Remigiusz Kinas's Basal is a family of open-weight typed-decision models for Polish and English that run on your own server. Using a downtime event in OmniMES, we walk through the request, the cost matrix, deferral to a human and an honest comparison with Jev.

How to build a Unified Namespace on an MQTT broker and Sparkplug B without replacing your controllers: an ISA-95 topic hierarchy, birth and death certificates, data contracts, broker high availability, IT and OT segmentation and a 30-day plan to get started.

Since 11 September 2026, software manufacturers, MES vendors included, must send an early warning of an actively exploited vulnerability within 24 hours and a full notification within 72 hours to the CSIRT and ENISA. Who Article 14 of the CRA covers and what factories should ask of their vendor.

A walk through the OmniEnergy module in the order you actually use it: connecting a meter, configuring SEUs and EnPIs, comparing periods, scheduling and exporting the report to PDF. Along the way we explain why the system records who entered every manual value, leaves incomplete snapshots off the charts and refuses to compare periods of different length — because that is what decides whether your data survives an ISO 50001 audit.

TypeSafe has released Jev — a model that writes no text at all and instead returns typed decisions with calibrated probability. The cost of a single judgment drops below the price of a few tokens, and latency from seconds to milliseconds. We look at what this actually changes in an MES: where such a model belongs on the shop floor, where it must never be used, and how to wire it in without breaking your audit trail.

From 18 February 2027 the product passport becomes mandatory for industrial batteries above 2 kWh, EV and LMT batteries. A passport cannot be assembled from a spreadsheet after the fact — it needs product genealogy captured by the MES while the goods are being made. Here is which data has to be linked, which standards now govern it (EN 18219, EN 18220), and what a mid-sized factory can realistically finish in five months.

Multiprojekt (the maker of OmniMES) and Relpol S.A. are launching a joint offering: certified Relpol RMM energy meters combined with the OmniMES system and the OmniEnergy module compliant with ISO 50001. One hardware vendor, one software vendor, one invoice, one point of technical support — instead of three separate integration projects.

From 18 February 2027 every LMT battery, every industrial battery above 2 kWh and every EV battery placed on the EU market must have a battery passport (Article 77 of Regulation 2023/1542). Based on the European Commission guidance of August 2026 we explain which data points are mandatory from February, which are not yet filled in (carbon footprint, recycled content, due diligence), what the Battery Pass, Volvo and GBA pilots showed, and what data your MES should prepare.

Black Hat USA 2026 and DEF CON 34 brought three threads that matter to manufacturers: the inside story of the December 2025 attack on Poland's energy sector and a manufacturing company (lateral movement into OT through a private APN), weaknesses in industrial protocols and engineering software (CC-Link IE TSN, OPC UA, open62541, Siemens S7), and vulnerabilities in local AI runtimes. We cover what was actually presented and propose a 30-day action list.

The 2 August 2026 date did not start enforcement of high-risk AI obligations. The Digital Omnibus (Regulation (EU) 2026/1744) moved them to 2 December 2027 for Annex III systems and to 2 August 2028 for Annex I. What a factory running an MES must comply with today (the workplace emotion-recognition ban, transparency duties, GDPR and DPIAs for operator monitoring), who supervises the AI Act in Poland, and how to use the time until December 2027.

OmniMES exposes an MCP (Model Context Protocol) endpoint — a bridge through which any AI assistant talks directly to your production system. We walk through, step by step, how to go from an empty system to a finished dashboard using a fictional WOODCRAFT furniture plant as an example — with no UI and not a single line of SQL.

For 50 years an industrial robot learned a single task through teach-pendant trajectory programming. In 2024–2026 three open foundation models for robotic manipulation arrived — Pi-Zero from Physical Intelligence, OpenVLA from Stanford and GR00T N1.5 from NVIDIA — promising to end that era. The operator shows the robot what to do in VR/AR through 5–10 demonstrations, and a generic model translates that into a real-time trajectory. This article walks through how it actually works today, where it is enough, where it is not, and the real implications for MES integrators across the EU.

TimescaleDB handles time-series from the last 1–90 days (hot) excellently, but at 5+ years of history the cost of PostgreSQL SSD and RAM grows faster than the business value. Apache Iceberg (a table format for the data lake) plus DuckDB (a local SQL engine) deliver a cheaper warm tier (90–730 days) and cold tier (>2 years) on S3 or MinIO. A 3-tier architecture cuts storage cost by 80–95% while keeping data queryable. This article shows how to build it for MES, where the limits are, what it costs, and how to migrate from a monolithic TimescaleDB.

