Salience Labs
Salience Labs develops fully integrated silicon photonics-based all-optical switches that address data movement bottlenecks in AI datacenters. The switches deliver 10 ns switching latency with 20x lower network latency, sub-1 W per port power consumption, and 10x lower cost per port versus electronic alternatives, enabling improved tokens per second per user for next-generation AI workloads.
- Semiconductor
- Hardware
- Artificial Intelligence
Funding
Raised to date
$170MRaised to date based on public sources. This may differ from the amount the company actually raised and is based only on what is publicly available on the internet.
$136.5M across 2 disclosed rounds
OS
AV
+6$125M · Series B
Announced Oct 2026 · from article date
Disclosed cumulative: $125M
Reported total $170M
Not on chart (date unknown): Seed $11.5M
Incomplete round history. $33.5M of the $170M reported total isn’t matched to a known round.
Founders
Product
Problem
AI datacenters face a fundamental bottleneck in data movement, where traditional electronic switches and interconnects limit processing speed and energy efficiency. Significant investment in optical pluggables and interconnects has been made, but without all-optical switches, the efficiency and latency gains of optical connectivity cannot be fully realized.
Solution
Salience Labs provides fully integrated, silicon photonics-based all-optical switches that scale network bandwidth and decrease latency, significantly improving AI workload performance. By replacing electronic switching with optical switching, the technology reduces network latency by 20x and power consumption by up to 8x per port, while cutting cost per port by 10x. The switches enable lower latency optical switching that improves tokens per second per user, supporting the interactivity required for next-generation AI workloads. Leveraging traditional semiconductor manufacturing processes, the company offers high-volume production with industry-leading cost advantages and a clear roadmap for scaling to higher bandwidths.
Target Audience
Primary customers are AI datacenter operators, hyperscalers, and cloud service providers who need high-bandwidth, low-latency networking infrastructure to scale AI training and inference workloads efficiently.
Features
- All-optical switches built on fully integrated silicon photonics technology
- 10 ns switching latency, achieving 20x lower network latency compared to electronic switches
- Sub-1 W per port power consumption, up to 8x lower than conventional solutions
- 200 Gbit/s high-speed interface per port
- 10x lower cost per port versus electronic alternatives
- Scalable all-optical interfaces designed to support multiple future bandwidth generations, including 128- and 256-port configurations
- Manufactured using standard semiconductor fabrication processes for high-volume, cost-effective production