# Manufacturers cite data labeling and edge compute as AI challenges

By MarketScale Newsroom · Published 2026-09-12 · Industrial IoT on MarketScale
Canonical: https://www.marketscale.com/industries/industrial-iot/factory-ai-budgets-are-shifting-from-models-to-tag-mapping-and-edge-compute

> Recent industry Q&As and interviews highlight persistent data-mapping and hardware issues facing manufacturers deploying AI on the plant floor.

## Key points

- Manufacturing plant data collected over decades from different vendors' equipment creates inconsistent labeling, siloed systems, and poor documentation that make it hard to use for AI projects.
- Timing mismatches between real-time sensor streams and quality results that arrive hours or days later pose a second, separate obstacle to using production data for AI.
- Industry 5.0 encourages manufacturers to consider resilience across suppliers, warehouses, logistics, and production rather than isolated automation projects, and expanding IIoT monitoring can involve coordinating data across multiple systems and owners.

A July 21, 2026 video interview published by Assembly Magazine features Sight Machine co-founder and CEO Jon Sobel discussing why manufacturing data remains difficult to use for artificial intelligence projects. According to Assembly Magazine, Sobel said plant data is often collected over decades from equipment made by different vendors, resulting in inconsistent labeling, siloed systems, and limited documentation of what individual data points represent.

## Why manufacturing data creates AI obstacles

Assembly Magazine reported that Sobel described a second, separate problem: timing mismatches between process data and quality results. Sensor data can stream in real time, while inspection or defect findings may not arrive for hours, days, or longer, according to the outlet. Sobel said that combining large volumes of sensor readings into a data lake does not produce useful insight unless the meaning of the data and its relationships are understood, Assembly Magazine reported.

Sobel's comments describe a challenge specific to Sight Machine's area of business, industrial data platforms, and were made in the context of a company-produced interview.

## IIoT scope and edge computing draw separate attention

A July 2026 Q&A published by ManufacturingTomorrow with Ilan Gluck, EVP and head of North America go-to-market at Digital Matter, described Industry 5.0 as encouraging manufacturers to consider resilience across suppliers, warehouses, logistics, and production, rather than isolated automation projects. Gluck's comments, as published by ManufacturingTomorrow, note that expanding IIoT monitoring into inbound materials and logistics can involve coordinating data across multiple systems and owners.

Separately, an August 2026 ManufacturingTomorrow Q&A with OnLogic, tied to the company's IMTS 2026 exhibit plans, said OnLogic is focused on providing industrial computing hardware intended to support artificial intelligence, advanced control, and real-time data collection at the production line. The Q&A is promotional content tied to OnLogic's trade-show presence and describes the company's own positioning rather than independently verified industry data.

A February 11, 2026 Automation World video segment featuring Austin Levin, lead automation engineer at system integrator ACS, discussed ways automation can be applied to address manufacturing workforce shortages while addressing safety and job satisfaction, according to Automation World.

## A related industry debate over AI labeling

An August 2026 ManufacturingTomorrow commentary by Tara Buchler, principal at JBF Consulting, argued that some capabilities marketed as AI on the plant floor are longstanding rules-based automation relabeled, and that manufacturers should ask vendors what specifically is new about a given method before treating it as a novel capability. This is Buchler's stated analysis and recommendation, published as commentary rather than reported news.

Taken together, these separate pieces describe distinct issues, data labeling and context, IIoT scope, edge hardware, and vendor claims about AI, that manufacturers have discussed in industry publications through 2026. None of the sources reviewed reports data showing a shift in aggregate capital spending from AI models toward data mapping or edge compute.

## Sources

- [Why manufacturing data breaks AI: Jon Sobel explains the challenge](https://www.assemblymag.com/articles/100237-video-why-manufacturing-data-breaks-ai-jon-sobel-explains-the-challenge) (Assembly Magazine)
- [Automation and the industrial workforce](https://www.automationworld.com/factory/workforce/video/55356853/automation-and-the-industrial-workforce) (Automation World)
- [Onlogic at IMTS 2026](https://www.manufacturingtomorrow.com/article/2026/08/onlogic-at-imts-2026/28063) (ManufacturingTomorrow)
- [How Industry 5.0 is changing what manufacturers expect from IoT on the factory floor](https://www.manufacturingtomorrow.com/article/2026/07/how-industry-50-is-changing-what-manufacturers-expect-from-iot-on-the-factory-floor/27920) (ManufacturingTomorrow)
- [How much of today’s AI on the plant floor is just yesterday’s software wearing a new label](https://www.manufacturingtomorrow.com/story/2026/08/how-much-of-todays-ai-on-the-plant-floor-is-just-yesterdays-software-wearing-a-new-label/28076/) (ManufacturingTomorrow)

Tags: manufacturing AI, industrial data, edge computing, IIoT, Sight Machine

---
Source: MarketScale, https://www.marketscale.com/industries/industrial-iot/factory-ai-budgets-are-shifting-from-models-to-tag-mapping-and-edge-compute. Published for AI indexing and citation; cite the canonical URL. Site guide for agents: https://www.marketscale.com/llms.txt