Aeva Unveils Optical Amplifier to Tackle AI's Data Center Crisis

MOUNTAIN VIEW, CA – January 15, 2026 – Aeva Technologies, a company primarily known for its advanced 4D LiDAR sensors for autonomous vehicles, has made a significant strategic move into the heart of the artificial intelligence boom. The firm today announced a new high-power semiconductor optical amplifier (SOA) designed to address the escalating power and data transmission challenges crippling the world's most advanced AI data centers.

This development signals a critical diversification for the Nasdaq-listed company, leveraging its deep expertise in silicon photonics to target a new, explosive growth market. As AI models become exponentially more complex, the infrastructure supporting them is straining under the immense pressure of unprecedented data loads and power consumption. Aeva's new component aims to be a key part of the solution, enabling the next generation of AI hardware to operate more efficiently and at a greater scale.

The AI Infrastructure Bottleneck

The rapid ascent of generative AI has created a voracious appetite for computational power, leading to what many industry experts call a data center infrastructure crisis. The core challenge is no longer just processing power, but the ability to move massive datasets between thousands of processors quickly and efficiently. Traditional copper-based electrical interconnects are hitting fundamental physical limits, becoming a major bottleneck in terms of bandwidth, latency, and especially, power consumption.

In response, the industry is aggressively shifting toward optical solutions. Technologies like Co-Packaged Optics (CPO), which brings optical components directly onto the same package as processing chips, are seen as the inevitable future. The goal of CPO is to drastically shorten data pathways, reduce energy-wasting electrical signaling, and enable denser, more powerful computing clusters. Market projections underscore this urgency, with the market for optical components in AI clusters expected to more than double from $5 billion in 2024 to over $10 billion by 2026. The CPO market alone is forecast to exceed $20 billion by 2036, growing at a staggering 37% compound annual growth rate.

This transition requires a new class of optical components that are not only powerful but also exceptionally efficient and thermally robust. Every watt of energy saved at the component level translates into significant reductions in overall power consumption and cooling costs for a hyperscale data center—a critical factor as AI workloads push energy grids to their limits.

Aeva's High-Performance Optical Engine

Aeva's new semiconductor optical amplifier is engineered specifically to meet these demanding requirements. The company reports that its SOA delivers an optical output power exceeding 28 dBm (approximately 630 milliwatts), a figure that positions it at the upper echelon of performance for such components. This high output power is crucial for driving optical signals across the complex, high-bandwidth fabrics that connect thousands of AI accelerators.

Perhaps more importantly, the device achieves this power with a wall-plug efficiency greater than 20% while operating at temperatures up to 50 degrees Celsius. This combination of high power, efficiency, and thermal robustness is a key differentiator. High efficiency directly combats the soaring energy consumption of AI hardware, while the ability to perform reliably at elevated temperatures simplifies thermal management, reduces system costs, and improves the long-term reliability of large-scale deployments.

"As AI infrastructure and autonomous systems continue to scale, the industry needs optical technologies that can deliver high performance without compromising efficiency, manufacturability, reliability, or cost," said Pradeep Srinivasan, VP of Photonics at Aeva, in the company's official announcement. The firm states its SOA is produced in modern semiconductor fabs, a process that enables high yields and cost-effective scaling to meet the surging market demand.

A Strategic Leap Beyond LiDAR

While the move into data center components may seem like a pivot, it is more accurately a strategic expansion that leverages Aeva's foundational technology. The company built its reputation on a vertically integrated silicon photonics platform, which is the same core competency behind its industry-leading Frequency Modulated Continuous Wave (FMCW) 4D LiDAR. This technology allows Aeva's sensors to measure velocity in addition to position, a key advantage in autonomous systems.

By applying this photonics expertise to the data center market, Aeva is diversifying its revenue streams and entering a sector with immense growth potential. It's a natural extension that allows the company to capitalize on its significant R&D investment in silicon photonics across multiple verticals. This dual-market strategy de-risks its business model and positions Aeva as a fundamental technology provider for two of the most transformative trends of the decade: AI and autonomy.

The new SOA is not just for data centers; Aeva has confirmed it is also designed to enhance the performance of advanced, multi-beam FMCW LiDAR systems. The increased optical power and efficiency can translate into longer detection ranges, higher resolution, and greater robustness for its perception sensors, further solidifying its competitive edge in the automotive and industrial automation markets.

This dual-use capability underscores the power of a platform-based approach. The same underlying innovation that helps an AI cluster process a large language model more efficiently can also help an autonomous vehicle perceive its environment more clearly. As detailed in the press release, Srinivasan noted the SOA "has the potential to enable a new class of scalable architectures across both AI data centers and FMCW LiDAR applications for automotive and Physical AI applications." Aeva is scheduled to present a paper with detailed technical results of its new amplifier at the upcoming SPIE Photonics West Conference on January 19, an event that will be closely watched by industry analysts and competitors.