Snowflake's CoCo platform, launched as an agentic control plane, is now running enterprise AI workflows for Fanatics, Thomson Reuters, and WHOOP.1 The platform gives builders a unified, governed environment to manage workflows across data, models, and applications simultaneously.1
The shift is architectural, not incremental. "Your existing tech stack was designed for human-operated, application-centric workflows," said Surojit Chatterjee. "It needs to be reconsidered when the actor is an AI agent operating at machine speed across multiple systems simultaneously."2
Prasun Shah puts the competitive stakes plainly: AI agents derive their value as connective tissue moving across technology layers, not as another layer within them. "That is where the next battleground will be,"3 he said.
EXL Service Holdings is positioning itself directly in that battleground. The analytics and AI firm generated nearly $300 million in free cash flow in 2025 and is redirecting capital allocation toward agentic platform development.4 EXL's investor day highlighted agentic AI as the core growth vector, reflecting broader industry convergence around autonomous process execution.
The physical infrastructure underpinning this transformation is scaling rapidly. NVIDIA GPU acceleration and Dell Exascale deployments are providing the compute substrate for enterprise agentic systems. Chinese competitors including Zhenwu's J900 and V900 chips are entering the same race, adding hardware supply competition to a market already under pressure.
Netskope's AgentSkope launch adds a security governance layer to the emerging stack — a signal that enterprises are moving past pilots and into compliance-grade deployments. The pattern across vendors is consistent: unified control planes, governed multi-system access, and workflow orchestration designed for non-human actors.
The organizational implications are direct. Traditional metrics, reporting structures, and headcount models built around human task execution face structural obsolescence as agentic systems take over cross-application coordination. Enterprises that delay rebuilding their operational architecture around machine-speed agents risk falling behind competitors who are already in production.
Sources:
1 Snowflake, finance.yahoo.com, June 2, 2026
2 Surojit Chatterjee, MIT Technology Review, May 26, 2026
3 Prasun Shah, MIT Technology Review, May 26, 2026
4 ExlService Holdings Investor Day, finance.yahoo.com, May 19, 2026
