOs, and founders, reveals accelerating adoption of agentic AI and a shift toward inference-heavy budgets. While leading organizations are realizing productivity gains and new capabilities from AI agents, many remain in early exploration phases, highlighting a growing divide in AI maturity heading into 2026.
The Currents report shows clear momentum in AI adoption among digital native enterprises. Organizations have progressed significantly from experimentation to production use, with more than half now embedding AI deeply into operations. This evolution reflects growing confidence in AI's ability to solve real business problems and drive efficiency.
A major trend is the reallocation of AI spending priorities. Nearly half of respondents direct the bulk of their budgets toward inference rather than model training. This shift indicates that the next wave of AI innovation and value creation will center on deploying and running models at scale, rather than initial development.
Most organizations avoid single-vendor stacks, opting instead for multi-tool environments that combine various models, data sources, and infrastructure. While this approach offers flexibility, it introduces significant hurdles. Pricing and ease of use rank as top selection criteria, yet common pain points include managing separate APIs, forecasting costs, orchestrating deployments, and securing distributed tools.
The adoption of AI agents is delivering tangible benefits for early implementers. More than half of users report meaningful time savings for teams, while nearly as many unlock entirely new business capabilities. These outcomes position agentic AI as a key driver of competitive advantage in operational efficiency and innovation.
Despite progress, fully autonomous agentic systems are rare. Only a small fraction of companies run agents without supervision in production environments. Human review of outputs continues to be widespread, and human-in-the-loop mechanisms serve as the primary safeguard, reflecting cautious but pragmatic approaches to risk management.
Looking forward, the report forecasts broader agentic AI adoption in 2026. A substantial portion of organizations that have not yet engaged with agents indicate plans to begin exploration or deployment next year. This anticipated wave underscores the widening gap: companies that delay risk falling further behind peers already capturing productivity and capability advantages.
Paddy Srinivasan, CEO of DigitalOcean, stated: “AI-native businesses today are being built on a foundation of inference and agentic AI, and those who are figuring out how to effectively integrate AI into their workflows are seeing real benefits. Respondents are in agreement that the real opportunity for AI lies in applications and agents, and modern businesses are in need of straightforward, comprehensive tools that pair traditional cloud services with AI infrastructure and platform tools.”
The findings emphasize that while agentic AI and inference are reshaping infrastructure needs and budgets, the path forward requires simplified, integrated solutions to overcome complexity and accelerate production readiness. Enterprises not yet planning agent workflows face increasing pressure to catch up as trends accelerate.
About Currents
Currents is DigitalOcean’s ongoing report on trends impacting growing AI-native and digital native enterprises around the world. The latest release builds on a February 2025 report, with updates on the emerging trend toward AI agents and inferencing. New findings indicate that finding the right AI solutions is dominating mindshare at companies, which are grappling with the complexity and cost of advancing the use of AI. The need for easy-to-use, end-to-end solutions that connect everything from agent creation to inference at scale is increasing.
About DigitalOcean
DigitalOcean is an inference cloud platform that helps AI and Digital Native Businesses build, run, and scale intelligent applications with speed, simplicity, and predictable economics. The platform combines production-ready GPU infrastructure, a full-stack cloud, model-first inference workflows, and an agentic experience layer to reduce operational complexity and accelerate time to production. More than 640,000 customers trust DigitalOcean to deliver the cloud and AI infrastructure they need to build and grow.