Point2 Technology, a developer of RF-based interconnect technology for artificial intelligence and accelerated computing infrastructure, has expanded its Series B financing to $136 million, adding strategic investment from Arm and continued participation from existing investor Maverick Silicon.
The latest financing was led by LB Investment and adds to a growing group of strategic and financial investors supporting Point2’s technology. The company’s investor base includes major names such as NVIDIA, Arm, UMC Capital, Molex, Bosch Ventures, Maverick Silicon and other new and existing backers.
The funding comes as AI data centers face increasing challenges in moving massive volumes of data between GPUs, accelerators, memory and other components. As AI workloads become larger and computing clusters continue to scale, high-speed interconnect technology is emerging as an increasingly important part of data-center infrastructure.
Point2 Targets the AI Data Center Interconnect Bottleneck
The rapid expansion of generative AI and accelerated computing has created demand for data-center systems capable of moving enormous amounts of data with minimal latency and power consumption.
According to Point2 Technology, conventional interconnect technologies are facing growing limitations as AI systems move toward terabit-per-second bandwidth requirements.
The company believes its e-Tube Technology Platform can address these challenges by using RF signaling over plastic waveguides to create a new category of Active RF Cables (ARCs).
Point2 Co-Founder and CEO Sean Park said the latest investment validates the company’s approach to addressing what it sees as one of the major bottlenecks in scaling AI infrastructure.
“As AI systems scale and bandwidth demands reach terabit-per-second speeds, interconnect has become the defining bottleneck,” Park said.
The company plans to use the new capital to expand its engineering, systems, operations and business teams while accelerating commercialization of its interconnect products.
Arm Joins Point2’s Growing Investor Base
The participation of Arm adds another major semiconductor industry player to Point2’s investor base.
Arm Senior Vice President of Strategic Ventures Paul Williamson said future AI infrastructure will require innovation throughout the technology stack, including compute, memory and interconnect.
“The next generation of AI infrastructure will rely on advances across the entire technology stack, from compute to memory and interconnect,” Williamson said.
For Point2, strategic participation from companies across the semiconductor, accelerator and connectivity ecosystem could provide more than financial support as the company moves toward commercialization.
The company is developing its technology for applications involving AI accelerators, rack-scale computing and high-bandwidth data-center architectures.
What Is Point2’s e-Tube Technology?
Point2’s e-Tube Technology Platform is based on an RF signaling architecture designed to transmit data through plastic waveguides.
The technology is intended to combine some of the advantages associated with copper and optical interconnects while addressing their respective limitations.
Point2 says its platform can deliver:
- 10X greater reach than copper at comparable cost
- 5X lower weight than copper
- 2X lower cable volume than copper
- 3X lower power consumption than optical solutions
- 3X lower cost than optical solutions
- 1000X lower latency than optics, according to the company
- Scalable bandwidth density for accelerator-to-accelerator and rack-scale computing applications
These characteristics are particularly relevant as data-center operators and AI infrastructure companies attempt to build increasingly dense computing clusters.
Active RF Cable Technology for AI Infrastructure
One of Point2’s key products is its Active RF Cable (ARC) technology.
Active RF Cables are designed to provide high-bandwidth connections between computing components while using RF signaling through plastic waveguides.
Point2 argues that this architecture can provide an alternative to traditional copper cables and optical interconnects for certain AI infrastructure applications.
The company is targeting applications where bandwidth, power consumption, latency, physical density and cost are becoming increasingly important.
As AI accelerator clusters grow, the physical interconnect between chips and systems can become a significant factor in overall data-center performance and energy efficiency.
Point2 Expands Product Roadmap With NPE and CPE
Beyond Active RF Cables, Point2 is developing multiple form factors through its e-Tube platform.
These include:
Active RF Cables (ARC)
ARCs are designed for high-bandwidth cable interconnect applications and represent Point2’s effort to establish RF-based cables as an alternative interconnect technology for AI and accelerated computing.
Near-Packaged e-Tube (NPE)
The company’s Near-Packaged e-Tube (NPE) approach is intended to bring the interconnect technology closer to the compute package, potentially enabling higher bandwidth density and shorter electrical paths.
Co-Packaged e-Tube (CPE)
Point2 is also developing Co-Packaged e-Tube (CPE) solutions designed for applications where interconnect technology can be integrated more closely with compute or networking components.
Together, these approaches form part of Point2’s broader strategy to create an interconnect platform that can scale alongside next-generation AI computing architectures.
Why AI Data Center Interconnects Matter
The AI infrastructure industry has largely focused on increasingly powerful GPUs, AI accelerators and networking technologies. However, the ability to efficiently move data between these components is becoming equally important.
Large AI models require enormous amounts of data to be exchanged between processors and memory. In distributed AI systems, thousands of accelerators may need to communicate continuously.
This creates several engineering challenges.
Bandwidth must increase to prevent compute resources from waiting for data.
Latency must remain low so that communication does not become a major performance bottleneck.
Power consumption must also be controlled because interconnects can represent a meaningful portion of the energy used by large data-center systems.
Finally, physical weight, cable volume and density become increasingly important as racks become more densely packed with accelerators.
These challenges are creating opportunities for companies developing alternatives to conventional electrical and optical interconnect technologies.
LB Investment and Maverick Silicon Back Point2
LB Investment CEO Kiho Park said the company’s technology could help address infrastructure challenges created by larger and more power-efficient AI computing clusters.
Maverick Silicon has also continued its investment in Point2.
Josh Miner, Principal at Maverick Silicon, said the firm’s continued investment reflects confidence in Point2’s RF-based approach and its potential to address the performance, efficiency and cost requirements of hyperscalers and accelerator vendors.
The continued participation of existing investors, combined with new strategic backing from Arm, gives Point2 additional resources as it moves toward commercial deployment.
Point2’s Growing AI Infrastructure Opportunity
The expansion of Point2’s Series B financing to $136 million highlights the growing investor interest in technologies designed to solve the infrastructure challenges associated with AI scaling.
The company’s strategy is centered on a relatively specific but increasingly important part of the AI data-center stack: high-bandwidth, low-power interconnects.
With AI clusters becoming larger and more densely integrated, the interconnect architecture connecting accelerators and other computing components could play an increasingly important role in determining system performance, efficiency and scalability.
Point2 plans to use the new financing to strengthen its engineering and commercial capabilities while accelerating the development and commercialization of its ARC, NPE and CPE solutions.
If its e-Tube platform can achieve the performance, cost and reliability characteristics claimed by the company at commercial scale, it could position Point2 as a significant player in the emerging market for next-generation AI data-center interconnect technology.
About Point2 Technology
Point2 Technology develops ultra-low-power and ultra-low-latency RF and mixed-signal interconnect system-on-chips designed for scalable, high-bandwidth-density, multi-terabit cable interconnects.
Its e-Tube Technology Platform uses RF signaling through plastic waveguides and is designed for AI and accelerated computing applications. The company’s product roadmap includes Active RF Cables, Near-Packaged e-Tube and Co-Packaged e-Tube solutions for next-generation data-center architectures.
With its Series B financing now totaling $136 million and strategic participation from companies including Arm and NVIDIA, Point2 is positioning its technology to address one of the critical infrastructure challenges emerging from the rapid expansion of AI computing.