Xscape Photonics Closes $81M Series A Round
Xscape Photonics has closed a $37 million funding round, bringing its total Series A investment to $81 million. The round was led by Addition, a new investor in the company, with continued participation from IAG Capital Partners and NVIDIA, among others. The fresh capital doubles the company’s valuation and deepens its runway as it advances a photonics platform built specifically for AI data center networks.
The latest investment extends the $44 million Series A Xscape announced in 2024. Alongside the funding, the company unveiled FalconX, its newest product: the first fully redundant External Laser Small Form-factor Pluggable device capable of generating up to eight wavelengths of light simultaneously.
The combination of new capital and a new product marks a clear step forward for a startup that has now raised nearly $95 million in total funding since its founding.
The Bandwidth Problem Holding AI Back
AI workloads are straining data center infrastructure in ways that were not anticipated even a few years ago. The core issue is bandwidth. As GPU clusters grow in size, the connections between accelerators, memory, and storage become a bottleneck. Today, many of those connections still rely on copper interconnects, which simply cannot move data fast enough to keep pace with what modern AI systems demand.
Optical interconnects offer a fundamentally different approach. By transmitting data as light rather than electrical signals, photonic solutions can carry far more information at lower power and over greater distances. Xscape Photonics is building at the center of this shift, developing platforms that bring multi-wavelength optical interconnects into the infrastructure layer of AI data centers.
“Rapidly increasing bandwidth, power and cost demands of AI workloads have created a critical hardware bottleneck, forcing developers to use just a fraction of their GPUs’ capacity. FalconX is the industry’s first Comb laser module in a pluggable form factor capable of generating eight wavelengths of light, powering high-speed data movement to allow the entire data center to function as one giant GPU.”
Vivek Raghunathan, CEO and co-founder of Xscape Photonics
What the New Funding Will Build
The $37 million will accelerate development of Xscape’s multi-color wavelength-division multiplexing fabric solutions, a technology that routes different wavelengths of light along the same optical path to multiply available bandwidth without adding new cables or hardware. The focus is on scaling this approach across larger and more complex AI cluster configurations.
A significant portion of the investment will go toward advancing ChromX, Xscape’s programmable multi-wavelength photonics platform. The company is targeting laser systems capable of generating 16, 32, and eventually 128 or more colors of light, with each additional color representing another lane of high-speed data. In August 2025, Xscape successfully demonstrated functional 16-color prototypes in collaboration with Tower Semiconductor. In June of the same year, it launched the EagleX Laser Evaluation Kit as a first step toward the broader ChromX platform.
FalconX: Eight Colors, Built for Reliability
FalconX is the product Xscape is bringing to market now. It uses the company’s proprietary CombX laser technology, which generates multiple wavelengths of light on a single silicon photonics chip. The result is a single pluggable module that produces more than 1W of optical power across eight colors, capable of powering multi-terabits per second of data bandwidth.
Reliability was central to the design. AI clusters have grown more than tenfold in size over the past two years. When a single laser module fails in a large cluster, it can stall an entire workload and drive up token processing costs. Hyperscalers now require failure rates ten times lower than what conventional laser modules deliver.
FalconX addresses this with built-in redundancy and more durable components, and it is designed to comply with industry MSA standards for both Scale-Up and Scale-Out fabric links, meaning it can be deployed within existing hyperscaler infrastructure.
The Team Behind the Technology
Xscape Photonics was founded in 2022 by Vivek Raghunathan and four photonics researchers from Columbia University: Alexander Gaeta, who serves as President; Yoshi Okawachi, Vice President of R&D; and Keren Bergman and Michal Lipson, both members of the Board of Advisors. The founding team brings a combination of deep academic research in photonics and direct experience building semiconductor systems.
The company’s platform, ChromX, is designed with three priorities in mind: performance, power efficiency, and cost at scale. As AI computing demands grow, the environmental and economic cost of powering data centers has become a real concern for operators. Xscape’s approach is built to reduce power consumption per bit of data moved, not just increase raw speed.
Since its founding three years ago, the company has attracted nearly $95 million in total investment and continues to expand its engineering team.
Investors Back a Specific Bet on Photonics
Addition, the new lead investor in this round, made its position clear.
“We see this as a compelling opportunity to invest in optical laser innovation at a moment when advanced photonics is becoming critical to AI infrastructure. Xscape is addressing the bandwidth bottlenecks that constrain AI cluster performance today.”
Lee Fixel, Founder of Addition
Existing backers IAG Capital Partners and NVIDIA also returned for this round. NVIDIA’s continued involvement is worth noting. As the dominant supplier of AI accelerators, its stake in a photonics company working on GPU cluster interconnects reflects where AI infrastructure investment is heading. Total funding across all rounds now stands at approximately $95 million.
Origianl Creator: Ekaterina Pisareva
Original Link: https://justainews.com/companies/funding-news/xscape-photonics-closes-81m-series-a-round/
Originally Posted: Wed, 11 Mar 2026 19:44:30 +0000












What do you think?
It is nice to know your opinion. Leave a comment.