The CBAM definitive period started on 1 January 2026, and 1 August 2026 is the deadline for the first periodic report under the new regime. European companies in steel, aluminium, cement, fertilisers, hydrogen and electricity — whether importing into the EU or exporting to carbon-aware markets — need to start computing embedded emissions per tonne of product. This article walks through what data MES + EMS must produce, how to map it to PCF (ISO 14067), which tools actually work, what it costs and what the penalties are. Plus a six-week roadmap to the deadline.

OmniMES uses TimescaleDB — the PostgreSQL extension with hypertables, native compression and continuous aggregates — as its database for industrial sensor time-series. The architecture handles 200 million readings per day, delivers 95% storage compression and 200–800 ms aggregations. This article walks through the hypertables concept, the OmniMES stack architecture, continuous aggregates and concrete production numbers.

Jensen Huang's Computex 2026 keynote (2 June) ran three hours and focused on industry — robotics, AI factories, digital twins. AMD and Intel countered the same day. From an MES perspective, five announcements actually change the technology map for 2027: GR00T N1.5 (robot foundation model), Cosmos 2.0 (world model for synthetic training data), the new Omniverse Replicator, AMD Ryzen AI Max+ as a Jetson alternative, and Intel Gaudi 3 for edge inference. This piece walks through each from a real deployment-pipeline perspective, not marketing.

The Polish amendment to the National Cybersecurity Act (KSC2), transposing the EU NIS2 Directive, enters real enforcement in the second half of 2026. Factories producing chemicals, food, machinery, automotive, electronics or medical devices — if they employ more than 50 people — are classified as "important entities" and fall under the full obligations package. This article walks through the ten Article 21 NIS2 requirements from an MES perspective: which functions already produce compliance evidence, what is missing, how every external API (OpenAI, cloud LLM, SaaS MES) grows your audit surface, and what to do concretely in the remaining months of 2026.

Two years after the first pretrained transformers for time series — TimesFM from Google, Chronos-Bolt from Amazon, Moirai-MoE from Salesforce — we finally have real answers to the question of whether a custom XGBoost pipeline for failure prediction in MES can be replaced by a zero-shot foundation model. This article walks through it without hype: what these models were trained on, the quality they deliver on real compressor sensor data, latencies on Jetson Orin and on a GPU server, when zero-shot is enough, and when fine-tuning is unavoidable — and what this means for the classic six-month ML project in a factory.

On 18 February 2027 the battery passport becomes mandatory for EV batteries, industrial batteries above 2 kWh, and LMT batteries — the first concrete moment when the Digital Product Passport (DPP) turns from EU concept into a hard production requirement. For an MES, this means exposing roughly seventeen attributes per unit: material origin, carbon footprint, recycled content, batch ID, durability, cell health data. This article walks through ESPR (2024/1781) and the Battery Regulation (2023/1542) without fluff: which MES functions already produce the data, what is missing, how to wire it up architecturally with GS1 Digital Link, and what to do in the nine months left before the deadline.

After a year of GPT-4 and Claude pilots in manufacturing, the honest question comes back: do we really have to ship process data to the cloud to get an MES assistant? In 2025–2026 the answer is no. Phi-4 (14B, Microsoft, MIT) at 4-bit quantization fits in 8 GB of VRAM, sqlite-vec gives you vector search in a single file with no server, and a Jetson Orin NX/AGX delivers 100–275 TOPS on the shop floor. This article walks through the concrete architecture, token-per-second benchmarks, 3-year TCO vs the OpenAI API, and what this means for AI Act, NIS2 and plant-level IT operations.

On 2 August 2026, the EU AI Act high-risk rules become fully enforceable. A subset of MES functions — predictive maintenance tied to machine safety, operator performance monitoring, AI in medical-grade quality control, and AI-assisted workforce decisions — may be classified as high-risk, with fines up to €15M for breaches. This article maps concrete MES modules to Annex I and Annex III, lists the seven obligations of high-risk operators, and gives a 3-month action checklist for plant owners.

