HighByte Briefing Note
Axendia was briefed by John Harrington, Chief Product Officer, and Torey Penrod-Cambra, Chief Communications Officer at HighByte, for an update on the company’s Industrial DataOps platform, including data strategy trends, IT/OT convergence and integration, emerging interoperability standards, and HighByte’s evolving efforts to support AI adoption in manufacturing. Axendia has been covering these themes, including the rise of agentic AI in this rapidly shifting architectural landscape.
The Data Strategist Arrives – Finally
For decades, manufacturers organized their technology investments around applications: SCADA administrators, MES leads, ERP teams. The data itself was a byproduct, not a priority. That has been changing for quite some time, but the trajectory has steepened. “You rarely saw a data strategist role in manufacturing,” Harrington observed. “That’s a change that’s happening right now in real time.”
This shift matters because the complexity of industrial data environments far exceeds that of conventional IT. Manufacturing data is physical and distributed – sensors, PLCs, SCADA systems, historians, MES, ERP, PLM – spanning on-premises machinery that won’t likely migrate to the cloud and cloud platforms where enterprise analytics increasingly live. Telemetry data, if not captured in real time, disappears. Transactional data lives in SQL-based systems of record. Time-series data resides in historians. Files ranging from CAD drawings to equipment manuals to AI-generated inspection outputs move in all directions across the stack. “Getting this data to the right consumer, at the right time, with sufficient context, is the core challenge HighByte Intelligence Hub is built to solve,” states Harrington.
IT, OT, and Domain Expertise, The Third Leg of the Stool
A persistent theme is the structural challenge of bringing IT and OT teams together – and then adding the business domain experts who actually consume the data. Harrington was direct about the stakes: “I tell our sales team, if they don’t have both teams together in the sales process, we won’t be successful. At some point along the path, the project will fail.”
The ideal champion, he explained, is someone who has moved from OT engineering into an IT architect role – someone who understands both worlds and has been tasked by the CIO with making plant-floor data accessible within the organization’s cloud platform of choice. The harder scenario is when IT arrives without any plant-floor knowledge or relationship with operational teams.
Penrod-Cambra added that domain teams such as quality and maintenance often represent a third leg of the buying decision. These are the end consumers of the data who define what “useful” actually looks like. This three-way alignment between OT engineers, IT architects, and business-side data consumers is increasingly the defining factor in whether Industrial DataOps deployments succeed at scale. As Axendia has reported, this same dynamic plays out in bio/pharma manufacturing, where quality and regulatory functions often drive data requirements that most IT and OT teams don’t always understand.
From Unified Namespace to Data Control Plane
HighByte Intelligence Hub is a low-code to no-code Industrial DataOps platform that connects OT and IT data sources, standardizes and contextualizes data across pipelines, and delivers it to analytics, AI, and application layers. Harrington described that its four core capability areas – orchestration, quality, observability, and governance – span the full data lifecycle, from edge devices to cloud data lakes.
The team had previously framed HighByte’s role around the unified namespace, providing a common, standardized view of industrial data. The company’s current framing has evolved: with the rise of AI agents, the unified namespace is no longer sufficient on its own. What manufacturers now need is a data control plane.
“The number of consumers of industrial data is going to dramatically increase by probably a factor of 100 over what they have been,” Harrington explained. “We expect to see thousands of agents within a factory – both running on-prem as well as in the cloud.” As agent proliferation accelerates, curating data delivery and managing which agents access which data sources becomes a critical infrastructure function, not just a data engineering task.
The platform supports streaming, transactional, and historical data movement, bulk transfer to cloud data lakes via Parquet files, and on-demand access via REST APIs. In the summer of 2025, HighByte extended its API layer to support MCP (Model Context Protocol) tools, enabling large language models and AI agents to query factory-floor data through standardized, discoverable interfaces – without requiring bulk data migration to the cloud.
An Emerging Standard: I3X
One of the more forward-looking developments discussed was HighByte’s involvement with CESMII (the Clean Energy Smart Manufacturing Innovation Institute) and its recently announced i3X protocol. Announced in beta, with a production-grade release expected this summer, i3X is a REST-based industrial data interoperability standard designed to address the fragmentation across application-layer interfaces.
