Zeki Gunay

Two announcements at CES 2026. Different scopes, different partners. But solving essentially the same problem. QNX/Vector unveiled Alloy Kore. Elektrobit launched EB civion. Both are betting, rightly I think, that OEMs are done with bespoke SDV development. The fragmentation era is (hurrah!) ending. Integration complexity has overwhelmed legacy OEMs. Every custom architectural decision compounds across generations. Even BMW and Mercedes are now working with QNX on foundational software—a clear signal that "build everything" is unsustainable. The shift is from custom development to pre-integrated platforms. Let vendors handle the complexity. Focus OEM resources on differentiation. This is industrialization, not capitulation. But it represents measurable progress toward platforms that might actually enable the software-defined vehicle at manageable cost and complexity. Read the full analysis in my TechInsights Daily Insights post (free registration required): https://lnkd.in/gCyS-GhZ #SoftwareDefinedVehicle #SDV #AutomotiveSoftware #CES2026 #AutomotiveTech (Image created by asking ChatGPT to "draw me an image that embodies the concept of reducing complexity and fragmentation, ideally with an automotive theme. ")

60

4 Comments

Any fab or pilot line in Europe which can be setup for 55nm open source technology? Is that imposible? When we started Tropic Square auditable chip was also considered imposible. Few years later it is different story... Short after we started #opentitan went public, open source PDK and ecosystem followed.

16

OEMs building scalable automated systems need perception that holds up in poor visibility. Speaking with SAE Media, Arbe’s CTO Dr. Noam Arkind explains why sensor fusion only works when radar resolution is high enough and how ultra-high-resolution radar improves motion estimation and free-space detection. Full story on SAE: https://lnkd.in/e74ed_dc

72

1 Comment

🔋 Baettery 2.0 Challenge — Episode 6 Battery Management System Maps & Look-up Tables (DMI / CMI) For Episode 6 of the Baettery 2.0 Challenge, we focus on a cornerstone of Battery Management Systems: 👉 Designing DMI (Discharge Maximum Intensity) and CMI (Charge Maximum Intensity) maps as functions of State of Charge (SOC) and Temperature (T). 🎯 The challenge BMS current limits must ensure safety, performance, and durability, yet they are often built: • With limited test data • Under tight development timelines • Across wide SOC and temperature ranges Creating reliable current limit maps early is therefore a critical — and non-trivial — task. ⚙️ What we explore in Episode 6 • Building DMI / CMI look-up tables with SOC and temperature dependencies • Using simple, physics-informed model structures to extrapolate sparse data • Ensuring smooth, consistent, and safe operating envelopes • Bridging cell behavior to BMS-ready control maps 🔍 How Baettery 2.0 helps Baettery 2.0 enables rapid creation of BMS-ready current limit maps, combining limited measurements, battery knowledge, and modeling to generate structured, consistent look-up tables for both cells and packs. 📩 Have a BMS use case? Send us a DM with your application — current limits, SOC/T envelopes, early BMS calibration, or cell-to-pack translation. 👉 We’ll explore it together with a trial run on Baettery 2.0. More episodes coming soon — always grounded in real battery engineering challenges. ⚡ #Baettery20Challenge #BMS #BatteryManagementSystem #BatteryModeling #ElectroThermal #EnergyTransition

15

Excited to share this news! Cadence Palladium Z3 Emulation and Protium X3 FPGA Prototyping system has been awarded “EDA/IP/Software Product of the Year” in the 2025 AspenCore World Electronics Achievement Awards for its continued innovation and unique strengths in verification field. This "dual power" system not only continues Cadence's legacy in hardware simulation and prototyping verification, but also sets a new standard for verification efficiency with its entirely new architecture: ⚡️ Capacity increased by more than 2 times, supporting design scales from 16 million gates to 48 billion gates ⚡️ Speed increased by approximately 1.5 times, enabling faster compilation, shorter deployment, and higher throughput ⚡️ Integrated migration architecture, allowing seamless switching between simulation and prototyping ⚡️ Modular design, flexibly expandable to meet the parallel verification needs of different teams. Their emergence marks a new stage from "functional verification" to "system-level acceleration." Discover more about our tools here: https://ow.ly/5bKG30sRTvr

