Point2 Technology, a Silicon Valley startup developing RF-based interconnect solutions for AI infrastructure, secured an extension to its Series B funding round led by Maverick Silicon, with participation from NVentures (the venture arm of NVIDIA) and UMC Capital, bringing total funding to $76 million. The investment targets commercialization of Point2’s Active RF Cable (ARC) platform and its broader e-Tube™ interconnect architecture aimed at next-generation AI data center deployments.
The company positions its e-Tube platform as an alternative to conventional copper and optical interconnects, which face scaling constraints in high-performance AI clusters. As hyperscalers deploy increasingly dense XPU fabrics, interconnect bandwidth, power efficiency, and reach have emerged as limiting factors. Point2’s approach uses RF signaling over plastic waveguides to address these challenges, targeting improved latency, reduced power consumption, and simplified system design compared to optical links that rely on lasers and complex alignment.
Funding will support expansion of engineering teams and productization across multiple form factors, including pluggable cables, near-package interconnects, and co-packaged solutions. The company reports that its ARC technology delivers up to 10x the reach of copper, with lower weight and volume, while reducing power consumption and latency relative to optical alternatives. The architecture targets accelerator-to-accelerator connectivity and rack-scale system design in large AI clusters.
• Series B extended to $76 million with Maverick Silicon, NVentures, and UMC Capital
• Focus on commercialization of Active RF Cable (ARC) and e-Tube platform
• Targets scale-up interconnect for AI clusters and rack-scale systems
• Claims: 10x copper reach, 3x lower power vs optics, reduced latency and system complexity
• Supports pluggable, near-package, and co-packaged interconnect architectures
“We are honored to welcome Maverick Silicon as the lead investor in this round, and deeply appreciate the support from NVentures and UMC Capital,” said Sean Park, CEO and Co-Founder of Point2 Technology. “As AI systems scale, interconnect is increasingly the limiting factor. Our e-Tube platform takes a fundamentally different approach—using RF techniques to deliver the reach, efficiency, and latency required for next-generation scale-up architectures.”
🌐 Analysis: Point2 enters a crowded but rapidly evolving interconnect landscape where copper is extending through linear drive and advanced materials, while optics continues to scale via pluggables and co-packaged optics. The participation of NVentures suggests strategic interest from NVIDIA in alternative scale-up interconnect approaches alongside NVLink and optical ecosystems. The emergence of RF-based cable architectures highlights ongoing industry exploration of hybrid electrical-optical paradigms for AI infrastructure.
| Profile: Point2 Technology | |
| Founded | 2016 (verified) |
| Headquarters | San Jose, California |
| CEO / Co-Founder | Sean Park (background includes Marvell, IDT, TeraSquare) |
| Latest Funding | $76M Series B (extended), led by Maverick Silicon with participation from NVentures (NVIDIA) and UMC Capital |
| Core Technology | e-Tube™ platform: RF signaling over plastic waveguides, enabling Active RF Cables (ARC) and future near-package and co-packaged interconnect architectures |
| Positioning | Alternative to both copper (reach/power limits) and optics (cost, power, complexity) for scale-up AI interconnects |
| e-Tube vs Copper | ~10x longer reach; ~5x lighter; ~2x lower cable volume; improved flexibility for dense rack environments |
| e-Tube vs Optics | Lower power (no lasers); simplified architecture (no optical alignment); lower latency path (no electro-optical conversion); competitive for rack/row-scale reach |
| Latency / Power | Designed for ultra-low latency (minimal DSP/FEC) and reduced power vs optical links (company cites ~3x lower power in target use cases) |
| Reliability | Eliminates laser components, reducing common optical failure points and thermal constraints |
| Product Roadmap | 1) Active RF Cables (ARC) – near-term 2) Near-Package e-Tube (NPE) – mid-term density scaling 3) Co-Packaged e-Tube (CPE) – long-term integration for rack-scale fabrics |
| AI Data Center Relevance | Addresses power density, cable congestion, and cost scaling challenges in large XPU clusters and rack-scale architectures |
| Manufacturing Approach | Leverages CMOS-compatible RF/mixed-signal design and polymer waveguide materials for potential cost advantages vs optics |
| Industry Recognition | Named a 2026 BloombergNEF Pioneer for sustainable, scalable data center infrastructure |

