SEMICON West 2025 marked a significant milestone for the semiconductor manufacturing industry as for the first time, the event was held in Phoenix, Arizona. This reflects the region’s rapid rise as a U.S. semiconductor manufacturing hub. Much of this growth is driven by TSMC’s continued expansion in the area, as well as other ecosystem investments spurred by the CHIPS and FABS Act.
This growth was evident to anyone attending. According to SEMI, attendance at this year’s event doubled compared to 2024, and there were 45% more exhibitors showcasing cutting-edge solutions across the supply chain. This energy mirrors the broader momentum of the industry as it accelerates toward becoming a $1 trillion global market by 2030.
Entering the era of Industry 5.0
A recurring theme at SEMICON West 2025 was the industry’s shift from Industry 4.0 to Industry 5.0. While Industry 4.0 centered around data-driven automation and “smart factories”, Industry 5.0 builds upon these foundations to emphasize human-centric, sustainable, and resilient manufacturing. The Smart Manufacturing Pavilion at SEMICON West highlighted these themes with sessions on humans collaborating with agentic AI and digital twins, resilient and scalable data platforms, and sustainability in the industry.
Digital Twins: Data-driven manufacturing at scale
One of the most discussed topics throughout the conference was the evolution of digital twins from isolated models to fully scalable, AI-driven platforms that power real-time decision-making across entire fabs.
Digital twins are now serving as the link between AI and manufacturing. Speakers at the Smart Manufacturing Pavilion, representing companies such as Intel, GlobalFoundries, and EMD Electronics, discussed how many current systems are not suitable for AI, so in order for this link to be strengthened and maintained, it is critical for digital twins to be backed by scalable data platforms.
Together, these tools can enable manufacturers to:
- Produce consistent material quality
- Reduce cycle time and scrap rates
- Deliver zero-defect manufacturing
- Lower production costs
- Decrease overall environmental impact
However, with these benefits comes the challenge of model drift. Over time, as equipment and processes evolve, model quality can degrade, reducing the reliability of digital twins. EMD Electronics underscored the importance of continuous monitoring and regular retraining to maintain model accuracy. This ensures better forecasting and helps prevent out-of-control (OOC) events before they impact yield or uptime.
Via Automation and Cohu’s sessions highlighted some additional challenges that plague the semiconductor industry. Unplanned downtime, new hire training, reduced equipment availability, and high repair times are all issues that are costing companies time and money. Therefore, some of the largest value of integrating digital twins and AI lies in improving yield, reliability, OEE, and the speed of new product introduction (NPI).
Other digital twin use cases were highlighted across the semiconductor lifecycle, including fault detection and classification, yield management and prediction, statistical process control (SPC), advanced process control (APC), and run-to-run (R2R) control—illustrating how central this technology has become to the industry’s growth.
Agentic AI: Empowering the workforce of the future
Alongside digital twins, agentic AI emerged as a cornerstone of the Industry 5.0 vision. Rather than simply generating insights, agentic AI can plan, reason, and take autonomous action, creating closed-loop systems that optimize operations in real time. The use cases for agentic AI in semiconductor manufacturing are virtually endless. Ranging from improving defect detection, predictive maintenance, scheduling, and virtual metrology, companies across the industry are finding ways for agentic AI to save them time and money.
A critical dimension of this year’s discussions was the human impact of this technology. The semiconductor industry faces a well-documented skills gap, with workforce shortages posing one of the biggest barriers to continued growth. Agentic AI is increasingly being positioned not just as an operational tool, but as a workforce development catalyst.
Through AI-powered copilots and task-oriented agents, organizations can onboard and train employees more efficiently, increase productivity, and bridge the knowledge gap between experienced engineers and newer staff. This has the potential to accelerate skill development across fabs, reduce errors, and allow human workers to focus on higher-value problem-solving.
As several presenters emphasized, the future of semiconductor manufacturing lies at the intersection of human expertise and AI, where intelligent agents support people and make employees more efficient, driving both operational and workforce transformation.
Are you interested in learning more about how Spotfire can take your semiconductor operations to the next level? Explore how Spotfire Copilot™ can help you accelerate time to insights.