Ayar Labs Raises $500 Million In Series E Funding Round

Listen to this article

Ayar Labs completed a $500 million Series E funding round, bringing total funding to $870 million and valuing the company at $3.75 billion. The round reflects strong investor confidence in optical interconnect technology for AI infrastructure, with participation from both financial and strategic players across the semiconductor ecosystem.

Ayar Labs’ latest Series E round was led by Neuberger Berman, with new investors including ARK Invest, Insight Partners, Qatar Investment Authority (QIA), Sequoia Global Equities, and 1789 Capital. Strategic participants such as AMD, Alchip Technologies, MediaTek, and NVIDIA also joined, building on prior relationships. The investment aims to expand high volume production, testing capabilities, and global operations, including a new office in Hsinchu, Taiwan. This aligns with broader AI infrastructure trends, where interconnect inefficiencies are seen as a key hurdle to scaling GPU clusters efficiently.

What is Ayar Labs?

Ayar Labs, founded in 2015, specializes in silicon photonics for AI and high performance computing (HPC). Its core technology replaces traditional copper wiring with optical interconnects, enabling faster data transfer using light, which could improve throughput by 4-20 times per watt compared to electrical systems. Products like the TeraPHY optical chiplet and SuperNova multi wavelength light source target data centers facing exploding bandwidth demands from AI workloads.

The funding seems likely to accelerate Ayar Labs’ shift toward commercial scale deployment, potentially positioning it as a leader in overcoming AI’s “power wall.” With backing from chip giants, it may foster deeper ecosystem partnerships, though success depends on navigating manufacturing complexities and market adoption rates. This could contribute to more efficient AI systems overall, but stakeholders should note ongoing debates around the timeline for widespread optical integration versus incremental copper improvements.

Ayar Labs, a pioneer in silicon photonics and co-packaged optics (CPO), has solidified its position in the AI infrastructure landscape with its Series E funding round of $500 million. This infusion, led by investment firm Neuberger Berman and including a mix of institutional and strategic investors, elevates the company’s total funding to $870 million and its valuation to $3.75 billion. The round underscores the escalating demand for innovative interconnect solutions that address the fundamental limitations of traditional copper based systems in data centers, particularly as artificial intelligence (AI) workloads push the boundaries of power consumption, bandwidth, and latency.

What is the Background of Ayar Labs Company?

Established in 2015 by CEO Mark Wade and a team from academic institutions including MIT, UC Berkeley, and CU Boulder, Ayar Labs emerged from research under the DARPA PIPES program. The company’s foundational technology stems from breakthroughs in embedding optical signaling directly within application specific integrated circuit (ASIC) packages, enabling data transmission via light rather than electrical signals. This approach, known as in-package optical I/O, promises to revolutionize how processors, memory, and accelerators communicate in high performance environments.

Ayar Labs’ product portfolio centers on the TeraPHY optical I/O chiplet, which integrates silicon photonics engines capable of 2 Tbps bandwidth through eight 256 Gbps channels. Complementing this is the SuperNova multi wavelength light source, which supports ultra high bandwidth, low latency connections optimized for AI training and inference. Manufactured in partnership with GlobalFoundries, these components are designed for co-packaging with GPUs, switches, or other processors, reducing power usage and latency compared to pluggable optics or copper links. By 2025, Ayar Labs had demonstrated the world’s first UCIe (Universal Chiplet Interconnect Express) optical interconnect chiplet, delivering 8 Tbps bandwidth while minimizing energy demands.

The company’s growth trajectory reflects the broader shift in semiconductor design toward chiplet architectures, driven by the end of Dennard scaling and challenges in advanced node yields. Strategic alliances with industry leaders like NVIDIA, AMD, Intel, and TSMC have been pivotal, providing not only capital but also technical validation. For instance, collaborations with Alchip for ASIC design and TSMC for advanced packaging (including COUPE and SoIC technologies) highlight Ayar Labs’ integration into the global supply chain.

Ayar Labs’ Funding History

Ayar Labs’ funding journey illustrates escalating investor interest in optical technologies for AI. Below is a comprehensive table of its known funding rounds, compiled from public announcements and industry reports:

This progression shows a shift from early stage venture backing to large scale strategic investments, with cumulative funding reaching $870 million by early 2026. The Series E round, in particular, aligns financial heavyweights with semiconductor ecosystem players, signaling confidence in Ayar Labs’ readiness for mass adoption. Neuberger Berman’s board observer role further emphasizes governance and scaling expertise.

