The Invisible Yield Killer: How a New Metrology Alliance is Rewriting Fab Rules

SEATTLE, WA – October 13, 2026 – The semiconductor industry is rapidly approaching the physical limits of silicon, but the most pressing bottleneck in sub-2nm manufacturing is no longer just lithography. It is the volatile, invisible engine that etches the circuits and deposits the materials: plasma. Controlling this highly reactive state of matter requires immense amounts of radio frequency (RF) power, delivered with microsecond precision. Yet, for decades, fabs have struggled with a structural blind spot. The power measured when a tool is calibrated in a pristine factory rarely matches the power actually delivered on the grueling, high-volume fabrication floor.

This discrepancy—a seemingly minor calibration drift—is a major source of yield loss in advanced wafer fabrication. Today, a new partnership aims to permanently close that gap. EHT Semi, a developer of advanced RF power delivery systems, has integrated RF measurement technology from the 80-year-old metrology stalwart Bird directly into its semiconductor power platforms. By embedding NIST-traceable Bird probes into systems like the Orion Direct RF platform, the collaboration establishes a standardized, unbroken chain of measurement from factory calibration to fab production.

This is not merely an equipment upgrade; it is a structural shift in how the industry approaches advanced process control. As semiconductor manufacturing becomes an automated, AI-driven enterprise, the quality of the data feeding those algorithms is paramount. Embedding primary metrology directly into the power supply is rewriting the rules of global fabrication competition, challenging entrenched power giants and setting a new baseline for yield optimization.

Taming the Plasma Chamber for Sub-2nm Yields

To understand the significance of this integration, one must understand the chaotic nature of plasma. Inside a semiconductor processing chamber, plasma acts as a dynamic, non-linear impedance. It is highly sensitive to operating conditions, meaning its electrical resistance constantly changes. Traditional RF power systems rely on matching networks—complex arrays of capacitors and inductors—to adjust to these changes and prevent power from bouncing back and damaging the generator.

However, as process windows tighten for sub-5nm and sub-2nm nodes, these traditional matching networks are simply too slow. They struggle to keep pace with rapid plasma fluctuations, especially during the short-pulse operations required for high aspect ratio etching. EHT Semi bypassed this limitation by developing a "matchless architecture." Its Orion system directly couples RF power into the plasma in real time, eliminating the matching network entirely for faster response and improved power coupling.

But controlling power without a matching network requires absolute measurement certainty. Minor variations in voltage, current, and phase angle can drastically alter ion energy and plasma uniformity, destroying a wafer worth tens of thousands of dollars. By embedding Bird's NIST-traceable sensors directly into the Orion system, EHT Semi ensures that power regulation remains within 2% of the set point even during the most violent, dynamic plasma load changes.

"Advanced process control depends on knowing that the power measured during system calibration corresponds to the power monitored in production," said Timothy Ziemba, CEO of EHT Semi. "Integrating Bird measurement technology gives engineers a common RF reference from factory to fab, while supporting the speed and precision required for rapidly changing plasma loads."

Fueling the AI-Driven Smart Fab

The integration solves a long-standing hardware problem, but its most profound impact will be felt in software. Modern semiconductor fabrication facilities are rapidly transforming into massive, AI-driven data centers. Advanced Process Control (APC) algorithms and predictive maintenance platforms are deployed to monitor thousands of variables simultaneously, making microsecond adjustments to maximize yield.

Yet, these machine learning models are only as effective as the data they ingest. Historically, fabs relied on external RF power measurements that provided an incomplete picture of the plasma's actual state. It was akin to driving a high-performance race car by only looking at the fuel gauge, rather than the RPMs and tire pressure.

The embedded Bird probes change this dynamic by providing high-fidelity, real-time data directly from the power supply. Utilizing technologies like Bird's BDS2 V-I-Φ measurement solution, the integrated system delivers granular data on voltage, current, phase angle, and impedance across fundamental, harmonic, and intermodulation frequencies. This data is captured with 1% accuracy and supports pulsed RF waveform analysis with a 500-nanosecond time resolution.

Customers will have direct access to this probe-generated data. For fab automation engineers, this is the missing link. By feeding real-time impedance tracking and time-synchronized measurements directly into neural networks, fabs can develop highly adaptive control strategies. The AI can detect minute process drifts and adjust plasma parameters dynamically, long before a defect is etched into the silicon.

"Bird is pleased that EHT Semi selected our RF measurement technology as an integral part of its semiconductor power platforms," said Bird's chief executive officer, Mike Simpson. "This integration will give equipment and process teams greater confidence in the data connecting calibration, development, and production."

A Strategic Alliance Challenging the Power Giants

Beyond the technical achievements, the EHT Semi and Bird partnership represents a calculated business maneuver in a highly consolidated market. The semiconductor power equipment sector is dominated by entrenched giants like Advanced Energy, MKS Instruments, and Comet. These legacy players offer comprehensive RF power solutions and have deeply established relationships with top-tier foundries.

Rather than attempting to build proprietary, in-house metrology to compete with these giants, the Seattle-based challenger opted for a strategic alliance. Partnering with a company founded in 1942—one that serves as the trusted standard for RF power measurement across defense, aerospace, and telecommunications—provides instant credibility. Fabs operate under strict SEMI international standards and ISO quality management systems; bringing in a NIST-traceable authority like Bird immediately satisfies these rigorous compliance requirements.

This move also contextualizes EHT Semi's broader supply chain strategy. In January 2026, the company secured a strategic minority investment and manufacturing collaboration with Jabil Inc., a global manufacturing powerhouse. By outsourcing its manufacturing scale to Jabil and its metrology standards to Bird, the power delivery developer is operating with a highly leveraged, agile business model. It allows the company to focus exclusively on its core intellectual property—matchless RF and tailored waveform pulsed DC power systems—while relying on best-in-class partners to deliver a globally scalable, fully compliant product.

The Future of Adaptive Manufacturing

As the global economy becomes increasingly dependent on automated manufacturing and digital infrastructure, the engines that power these transformations are being fundamentally redesigned. The semiconductor industry's shift toward sub-2nm architectures is forcing a reckoning in how energy is delivered, measured, and controlled at the atomic level.

The days of isolated, uncalibrated power delivery are over. The future of global fabrication competition will be defined by those who can seamlessly link physical power generation with digital intelligence. By bridging the gap between lab calibration and high-volume manufacturing, and by providing the raw, high-fidelity data required for autonomous process control, this new alliance is laying the groundwork for the next generation of smart manufacturing. The integration of legacy metrology with next-generation power architectures proves that the most critical innovations often happen not on the surface of the silicon, but in the unseen electrical currents that shape it.