---

SiTime has launched a new MEMS-based timing device aimed at improving GPU utilization in AI data centers, where synchronization errors can significantly reduce computing efficiency.

The company’s Elite 2 Super-TCXO is designed to provide sub-nanosecond synchronization accuracy across AI clusters, helping hyperscalers and system architects reduce wait cycles, avoid GPU timeouts, and increase throughput in large-scale AI training systems.

For eeNews Europe readers, the launch highlights how timing technology is becoming a critical part of AI infrastructure design, particularly as data center operators look for ways to improve performance per watt and maximize expensive GPU resources. It also underlines the growing role of MEMS timing devices in applications traditionally dominated by quartz technology.

Timing bottlenecks in AI clusters

According to SiTime, GPU utilization rates in AI clusters can fall as low as 20% to 40% because distributed AI workloads require extremely precise synchronization between processors.

“Industry reports show GPU utilization in AI clusters can be as low as 20 to 40 percent—a large and largely hidden tax on AI infrastructure,” said Piyush Sevalia, chief business officer at SiTime. “AI workloads are distributed across GPUs in tightly orchestrated time slots. Even small timing errors force wait cycles to avoid data corruption, and in extreme cases can trigger GPU timeouts and system restarts. Poor synchronization directly caps GPU utilization.”

The company says the industry is moving from synchronization targets of around 1 microsecond today toward 10 nanoseconds across AI clusters. SiTime claims its Elite 2 device delivers synchronization accuracy of 1ns, which it says is 10x better than the industry target.

Sevalia added: “To address this, the industry is driving towards a target of 10 nanoseconds time synchronization across an AI cluster, down from 1 microsecond today. We collaborated closely with leading AI system architects at hyperscalers and silicon providers and concluded that the right oscillator can significantly improve cluster-wide synchronization.”

Smaller footprint, higher stability

The Elite 2 Super-TCXO is available in 3.2 mm × 2.5 mm and 5.0 mm × 3.2 mm packages, with claimed improvements in thermal stability, Allan deviation and frequency stability over competing solutions.

The device also includes digital frequency tuning and withstands shock, vibration and board bending, areas where MEMS devices typically outperform quartz-based timing components.

Sameh Boujelbene, vice president at Dell’Oro Group, said: “AI networks must operate with extremely high efficiency to fully utilize expensive GPU resources. As AI back-end infrastructure refreshes at a much faster cadence than traditional non-accelerated infrastructure, time synchronization accuracy becomes increasingly important to sustaining performance across rapidly evolving data center architectures.”

SiTime says the Elite 2 Super-TCXO is sampling now, with volume production expected in the third quarter of 2026. The company estimates the addressable market for AI synchronization timing devices could reach a cumulative $1.5bn by 2030.

---

---