Edge Computing: Why the Cloud Isn’t Always Enough for SMEs
6 min. read
Edge computing moves data processing to where it is needed: to the machine, onto the factory floor, into the vehicle. Gartner predicts that by 2025, more than 75 percent of enterprise-generated data will be created outside traditional data centers. For Germany’s Mittelstand, with its latency-sensitive manufacturing processes, edge computing is not a luxury, but a necessity that the cloud alone cannot cover.
The key points at a glance
- 75 percent of data at the edge: Gartner predicts that the vast majority of enterprise data will be processed outside central data centers.
- USD 232 billion market by 2030: The global edge computing market is growing at a CAGR of 33 percent (Grand View Research).
- Latency below 10 ms: Manufacturing processes such as welding robotics, camera-based quality inspection and real-time PLC control require response times that cloud latency cannot deliver.
- 5G as an enabler: 5G campus networks provide the bandwidth and latency that edge applications need on the factory floor. Deutsche Telekom, Vodafone and private providers are rolling out campus networks.
- Data stays local: Edge computing enables on-site data processing and reduces data transfer to the cloud. This simplifies GDPR compliance and lowers transmission costs.
Why the cloud does not solve everything
Cloud computing has fundamentally changed IT infrastructure for the Mittelstand. Scalability, pay-per-use and the elimination of in-house server rooms are real advantages. But the cloud has a physical limit: latency. Data has to travel from the end device to the nearest cloud data center and back. Even with the best connections, that takes 20 to 50 milliseconds. For business applications, email, CRM and ERP, that is irrelevant. For real-time control in manufacturing, it is not.
A welding robot that uses a camera system to inspect seam quality in real time needs response times below 10 milliseconds. Optical quality control on an assembly line has to decide within fractions of a second whether a part is acceptable or should be rejected. An autonomous transport system in a logistics hall cannot wait 50 milliseconds for the cloud before taking evasive action. In these use cases, latency becomes the showstopper.
Edge computing solves this problem by bringing computing power to where the data is generated. An edge server on the factory floor processes the camera data from quality control locally. Only aggregated results, statistics and anomalies are uploaded to the cloud. This reduces latency to below 5 milliseconds and cuts data traffic to the cloud by 80 to 90 percent.
“The future of computing is hybrid. Edge and cloud are not competitors, they complement each other. The question is not where you compute, but what you compute where.“
Satya Nadella, CEO Microsoft (Ignite 2025)
5G Campus Networks: The Infrastructure for Edge
Edge computing in manufacturing needs a network infrastructure that guarantees bandwidth and latency on the factory floor. Wi-Fi quickly reaches its limits in industrial environments: interference from metal structures, limited device capacity, and no guaranteed latency. 5G campus networks solve these problems.
By allocating 3.7 to 3.8 GHz frequencies to industrial companies, Germany has created a foundation that is unique in Europe. More than 200 companies have acquired their own 5G campus licenses from the Federal Network Agency. Bosch operates campus networks in several plants, Siemens uses 5G in automotive manufacturing, and BASF is testing 5G-supported sensor technology at its chemical park.
For mid-sized companies, Telekom, Vodafone, and specialized providers such as Ericsson and Nokia offer turnkey campus solutions. A typical campus network for a production hall costs between 50,000 and 200,000 euros to set up, plus monthly operating costs. The investment pays off through reduced downtime, improved quality control, and lower cloud costs.
Use Cases: Where Edge Computing Creates Value
● Visual quality control: Camera systems on the assembly line that use AI to detect defects. On-site processing in under 5 ms. Scrap reduction of 20 to 35 percent.
● Predictive maintenance: Vibration, temperature, and current sensors on machines. The edge server detects anomalies in real time. Only alerts are sent to the cloud. Reduction of unplanned downtime by 30 percent.
● Autonomous logistics: Automated guided vehicles (AGVs) in warehouses and production areas. Collision avoidance and route optimization must be calculated locally.
● Augmented reality: Remote maintenance and training via AR glasses. Overlay computation must take place locally to prevent latency-induced nausea.
● Energy management: Real-time control of load distribution, solar feed-in, and battery storage systems. Delayed responses can cause grid instability and penalty charges.
Edge plus Cloud: The Hybrid Architecture
Edge computing does not replace the cloud. It complements it. The architecture follows a clear principle: time-critical processing at the edge, long-term storage and analysis in the cloud. The edge processes sensor data in real time, filters out noise, detects anomalies and triggers immediate actions. The cloud receives aggregated data for machine learning, trend analysis and company-wide dashboards.
AWS (Outposts, Wavelength), Microsoft (Azure Stack Edge) and Google (Distributed Cloud Edge) offer hybrid solutions that seamlessly connect edge and cloud. For SAP environments, SAP has introduced Edge Integration Cell, a solution that can run SAP processes locally at the edge.
For mid-sized companies, a step-by-step approach is recommended: start with a specific use case (for example, quality control), place an edge server on the production floor, gather experience and then scale. The platform decision should match the existing cloud provider in order to minimize integration effort.
Conclusion: Edge Computing Closes the Cloud’s Gap
The cloud has limits, and those limits become visible as soon as milliseconds count. Edge computing closes this gap, not as a competitor to the cloud, but as a necessary complement. This is especially relevant for Germany’s mid-sized companies, whose value creation lies in manufacturing. The combination of 5G campus networks, edge servers and a cloud backend enables factory control that, five years ago, was reserved for large corporations. The technology is here. The infrastructure is being rolled out. Getting started begins with one machine and one use case.
Frequently Asked Questions
What is edge computing, in simple terms?
Edge computing moves data processing from centralized cloud data centers closer to where the data is generated: at the machine, on the factory floor, or in the vehicle. Instead of sending all data to the cloud, it is processed locally. This reduces latency, saves bandwidth, and enables real-time responses.
Does edge computing replace the cloud?
No. Edge and cloud complement each other. Time-critical tasks are processed at the edge, while long-term storage, machine learning, and company-wide analytics remain in the cloud. The architecture is hybrid: the edge server filters and processes data in real time, while the cloud receives aggregated results.
How much does edge computing cost for SMEs?
An entry-level project with one edge server for a single use case (e.g. quality control) costs between 10,000 and 30,000 euros for hardware plus setup. A 5G campus network adds another 50,000 to 200,000 euros. Payback typically comes from reduced downtime and improved quality metrics.
Do I need 5G for edge computing?
Not necessarily. Edge computing also works via Ethernet, industrial Wi-Fi, or LTE. However, 5G campus networks offer the best combination of bandwidth, latency, and device capacity for industrial applications. For getting started, a wired connection to the edge server is often sufficient.
How secure is edge computing?
Edge computing can improve data security because sensitive production data is processed locally and does not have to travel over the internet to the cloud. At the same time, every edge location expands the attack surface. Physical security, encrypted communication, and regular updates for edge servers are mandatory.
Further Reading
- → ERP Cloud Migration: Replatforming as a Strategy – Why the cloud decision should include hybrid architecture (MyBusinessFuture)
- → Disaster Recovery in the Cloud – Edge as part of the resilience strategy (cloudmagazin)
- → Cybersecurity Trends 2026 – IoT and edge security as a critical trend (SecurityToday)
Title image source: Panumas Nikhomkhai / Pexels
