On August 4, 2026, Xu Hui, Director of Solutions at Glorysoft, delivered a keynote speech titled Post-Moore Era: AI Reshaping Semiconductor CIM Intelligent Manufacturing at the New Intelligent Productivity Forum during the 20th Frost&Sullivan Global Growth, Innovation and Leadership Summit.
Sullivan is a world-leading growth consulting firm with over 28 years of presence in the Chinese market. It has served more than 10,000 enterprises and consistently ranked first in market share as a professional industry advisor for Chinese companies seeking Hong Kong and overseas listings. This year’s summit marks the 65th anniversary of Sullivan’s global founding. It brought together over 300 distinguished guests from home and abroad and more than 5,000 professional attendees, serving as a vital benchmark for insights into cutting-edge technologies and industrial trends. Faced with the industry challenge that semiconductor processes are approaching physical limits in the post-Moore era, Xu Hui systematically elaborated on how artificial intelligence holds the key to breaking through bottlenecks, and shared Glorysoft’s complete technical roadmap and implementation results for intelligent upgrading of CIM.
Xu Hui, Director of Solutions at Glorysoft, delivers the speech.
01
Four Driving Forces Accelerate Growth in China’s Pan-Semiconductor Intelligent Manufacturing Software Solution Market
According to Frost&Sullivan’s market research report: Sales revenue of China’s pan-semiconductor intelligent manufacturing software solution market is projected to grow from RMB 1.3 billion in 2021 to RMB 7.2 billion in 2030, representing a Compound Annual Growth Rate (CAGR) of 25.0% between 2026 and 2030. By 2030, the market size of Chinese enterprises is expected to expand from RMB 1.5 billion in 2025 to RMB 5.5 billion, with a CAGR of 29.7% from 2026 to 2030. The accelerated market growth is mainly fueled by four drivers:
Policy-driven: Initiatives including Made in China 2025 and the Action Plan for Innovative Development of the Industrial Internet identify intelligent manufacturing as a core priority.
Technological transformation: Accelerated deployment of 5G+ Industrial Internet,large AI models, edge computing and digital twin technologies.
Elevated demand: Demand for customized production, short lead times, high quality and low costs pushes forward flexible and intelligent manufacturing.
Data explosion: Large-scale deployment of IIoT equipment drives annual industrial data growth exceeding 30%, highlighting the value of data as a production factor.
02
Evolution of Intelligent Manufacturing in the Post-Moore Era: Glorysoft Reimagines the Future of CIM with GloryIGI + LOFA
Drawing on nearly 20 years of experience in the pan-semiconductor industry, Glorysoft keeps pace with technological trends of the post-Moore era. Building on traditional CIM, the company has launched AI products including the GloryIGI industrial large model. The goal is to empower rather than replace conventional capabilities with intelligence, forming a more comprehensive intelligent production product portfolio that has achieved real-world implementation.
Based on Glorysoft’s implementation experience, the evolution of CIM toward intelligent manufacturing proceeds through three phases:
Digital foundation building: Data accumulation and big data application
Intelligent upgrading: Deployment of scenarios including GloryIGI and the LOFA Light Off Factory Assistant
Autonomous evolution: Fully intelligent factories
To illustrate vividly: data acts as the blood of intelligent manufacturing, CIM serves as the skeleton, and AI functions as the brain. The integration of the three unlocks overall vitality and enhances operational capabilities.
03
iDEP Platform: Turning Industrial Big Data from Storage Costs into Decision-Making Assets
Glorysoft’s iDEP Intelligent Big Data Platform addresses four characteristics of manufacturing big data: massive volume, rapid growth, diverse formats and untapped value. Leveraging well-established data models and mature governance experience accumulated over nearly 20 years, the platform delivers end-to-end closed-loop management covering data access, governance and analysis. It effectively supports the full closed loop from yield deviations occurring on production lines to root cause diagnosis, and codifies analytical experience and rules into an AI knowledge base. This enables rapid knowledge reuse, improved yield, stable quality and efficient talent development.
