New Agil’EDA software targets chip-fab data demands, without risking costly tool downtime

As chipmakers push harder into AI-driven manufacturing, they’re demanding something many equipment suppliers still struggle to deliver: fast, structured, high-frequency machine data, without destabilizing the tools that print money by the minute.

Agileo Automation says it has an answer. On March 25, 2026, the company unveiled Agil’EDA, a new Equipment Data Acquisition (EDA) software package designed for OEMs that build semiconductor manufacturing equipment. The pitch is straightforward: meet top-tier fab connectivity requirements, especially in advanced packaging, while keeping real-time machine control rock-solid.

Why “SEMI EDA” is becoming non-negotiable for equipment makers

For decades, many fabs relied on SECS/GEM, the long-standing communications standard used to send commands and monitor equipment. But leading-edge fabs now want EDA alongside it, because modern manufacturing runs on richer data: tightly timestamped events, consistent measurement sets, and detailed traces that analytics systems can actually use.

That shift is showing up in procurement. OEMs increasingly face RFPs that don’t just ask whether a tool can connect, they specify the structure, speed, and usability of the data coming off the machine. In other words, “data-ready” is moving from a nice-to-have to a baseline requirement.

Agileo is positioning Agil’EDA as a plug-in building block OEMs can integrate into equipment controllers without ripping up their existing software stack. The company is explicitly aiming at “tier-one” fabs, industry shorthand for the biggest, most demanding chip manufacturers, and at advanced packaging sites, where process complexity and sensor counts keep climbing.

The core technical bet: split control traffic from data traffic

Agileo’s central claim is that Agil’EDA separates the control stream from the data stream, an idea baked into SEMI EDA, also known as “Interface A.” The goal is to prevent high-frequency data publishing from interfering with the equipment’s critical operations.

That matters because fabs are pulling more data than ever: chamber parameters, recipes, alarms, counters, state transitions, sensor readings, the list grows with every new node and packaging technique. If data collection competes for the same resources as real-time control, the tool can become less predictable, and that’s when tensions flare between automation engineers, IT teams, and production managers.

Agileo is selling insurance against that scenario: keep the control loop stable even when the fab is “yanking” data for near-real-time monitoring and analytics.

Freeze 2 and Freeze 3: the standards roadmap points to HTTP/2 and gRPC

Within the SEMI EDA ecosystem, “Freeze 2” and “Freeze 3” mark major milestones in how the standard evolves. Agileo says it’s working on client and server testing around Freeze 2 and is pointing customers toward what Freeze 3 is expected to bring.

Freeze 3 is associated with modern transport technologies like HTTP/2, gRPC, and protocol buffers, tools familiar to many software engineers, but still a leap for parts of industrial connectivity that historically leaned on older approaches such as SOAP/XML. The promise: lower latency and higher throughput, which becomes critical when fabs want to feed analytics pipelines continuously without bogging down equipment controllers.

The broader industry problem is consistency. Predictive maintenance and yield optimization depend on clean time-series data that’s comparable across tools. If every OEM exposes data differently, fabs pay the price in custom integration work. SEMI EDA is meant to reduce that friction, and OEMs that lag risk looking outdated in competitive bids.

How Agil’EDA fits into Agileo’s broader connectivity “stack”

Agileo isn’t launching Agil’EDA in isolation. The company is framing it as part of a larger suite built on its A²ECF-SEMI framework, alongside Agil’GEM and Agil’GEM300, products aimed at the “classic” SECS/GEM and GEM300 connectivity that’s still widely deployed across fabs.

The practical message to OEMs: you don’t have to throw out what already works. Keep SECS/GEM for certain command-and-control exchanges, and add an EDA channel for structured, high-frequency data collection. That hybrid approach may sound clean on paper, but it introduces real integration work, aligning identifiers, states, event consistency, and versioning across interfaces.

Agileo is also leaning on support as a selling point. SEMI standards can be unforgiving, and interoperability testing can make or break a qualification timeline. For an OEM, a delayed qualification can mean a tool that doesn’t get accepted into production, an expensive problem that dwarfs the cost of software.

CEO Marc Engel’s pitch: better yields, faster innovation, and AI that actually works

Marc Engel, Agileo Automation’s CEO, is tying the product to the industry’s bigger ambitions: higher yields, faster innovation cycles, and stronger competitiveness. The logic is familiar in semiconductor manufacturing, more reliable, actionable data helps engineers spot drift earlier, compare lots more effectively, and tune processes before problems turn into scrap.

Agileo is also explicitly linking EDA to AI and advanced analytics. The company’s argument is blunt: AI is only as good as the data it learns from. If equipment data is inconsistent or poorly structured, models degrade and alerts become noise. Standardized EDA-style data streams are meant to make analytics more scalable across an entire fab.

Still, the company acknowledges, implicitly through its emphasis on modeling and integration, that EDA isn’t a magic switch. OEMs and fabs still have to decide what signals to publish, how often, in what structure, and how to manage changes over time. The direction of travel, though, is clear: the most advanced fabs want structured, high-performance data as a default expectation, and equipment makers are being pushed to keep up.