30

A quick note on magnetics in Class D amplifier power supplies When working on Class D designs, I’ve found that transformers and inductors are often where things really get decided. The main SMPS transformer needs to survive fast load changes without adding noise. PFC and output inductors aren’t just about inductance value — saturation, DCR, and thermal behavior show up quickly under real music signals. And the output filter inductor? That one can quietly affect distortion and EMI more than expected. In many cases, issues don’t come from a single component, but from how the magnetic parts interact as a whole system. That’s usually where most tuning time goes.

7

2 Comments

TSMC has announced a major efficiency breakthrough, reducing peak power consumption of its EUV lithography tools by 44%, which is expected to save around 190 million kilowatt-hours of electricity by 2030. This move not only lowers operational costs but also reinforces TSMC’s commitment to sustainable semiconductor manufacturing as it scales toward advanced nodes like 2nm and beyond. #TSMC #TechGiants #Foundry #Foundries #EUVLithography #SustainableTech #Tech #Chips #Semiconductors #GreenManufacturing #2nm #Innovation #ChipManufacturing #SemiconductorIndustry #TechnologyNews https://lnkd.in/d3zNkn8p

7

2 Comments

The company I co-founded, Vertical Semiconductor, has raised $11M to build the foundational power layer for the next era of AI compute. AI is outgrowing today’s power systems. So we’re building what comes next, with high density, high efficiency, high voltage Vertical GaN power electronics.  Today’s funding news – a seed round led by Playground Global with participation from JIMCO (Jameel Investment Management Co.), milemark•capital, Shin-Etsu, Kastor Capital, LLC is a major milestone in our mission. Vertical Semiconductor was born from eight years of research at Massachusetts Institute of Technology, led by my co-founders Dr. Josh Perozek and Prof. Tomás Palacios. Together, we’re redefining what power delivery will look like for decades to come. It has been an immense personal privilege to work with Josh and Tomas to bring fundamental innovations in semiconductor design and fabrication to solve the energy industry’s greatest challenges.  Their pioneering work in gallium nitride (GaN) led to a breakthrough vertical architecture — one that unlocks the material’s full potential for scalable, efficient, and high-voltage power systems. First-in-the-world manufacturing advancements developed by Josh makes this technology ready for the massive scale demanded by the market. Our journey begins in data centers, where rethinking power delivery is essential to unlocking the next generation of compute. As AI data centers moves toward megawatt-scale racks, the bottleneck is no longer compute — it’s power. The infrastructure that moves energy from the grid to the GPUs can no longer keep up. Vertical Semiconductor will enable the power delivery needed for the future of AI. I’m deeply grateful to Matt Hershenson, Bruce Leak, Pat Gelsinger, Peter Barrett, Rahul Meka and the entire Playground Global team for believing in us and our vision of the future. We are incredibly lucky to partner with investors that bring such deep technical expertise and high conviction in building deep technology companies that matter. We’re also fortunate to have the support of all of our investors, as well as our earliest champions — Advanced Research Projects Agency-Energy (ARPA-E), Martin Trust Center for MIT Entrepreneurship, MIT Proto Ventures, Massachusetts Clean Energy Center, MIT Climate and Energy Prize, Dr. Luis Perez-Breva, PhD and beyond — who helped turn years of research into reality. The future of compute depends on the future of power. With this milestone, we’re accelerating the charge at the frontier of the Vertical GaN revolution — and we’re just getting started.

799

158 Comments

Voice AI is increasingly important in #physicalAI and #AgentiveAI systems, and user experiences are getting better every day. See our latest innovation to help our customers get right-first-time solutions on #XCORE #GenSoC technology.

5