Valens Semiconductor Interview: AI-Powered Conference Room Connectivity and the Future of Enterprise Collaboration

Semicon Leaders Asia Interviews: Eli Ofek, VP of Products for Valens Semiconductor, on AI-Powered Conference Room Connectivity and Enterprise Collaboration

1. AI-powered features such as automatic camera switching, speaker tracking, and intelligent framing are becoming standard in modern conference rooms. What are the biggest connectivity and infrastructure challenges organisations face when deploying these capabilities at scale?

Leveraging these features requires multiple high-res cameras to send uncompressed video to AI processors in real time, which creates myriad infrastructure challenges, including:

• Bandwidth issues: Using multiple cameras generates massive data. Traditional USB and HDMI connections have distance limitations, so meeting room designers have to add extenders, converters, or compression, which increases costs and complexity.

• Too many devices: A lot of these tools require separate processing boxes, AV-over-IP equipment, and power supplies. Every new device introduces another potential point of failure while also making installation and troubleshooting more difficult.

• Long cable runs: Many enterprise conference rooms have cameras mounted a distance from the computers for optimal coverage. It’s not easy to maintain low-latency, high-quality video over these distances without signal degradation.

• Power, space, and heat: As more devices are added, power consumption increases while space becomes limited, often leading to overheating.

• Reliability and maintenance: Loose cables, extender failures, firmware incompatibilities, and synchronisation issues can disable intelligent features and ruin meetings.

• Scalability: What works in a single conference room might not scale to hundreds of meeting spaces. 

2. Many AI-enabled conferencing systems require additional extender boxes, cables, and power supplies. Why has the industry evolved this way, and what are the key drawbacks of this approach from both a cost and reliability perspective?

Conference room components evolved independently from each other. Also, cameras, displays, microphones, and compute platforms were originally meant for consumer electronics, not large commercial installations. As rooms became larger and AI features demanded more cameras and higher-resolution video, installers compensated by adding extenders, converters, active cables, and separate power supplies.

As I mentioned earlier, the key drawback is that every connector, cable, and powered device represents another possible point of failure. If a meeting room isn't working, it's often because one small hidden component or cable somewhere in the signal chain isn't working correctly.

3. You mentioned that embedded connectivity technology can simplify conference room deployments. How does this approach reduce hardware complexity while still supporting advanced AI-driven collaboration features?

Valens’ technology allows high-speed connectivity to be embedded directly into endpoint devices like cameras and displays. This removes the need for external extender boxes, converters, and additional power supplies, which have traditionally been used to transfer high-bandwidth video over long distances. Now, instead of having to pass signals through multiple devices, the video travels directly over a single cable while maintaining the necessary bandwidth and low latency for AI-driven features like speaker tracking, intelligent framing, and automatic camera switching.

This makes the system architecture much simpler, with fewer potential points of failure. The result is a cleaner and more scalable conference room design that's easy to deploy and maintain while still supporting advanced AI capabilities.

4. Valens Semiconductor suggests that embedded technology can reduce hardware costs by as much as 30%. What factors contribute to these savings, and where do customers typically see the greatest return on investment?

The main factor is eliminating unnecessary hardware.  You no longer need to pay for as many external extenders, converts, power supplies, and cabling (potential to reduce the number of cables from up to six to just one).  Installation is also much quicker because there are fewer devices to mount, configure and troubleshoot. This type of set-up also pays off in the long run by simplifying maintenance and reducing ongoing support costs.

5. Beyond lower installation costs, how does reducing the number of external components improve long-term reliability, maintenance, and the overall lifecycle of conference room systems?

Fewer service calls, faster troubleshooting, and less downtime for employees who depend on meeting rooms every day. It also makes future upgrades easier because fewer interconnected devices need to be replaced or reconfigured. From an IT perspective, simplifying the infrastructure means spending less time maintaining conference rooms and more time supporting higher-value initiatives.

6. As enterprises continue investing in AI-powered collaboration spaces, what trends do you expect to shape the next generation of conference room connectivity, and how is Valens Semiconductor preparing for these evolving requirements?

We will see more and more systems that incorporate multiple cameras, beamforming technology, AI processors, and intelligent displays, all working together in real time.

At Valens, we're focused on enabling that transition by providing embedded high-speed connectivity technologies that allow device manufacturers to build simpler, more scalable AI-enabled conferencing systems. Our experience transporting high-speed sensor data in demanding environments has positioned us well to support the next generation of intelligent collaboration.

7. For organisations planning to upgrade or build AI-enabled meeting spaces, what key considerations should they evaluate when selecting connectivity technologies to ensure scalability, performance, and future readiness?

They need to consider how the room will evolve over the next five to ten years. Will the connectivity tech be able to support camera resolution increases, evolving AI processing requirements, and new sensors without requiring major infrastructure changes? 

They should also aim for simplicity. A solution with minimal external hardware will be much easier to deploy and maintain.

Finally, I recommend that they consider interoperability and standards-based technologies that allow equipment from different manufacturers to work together. Conference rooms have become business-critical infrastructure, and connectivity should enable organizations to add new AI capabilities without adding unnecessary complexity.