This article covers Ayar Labs‘ $500 million Series E funding round for its co-packaged optics (CPO) technology. It digs into how this investment could speed up high-volume production, global expansion, and real-world deployment in AI data centers.
You’ll also get a look at the company’s silicon photonics chiplets and key components. There’s a bit about what this means for the broader data center and AI infrastructure world too—big implications, honestly.
Ayar Labs Secures a Major Series E to Scale Co-Packaged Optics
With total capital raised now close to $870 million and a reported valuation of $3.75 billion, Ayar Labs plans to use the new funds to ramp up manufacturing and testing for its co-packaged optics solution. The money will help the company expand globally, especially with new growth in Hsinchu, Taiwan, and push adoption in AI data centers.
Neuberger Berman led the Series E and will take a board observer seat. New investors joining in include Alchip Technologies, ARK Invest, Insight Partners, MediaTek, Qatar Investment Authority, Sequoia Global Equities, and 1789 Capital.
Funding Details at a Glance
- Series E: US$500 million; total capital about US$870 million; valuation around US$3.75 billion
- Lead investor: Neuberger Berman (taking a board observer role)
- New participants: Alchip Technologies, ARK Invest, Insight Partners, MediaTek, Qatar Investment Authority, Sequoia Global Equities, 1789 Capital
How Ayar Labs’ Co-Packaged Optics Work
Ayar Labs makes silicon photonic “chiplets” that put electrical-to-optical conversion right next to GPUs. This move tackles the notorious power wall that’s been holding back copper interconnects in AI workloads.
By bringing optical I/O straight to the processing units in data centers, Ayar Labs wants to boost throughput, cut latency, and trim power use. The goal? Let thousands of GPUs work together as one efficient system for AI inference and training.
The company uses its GlobalFoundries-manufactured TeraPHY optical I/O chiplets alongside its SuperNova off-chip multi-wavelength light sources. Together, these components aim to move data at high bandwidth while lowering both total cost of ownership and energy usage—a pretty big deal as AI models keep getting larger and more demanding in today’s data centers.
Key Components and the Path to Manufacturing Scale
At the heart of Ayar Labs’ approach, you’ll find two standout components: the TeraPHY optical I/O chiplets and the SuperNova light sources. The TeraPHY chiplets come from a collaboration with GlobalFoundries and pull optical signaling closer to GPUs.
That move cuts down the copper length, which so often slows down AI systems. The SuperNova sources offer off-chip multi-wavelength capability, making data transfer more efficient and scalable.
Together, these components tackle key metrics for AI data centers: throughput, latency, and power efficiency. It’s a pretty direct shot at the old pain points that have plagued high-performance workloads for years.
Ayar Labs aims to help data center architectures keep up with the growing demands of large-language models, real-time analytics, and AI-driven workloads. They want to do this while also easing energy and cooling headaches.
By zeroing in on an industry where copper interconnects are hitting their limits, Ayar Labs tries to carve out a vital spot in the next generation of AI infrastructure. It’s a bold move, but the timing feels right.
Strategic Investors and Global Ecosystem Growth
The Series E round really cements Ayar Labs’ ties with a wide mix of strategic and financial backers. It reflects a bigger industry push toward energy-efficient optical interconnects for AI ecosystems.
Investors like MediaTek and Alchip highlight strong connections to the custom ASIC design world and downstream manufacturing. Meanwhile, funds such as ARK Invest and Insight Partners help fuel Ayar Labs’ growth and expansion plans.
Neuberger Berman’s Gabe Cahill pointed out Ayar Labs’ solid execution and market conviction. He called the company’s solution production-ready and praised its deep ecosystem integration.
This investment aims to speed up a shift in AI infrastructure, moving toward energy-efficient optical interconnects that could sidestep the latency and power issues of copper-based signaling in dense GPU setups. The round comes after a December 2025 Series D, which included AMD Ventures, Intel Capital, and Nvidia.
That earlier round reinforced how strategically important Ayar Labs’ technology is as AI centers keep expanding, especially in Asia-Pacific economies like Taiwan. The momentum seems to be building fast.
Here is the source article for this story: Ayar Labs Raises US$500 Million