Chiplets, AI Agents and Silicon: Highlights From the International Microelectronics Forum 2026

The second day of Silicon Mountains 2026 in Yerevan was dedicated to the International Microelectronics Forum, which brought leading international and Armenian microelectronics and semiconductor companies to the stage. The talks shared one theme: how next-generation systems are designed, tested and managed when complexity outgrows conventional approaches.

Chiplets and the future of automotive compute

Dr. Katayoon Basharkhah, Lead Technologist at IMEC (Belgium), presented the future of automotive computing platforms. The industry, she said, is moving toward heterogeneous, chiplet-based systems driven by the demands of AI, performance and scalability. Standardization and interoperability are essential on this path — and the key question is no longer “can we build this?” but “how do we design and validate the right chiplet system before first silicon?”

Testing across the whole silicon lifecycle

Dr. Artur Jutman, founder and CEO of Estonia’s Testonica, argued that classic testing is no longer enough given the complexity of high-performance and AI systems. In his approach, test solutions should be reused throughout the entire silicon lifecycle — from manufacturing to in-field monitoring — and the data collected on-chip should be turned into fault management, from diagnosis all the way to autonomous recovery.

AI agents in chip design

Faisal Shturu, Technical Account Manager at Siemens EDA, showed how AI is changing design work. His three main points: agentic systems flag risks before they become problems; AI agents explore several directions at once; and expertise gets captured in the tool rather than staying with individual specialists.

Synopsys: 100% of graduates employed in microelectronics

Prof. Vazgen Melikyan, Director of the Educational Department at Synopsys Armenia, presented results of the Synopsys educational program. According to his data, 100% of the program’s graduates have found jobs in microelectronics, 77% of them at Synopsys Armenia, and 56% of the company’s employees have come through the program. In total, the program has trained 95 PhDs, 1,161 master’s and 1,588 bachelor’s graduates.

AI networks start with silicon

Arsen Manukyan, Head of Cisco Armenia, spoke about the evolution of AI networks, reminding the audience that it all starts with silicon. He presented systems powered by the Cisco Silicon One G300 platform, featuring 100% liquid cooling and around 70% higher energy efficiency.

Validation, timing and data-center silicon

• Manuel Ohanyan of National Instruments stressed that in semiconductor validation, AI mainly helps manage and improve existing workflows — speeding up the scaling of test code and anomaly detection — while human involvement remains the foundation of product optimization.
• Sergey Ghahramanyan, Lead QA Engineer at Cadence Design Systems, presented the TimeVision AI Agent, part of the InnoStack AI Super Agent, which helps catch SDC errors before final sign-off.
• Harutyun Harutyunyan, General Manager of Xsight Labs Armenia, presented the company’s open, programmable silicon solutions for AI data centers — from switch to DPU — including a lightweight version of the E1L.

Silicon Mountains is Armenia’s leading technology platform, organized since 2019 by the Union of Employers of ICT (UEICT) with the support of Armenia’s Ministry of High-Tech Industry and Artificial Intelligence. More information: silicon-mountains.org.

 7 просмотров всего,  7 просмотров сегодня