Point2 secured $136 million to address AI's wiring challenge. Its RF cables outperform copper in terms of distance and surpass optics in latency.

      Point2 has secured $136 million in Series B funding, led by LB Investment, with participation from Arm and Maverick Silicon. Supporters also include Nvidia, Bosch, and Molex. Their e-Tube RF cables offer ten times the reach of copper, three times lower optics power, and one thousand times less latency, aimed at rack-scale AI computing interconnect.

      Point2 Technology is using the Series B funding to commercialize a novel cable type designed to replace both copper and optics in AI data centers. The latest investment round saw LB Investment leading, with Arm joining as a new strategic partner and Maverick Silicon continuing its support. The list of investors also includes Nvidia, UMC Capital, Molex, and Bosch Ventures. The involvement of companies creating AI chips, chip architectures, connectors, and automotive sensors indicates a shared concern that wiring is now a critical bottleneck.

      The e-Tube platform from Point2 employs RF signaling over plastic waveguides rather than traditional copper or optical fiber. Its advantages are significant: it boasts ten times the reach of copper at a similar cost, five times lighter weight, and reduced cable volume by half. Compared to optical solutions, it claims to offer three times lower power usage, three times reduced costs, and one thousand times lower latency without the issues associated with failing lasers. Earlier this month, Lumilens raised $700 million for optical interconnects, believing light is the solution, while Point2 asserts that RF can provide the same performance at a lower cost, with improved cooling and speed for large-scale operations.

      "As AI systems grow and the demand for bandwidth escalates to terabit-per-second levels, interconnect has emerged as a crucial bottleneck," stated CEO Sean Park. Arm's SVP of strategic ventures, Paul Williamson, positioned the investment concerning the complete technology stack: "The future of AI infrastructure will depend on breakthroughs across all layers, from computing to memory and interconnect." This funding aims to expedite the commercialization of Point2’s Active RF Cable, near-packaged e-Tube, and co-packaged e-Tube solutions for large-scale computing systems.

      The expansion of AI data centers has utilized hundreds of billions of dollars in chips, energy, and cooling resources. Due to power constraints, data center builders are seeking financial support, making every watt saved on interconnect critical for computation. Point2 argues that while the industry is investing in faster chips, the connecting cables have not evolved at the same pace. The challenge remains whether RF over plastic waveguides can meet the scaling needs of hyperscalers, but the current investor roster indicates a collective belief that it can.