Model Context Protocol (MCP), released by Anthropic in November 2024 and adopted in 2025 by OpenAI, Google and Microsoft, has become the de facto standard for connecting LLM agents to external data and tools. For the MES world, it ends the era of "one integration per model" and starts an architecture in which a plant exposes its data once — and Claude, ChatGPT, Gemini, and home-grown agents all plug in. This article explains how MCP works, what to expose from an MES, how to design it safely (lethal trifecta, IT/OT zoning), and when MCP simply does not make sense.

One of the recurring questions on MES projects is: how do you build analytical dashboards on data that lives in MongoDB, without rewriting the backend to a SQL database? In OmniMES we use Redash for this — for two reasons. First, Redash can consume a REST API as a data source. Second, it has a Query Results mechanism that lets you write full SQL on the result of any earlier query. The outcome is a combination that proves to be a game changer for many MES deployments: full SQL power (JOIN, CTE, window functions) over a NoSQL backend, with zero changes in the production layer. This article walks through the setup step by step — from configuring Redash, through querying the OmniMES REST API, to a SQL query that computes a vibration trend and a ready dashboard.

Agentic AI is redefining the trajectory of Industry 4.0. Traditional systems have largely been reactive—analyzing data and supporting human decision-making. Today, a new paradigm is emerging: autonomous agents that independently make decisions, optimize processes, and adapt to changes in real time. In modern smart factories, these systems integrate with MES, IIoT, and ERP platforms to create self-optimizing environments. Machines no longer wait for instructions—they identify issues, anticipate disruptions, and act before production is affected.

The concept of Industry 4.0 is based on continuous monitoring of production processes and decision-making driven by real-time data. Without reliable information, it is impossible to implement predictive maintenance, energy optimization, automated planning, or meaningful OEE analysis. If data is delayed, incomplete, or inconsistent, digitalization becomes only superficial. IT systems may be in place, but they do not deliver real business value. That is why it is critical to build an architecture in which data is collected automatically, consistently, and centrally.

Implementing a Manufacturing Execution System (MES) rarely begins with a full-scale rollout across the entire production environment. In practice, the most effective approach is incremental—starting with a pilot implementation on a selected production line or area. This allows companies to validate the system in real operating conditions, collect data, and fine-tune the configuration before scaling across the entire machine park. More and more organizations choose a pilot not only for technical reasons but also for organizational ones. MES impacts the daily work of operators, production managers, maintenance teams, and planners. Therefore, implementation should be a controlled process rather than a one-time deployment.

In Industry 4.0 and 5.0 discussions, a common argument appears: AI models reach around 90% accuracy, and in industrial environments a 10% error can cost millions. The reasoning sounds compelling because it is framed numerically and linked to operational risk. However, this argument assumes something critical — the existence of a fully digitized factory. In reality, across many manufacturing plants in Poland and Europe — especially in the SME sector — that level of digital maturity simply does not exist.

Why Connect Older Machines to an MES System? Legacy machines are often a crucial part of the production fleet, yet their actual performance frequently remains unknown. Data is collected manually, reports are delayed or missing entirely, and operational decisions rely on experience rather than facts.

Digital twins are increasingly appearing in modern factory strategies. Process simulations, virtual production lines, scenario testing without the risk of stopping production – this sounds like the future of manufacturing. One of the most recognizable tools in this area is NVIDIA Omniverse. The problem begins when a digital twin is supposed to stop being a visualization and become a reflection of actual production. For this, data is needed. And this is where MES-class systems play a key role.

Nvidia and the data problem at a broader scale

For the last two decades, most digital innovation has focused on mass-market IT: e-commerce, social platforms, and office SaaS. But the largest untapped opportunities now lie in highly specialized, industrial domains — where processes are physical, regulated, and operationally critical. Manufacturing, energy, logistics, and infrastructure do not need more generic software. They need systems that understand how the real world of machines actually works. That is where the next decade of industrial innovation will be built.

System MES (Manufacturing Execution System) - If someone in 2025 still thinks of an MES system as "terminals at workstations and reports from the department," it's about as current as a fax machine in OT-IT integration. The Manufacturing Execution System has ceased being a shop floor application. MES has become the operational layer of truth between the world of automation (OT) and the business world (IT). MES systems in 2025 are operational platforms that determine whether a company thrives on availability, quality, lead time, and energy efficiency. The MES system has stopped being a cost—it has become a mechanism for steering competitiveness.