Today, every application in the industrial stack has its own SDK, its own payload format, and its own connection protocol. HighByte is a founding contributor to the i3X working group, and Harrington sees it as a necessary step toward reducing that integration overhead. “Manufacturers are looking for standards, and the smaller vendors are looking for standards,” he noted. “We have seen that the more established vendors aren’t always as invested. HighByte prides itself on being good at tearing down moats and exposing data.”
Putting Industrial DataOps to Work
Two publicly disclosed engagements illustrate how the Intelligence Hub translates industrial DataOps into measurable outcomes.
At Alcon, a global eye care company, the challenge was scale: a contact lens factory with hundreds of small work cells, each requiring data collection, and a projected timeline of a full year to deploy a single predictive maintenance use case. Using the Intelligence Hub alongside Amazon S3, Amazon Timestream, and Seeq, Alcon compressed that timeline to one month.
With the integration infrastructure already in place, a subsequent use case, AI-assisted defect monitoring for yield improvement, progressed even more quickly, from a projected three months to one week. As Alcon’s Data Analytics Manager John Patanian has stated publicly: “Rather than be limited to one or two projects per year, we leveraged HighByte Intelligence Hub to systematize and scale the data mapping process for new applications.”
Bayer Pharmaceutical faced a different problem: more than 15 years of validated process data locked in a legacy historian, accessible only through static dashboards on computers located in the cleanroom. HighByte connected the Intelligence Hub to Bayer’s Asset Framework, Event Frames, and raw archive data, exposing these data sets as MCP tools. Bayer’s internal AI assistant connects those tools to AWS Bedrock, giving engineers and operators conversational access to process data in seconds rather than hours, an aspiration that Penrod-Cambra emphasized the importance of: “More fluidity, with operators on the plant floor creating ad hoc requests and reports as needed.” Because HighByte serves as the abstraction layer, the same MCP tools and agent can extend to sites running different historians without rebuilding the integration from scratch.
The CMO/CDMO data access challenge is a related open question: regulatory traceability requirements apply regardless of whether a manufacturer owns the facility. HighByte is working with leading CDMOs on outbound data publishing approaches, with Snowflake’s cross-organizational sharing as one viable mechanism. Closing the loop from insight to validated action, i.e. issuing a work order through a GxP-compliant CMMS based on a predictive finding, remains an area to watch.
In Brief
The shift HighByte describes – from unified namespace to data control plane – reflects a real change in what industrial AI demands. As agent proliferation accelerates, governing which datasets reach which consumer becomes as critical as moving the data in the first place. The Alcon and Bayer case studies demonstrate that when integration infrastructure is in place, the pace of new use case deployment changes dramatically. With life sciences organizations still working through heterogeneous OT environments and data access challenges, that is a force multiplier worth examining.
The CESMII i3X standard is early but worth watching. If it gains traction, it could address the application-layer fragmentation that has kept smart manufacturing use cases stuck at the pilot stage for too long. HighByte’s role as a founding contributor – and its multi-protocol fallback if standardization stalls – positions it well either way.
We will continue to provide updates on Highbyte as they become available.
Related Content:
- Is AI in Quality an Evolution or a Revolution?
- Moving Data with Purpose to Drive Value
- Industrial DataOps: The Backbone of AI-Ready Manufacturing Data
Stay Connected with Axendia!
Stay at the forefront of innovation and technology in life sciences, healthcare, and space-enabled research by connecting with Axendia.
- Follow us on LinkedIn: Join our professional network and stay updated on the latest insights, trends, and case studies. Connect with Axendia on LinkedIn
- Subscribe to Our Updates: Get exclusive insights and thought leadership articles delivered straight to your inbox. Sign up for our newsletter
- Share Your Thoughts: We’d love to hear from you! Contact us to discuss your ideas, challenges, or opportunities. Email Axendia
Let’s drive innovation together!
Ready to unlock new opportunities and drive success? Schedule an Analyst Inquiry
The opinions and analysis expressed in this post reflect the judgment of Axendia at the time of publication and are subject to change without notice. Information contained in this post is current as of publication date. Information cited is not warranted by Axendia but has been obtained through a valid research methodology. This post is not intended to endorse any company or product and should not be attributed as such.