Ayar Labs’ Technology: Co-Packaged Optics for AI

At the heart of Ayar Labs’ value proposition is co-packaged optics (CPO), which integrates optical engines directly onto a common substrate with compute elements. Traditional copper interconnects, while reliable, face physical limits: as data rates exceed 100 Gbps per lane, signal degradation, power loss, and heat generation become prohibitive. Optical interconnects, using photons instead of electrons, offer advantages in speed (near light speed transmission), efficiency (4-20x better throughput per watt), and scalability (enabling clusters of thousands of GPUs as a unified system).

Key components include:

• TeraPHY Chiplet: Provides bidirectional bandwidth up to 2 Tbps with low power (under 5 pJ/bit), supporting standards like UCIe for seamless integration.
• SuperNova Light Source: A multi wavelength laser delivering 16 wavelengths at 128 Tbps, enabling dense wavelength-division multiplexing (WDM) for high capacity links.
• Integration Ecosystem: Partnerships with TSMC for advanced nodes and Alchip for custom ASICs ensure compatibility with next gen AI accelerators.

In AI contexts, this technology addresses the “power wall” – where interconnects consume up to 50% of data center energy. For hyperscale AI, CPO could reduce latency by 10x and power by 5x, facilitating larger models and faster inference. Demonstrations, such as those with NVIDIA’s NVLink, illustrate potential for hybrid systems blending copper for short reaches and optics for longer ones.

Opportunities and Challenges in Optical Interconnects

The optical interconnect market is projected to grow from $18.25 billion in 2024 to $67.4 billion by 2034, at a CAGR of 14.6%, driven by AI, 5G, and cloud computing demands. AI infrastructure alone represents a massive opportunity, with data center interconnects emerging as a bottleneck amid exploding bandwidth needs (e.g., training models like GPT-4 requires petabytes of data movement). Key drivers include:

• AI Scale-Up: GPU fabrics from NVIDIA and AMD require efficient interconnects; optics could expand cluster sizes beyond current copper limits.
• Energy Efficiency: With data centers consuming 2-3% of global electricity, optics offer a path to sustainability.
• Standardization Efforts: Initiatives like UCIe and collaborations with bodies like the Optical Internetworking Forum accelerate adoption.

However, challenges persist:

• Manufacturing Complexity: Scaling silicon photonics involves thermal management and yield issues, though partnerships with foundries like GlobalFoundries mitigate this.
• Cost Barriers: Initial deployment costs are higher than copper, but long term savings in power and performance could offset this.
• Ecosystem Readiness: Widespread adoption depends on integration with existing infrastructure, where incumbents like Broadcom and Intel hold sway.

Recommended: Alva Energy Raises $33M In Funding Led By Playground Global

Competitive Landscape:

Ayar Labs operates in a dynamic field of startups and incumbents vying for dominance in optical interconnects. The table below compares key players based on focus, funding, and differentiators:

Ayar Labs stands out for its maturity and ecosystem ties, but faces pressure from Lightmatter’s higher valuation and Celestial AI’s broader offerings. Incumbents like Intel (with silicon photonics R&D) and Broadcom (high speed modules) add competition, often through acquisitions or internal development.

Looking ahead, Ayar Labs’ Series E positions it for potential IPO within 1-2 years, capitalizing on AI’s capital deployment boom – estimated as one of the largest in a generation. The funds will likely prioritize production ramp-up at facilities like the new Taiwan office, targeting hyperscalers and AI focused enterprises. Success could reshape data centers, enabling “beyond the rack” scaling where optics unify massive GPU arrays.

Yet, the transition isn’t guaranteed; copper’s entrenchment and hybrid approaches may delay full adoption. Broader implications include reduced global energy use in AI (potentially saving gigawatts) and accelerated innovation in HPC. For investors, this round signals optics as a high growth sector, though risks like supply chain disruptions or slower AI market maturation warrant caution. Ayar Labs’ trajectory, backed by quotes like CEO Wade’s emphasis on overcoming copper bottlenecks, suggests a paradigm shift toward light based computing is underway.

Please email us your feedback and news tips at hello(at)superbcrew.com