The role of a production manager is becoming more complex every year. Modern factories demand not only smooth operations and on-time execution, but also cost optimization, rapid reaction to deviations, and building a culture of continuous improvement. In such an environment, traditional Excel sheets, manual reports, or intuition-based decisions are no longer enough. This is where OmniMES proves its value — a system that gives the production manager tools to make informed, fast, and accurate decisions.

Every manufacturing company has activities that do not add value to the product yet consume time, resources, and employee energy. In the book Toyota Production System: Beyond Large-Scale Production (1978), Taiichi Ohno identified seven of the most common types of waste. In Lean Manufacturing philosophy, these activities are called Muda – meaning wastefulness or uselessness (Japanese: muda = useless, unnecessary). Japanese companies, led by Toyota, have been effectively eliminating Muda for decades, achieving high production efficiency and flexibility. Understanding the seven classic wastes of Muda helps identify where productivity may be leaking in your company — and how to fix it.

OmniMES is a modern Manufacturing Execution System (MES) designed on a microservice architecture. This model ensures scalability, flexibility, and independent development of individual modules — from machine communication to data analytics and production management.

In modern manufacturing plants, every second of machine operation matters. Even short, seemingly insignificant interruptions – so-called micro-downtimes – can generate substantial losses on a production line. Research (Aberdeen Research) shows that unplanned downtime can cost manufacturing companies from hundreds to even thousands of dollars per minute. This is one of the reasons why more and more enterprises implement MES (Manufacturing Execution Systems). They support maintenance, enable real-time machine performance analysis, and help minimize the negative impact of micro-downtimes.

Industry 5.0 is not just another stage of digital transformation – it's a fundamental shift in approach to manufacturing that places humans and sustainable development at the center of advanced technologies. In this new reality, analytical platforms like BigQuery AI become a key component of intelligent production systems, enabling the transformation of vast amounts of data into concrete business insights.

In the era of industrial digital transformation, the global human-machine interface (HMI) market reached USD 24.5 billion in 2024 and is projected to grow to USD 55.2 billion by 2033, with a compound annual growth rate (CAGR) of 9.7%. At the same time, the simulation software market expanded from USD 19.95 billion in 2024 to an expected USD 36.22 billion by 2030 at a CAGR of 10.4%. These dynamic trends reflect a fundamental shift in the approach to optimizing production resources.

The integration of the Sparkplug B protocol in MES systems enhances energy efficiency and reliability in production.

The digital transformation of manufacturing requires a strategic approach to managing production batches. The use of Manufacturing Execution Systems (MES) combined with Industry 5.0 technologies can reduce downtime by 50% and cut maintenance costs by 40%, according to the latest Deloitte research.

Modern industry is at the heart of digital transformation, where Manufacturing Execution Systems (MES) with real-time data analytics form a cornerstone of Industry 5.0. According to a 2023 McKinsey Global Institute report, companies that deploy advanced analytics in production achieve, on average, a 15–20% productivity increase and a 25% reduction in defects.

ROI: up to 15% efficiency gains and 10–20% better energy storage performance – this is the real power of AI in modern energy. In the era of Industry 5.0, where technology merges with a human-centric approach, the renewable energy sector is growing at an unprecedented pace – with 18.6 GW of additional solar capacity installed in the first nine months of 2024. Manufacturing Execution Systems (MES) enhanced with machine learning are becoming a cornerstone of this transformation, delivering both business and technical benefits.

In the age of Industry 5.0, where energy efficiency becomes a key differentiator, manufacturing enterprises face unprecedented challenges. According to the new EU Directive 2023/1791 on energy efficiency, member states must achieve an average annual energy savings rate of 1.49% between 2024–2030. This means that companies must radically rethink their approach to energy management. A crucial tool in this transformation is the optimization of the OEE (Overall Equipment Effectiveness) index through intelligent energy management in MES-class systems. In both Industry 4.0 and 5.0, OEE plays a key role in ROI planning, reliability, and production stability.

Integrate MES with IoT to enhance quality control and reduce production downtime.

Comparison of MQTT and Sparkplug B Communication Protocols in the Context of Industrial Internet of Things (IIoT). Learn about their advantages, differences, quality of service (QoS), and message retention. Discover which protocol better meets the requirements of your IoT project.