Cloud-Native DCS

Our vision

Connect your AI to the plant floor

The industry is moving from a pyramid to an hourglass

Cloud-Native DCS runs deterministic batch sequencing at the edge below, and one standard API to the live plant serves the software, data, and AI above.

Request a demo
Industrial automation is shifting from a pyramid to an hourglass. An hourglass drawn against three tiers. It is pinched to a dashed outline at the proprietary PLC and DCS control layer in the middle and widest at the two ends, where smart field devices below and software, data, and AI above now hold the value. The top end holds cloud, compute, database, and AI icons; the pinched middle holds the proprietary layer's own PLC icon; the base holds machine vision, a smart sensor, an I/O module, a drive, and a robotic arm. When animation is available, the shape first appears as the traditional automation pyramid, widest at the smart-field base and narrowing up through the control layer, holding only a lone database at its narrow top, an I/O module and a drive at its base, and the PLC icon in its middle band. It then transforms into the hourglass as the other value icons arrive in the widened ends: cloud, compute, and AI joining the database above, and machine vision, smart sensors, and robotics joining the I/O module and drive below. The hourglass Software · data · AI Proprietary PLC / DCS Smart field devices The pyramid The hourglass Software · data · AI Proprietary PLC / DCS Smart field devices
The three tiers of the automation stack, with value gathering at the two ends while the proprietary middle commoditizes

Standard interfaces, up and down

An AI agent reads the plant model directly through one standard API

Every unit, every recipe, and every running batch is a typed API object. There is one copy of each, so a query from an AI agent and an edit from a plant engineer land on the same object.

A distributed control system (DCS) is the software that drives a batch process from raw materials to finished product. Cloud-Native DCS sits where a traditional DCS sits, between the enterprise software above and the field instrumentation below.

Where Cloud-Native DCS fits in the plant. A vertical stack of three tiers. At the top, in neutral, the enterprise systems: ERP, MES, LIMS, a data lake, and AI agents. In the middle, Cloud-Native DCS, drawn as a tinted region whose top and bottom boundaries are both emphasized in blue and given equal weight. The top boundary is labelled one standard API and the bottom one Modbus TCP and OPC UA. Inside the region the plant model, held as API objects, faces up, and the controllers, running IEC 61131-3 logic at the edge, face down. At the bottom, in neutral, the I/O and field: sensors, actuators, valves, and pumps. A double-headed arrow crosses each boundary, so traffic runs both ways at both. Enterprise systems ERP · MES · LIMS · data lake · AI agents one standard API Cloud-Native DCS Plant model as API objects Controllers IEC 61131-3, at the edge Modbus TCP · OPC UA I/O and field sensors · actuators · valves · pumps

The pilot that travels

Data access is the bottleneck for industrial AI

Operations are moving toward AI-driven decision-making, and the largest automation vendors are racing to build for it. They keep meeting the same obstacle. The plant data an AI needs is scattered across proprietary systems, each reachable only through a custom integration.

Cloud-Native DCS removes the obstacle at the root. The plant is read through standard interfaces that modern data and analytics tooling already speaks.

One AI pilot travels plant to plant through the same open API. A proprietary system reaches each next plant only after its integration is rebuilt. Two lanes, each with the same AI and analytics tool and the same three plants. In the left lane, Cloud-Native DCS: the tool's trunk runs down the lane and each plant hangs off it as an identical short branch with a tap dot. The connector is labeled "credentials and a query," and the lane reads "The same tool reads every plant." In the right lane, proprietary systems: the tool's trunk reaches all three plants the same way, but each plant sits behind its own integration bead on the trunk before its branch, annotated "rebuilt from scratch." The lane reads "The integration is rebuilt for each plant." When animation is available, the connections draw themselves out from the tools. The left trunk docks all three plants in quick identical taps. The right trunk waits at every plant, the first included, while its integration flashes under construction and settles, and docks only then. Both lanes end fully connected, and the proprietary lane gets there only by building a new integration for every plant on the way. Cloud-Native DCS AI & analytics reads live state credentials and a query The same tool reads every plant Proprietary systems AI & analytics rebuilt from scratch integration integration integration The integration is rebuilt for each plant
An AI pilot proves out on one plant and travels to the next as credentials and a query, while a proprietary integration is rebuilt for each plant

Where it comes from

Cloud-Native DCS began with Dhananjay Bhaskar, an engineer who has worked both sides of the gap it closes. He spent eight years at one of the world's largest pharmaceutical manufacturers, first in automation engineering on DeltaV and PLC systems, then in platform engineering on Kubernetes. At the same company, provisioning a production control system took months. Standing up cloud infrastructure took minutes.

See what it does Request a demo