Take declining product yield as an example. In the past, engineers had to go through a lengthy workflow: checking reports, gathering data, consolidating Excel spreadsheets, conducting analysis, verifying statistical standards and compiling documents to reach a diagnosis on yield loss.Unlike traditional data platforms that merely display reports, engineers can simply ask directly: “Why did line yield drop last month?” Powered by historical data, iDEP delivers diagnostic conclusions and recommendations. Its core value lies in transforming data — once treated purely as storage overhead — into valuable decision-making assets for enterprises.
iDEP enables data to become genuine decision-making assets.
04
AI Empowers the Full Intelligent Manufacturing Chain: Practices and Outcomes of Glorysoft’s GloryIGI and LOFA Light Off Factory Assistant
Centered on three core priorities for manufacturing — quality, efficiency and cost control — Glorysoft’s GloryIGI industrial large model leverages the combination of GloryIGI and the LOFA Light Off Factory Assistant to fully empower production management.
Quality Management
· Machine vision inspection achieves a detection rate above 99.5% and a defect escape rate below 0.1% via graphical training, drastically cutting labor requirements.
· Root cause analysis tracks abnormalities in real time and pinpoints issues rapidly, institutionalizing lessons learned into closed-loop prevention mechanisms to sustain yield improvement.
Efficiency Improvement
· Intelligent production scheduling and optimization handles more than one thousand process steps (including critical lithography tools) with minute-level optimization, and has been deployed at multiple 300mm fabs.
· Predictive equipment maintenance implements condition-based maintenance based on LSTM and RUL regression prediction to avoid equipment failures in advance.
· Process parameter optimization reverses adjustments to upstream process parameters using test results
Cost Control
· Intelligent warehouse material management boosts delivery efficiency and accuracy through path planning to reduce labor costs.
· Intelligent energy consumption optimization achieves load balancing and off-peak energy utilization by analyzing historical data, lowering unit energy consumption and extending equipment service life.
· Demand forecasting and supply chain coordination calculates optimal procurement timelines for critical materials with procurement cycles spanning six months to one year, and coordinates with suppliers. While stabilizing delivery performance, it effectively reduces inventory and procurement costs, bringing dual improvements in economic benefits and operational efficiency.
Beyond the above eight applications, Glorysoft’s LOFA Light Off Factory Assistant is a field-proven intelligent application that brings AI capabilities down to the production shop floor.
Automatic measurement by tools generates numerous failed measurement points requiring manual re-inspection; tools will halt operation without timely intervention. Through training on massive image datasets and iterative human correction, the LOFA Light Off Factory Assistant realizes automatic visual defect inspection. It cuts manual inspection workload by 75% with a detection rate exceeding 95%.
Previously, engineers manually created measurement recipes for THK tools, which was cumbersome, labor-intensive and inefficient. Supported by intelligent matching and OCR automatic recognition technology, the LOFA Light Off Factory Assistant shortens recipe debugging time for new products from 20 minutes to 2 minutes — a 10-fold efficiency gain. In addition, it saves a 30-minute production cycle for same-batch lots entering the station, delivering significant value for semiconductor front-end processes.
To date, the LOFA Light Off Factory Assistant has been deployed in 24 scenarios across semiconductor front-end manufacturing and advanced packaging back-end processes. This marks its transition from pilot trials to large-scale rollout, effectively driving intelligent upgrading of production lines.
05
Evolution Roadmap: The Leap from Awareness to Foresight
Towards the end of the speech, Xu Hui stated that the evolution of Glorysoft’s AI technologies focuses on coordinated progress across three dimensions: form, operation and technology.
Form: Drive full digitalization of the physical world, enabling digital twins to evolve from passive replication to proactive simulation and deduction.
Operation: Upgrade from “passive response” to “active prediction” to realize efficient innovation through human-machine collaboration.
Technology: Take industrial large models as the core engine to advance CIM + AI toward its ultimate form — shifting from “seeing” to “foreseeing”.
Future smart factories will function as intelligent living entities capable of self-perception, autonomous decision-making and self-optimization. This represents Glorysoft’s long-term strategic direction as it continues to deepen its presence in general semiconductor CIM.
This technology vision aligns closely with the theme of this year’s Sullivan Summit: Anchor the Innovation Boom, Jointly Shape a New Future of Transformative Growth. The presentation drew strong attention and resonant feedback from the audience. Moving from awareness to foresight, Glorysoft will continue to deliver lasting value to more outstanding enterprises.
