Eridu Raises Over $200M in Funding to Develop AI Networking Infrastructure

AI right now feels like a stadium packed to the rafters. GPUs are the all stars putting up absurd numbers. Models are the headliners drawing the crowd. But behind the curtain, the network has been quietly struggling to keep pace with the show. When the infrastructure feeding those GPUs starts gasping for air, the entire system feels the drag. That tension is exactly where Eridu stepped in.

Now the company has emerged from stealth with $200M+ in funding and a clear message to the AI infrastructure world. The network wall slowing massive AI data centers is not a law of physics. It is an engineering problem waiting for the right crew. Credit to Drew Perkins, founder and CEO, and Omar Hassen, co founder and CPO, for assembling a team that decided the old map was not good enough for the territory AI is about to cover.

The capital came from a heavyweight table. The Series A was led by Socratic Partners, John Doerr, Hudson River Trading, Capricorn Investment Group, and Matter Venture Partners, with participation from Bosch Ventures, Eclipse Capital, Fusion Fund, Osage University Partners, SBVA, TDK Ventures, VentureTech Alliance, Zelda Ventures, and others including Black Opal Ventures, Catapult Ventures, Fathom Fund, Godfrey Capital, Hyperlink, Leslie Enterprises, MediaTek, Modi Ventures, Ohio Innovation, Open Field Capital, Perkins Enterprises, Pierre Lamond, Rice Management, Struck Capital, and Triple Point Capital. When that many sharp minds lean forward, the signal is loud.

What Eridu is building is not another incremental patch for data center networking. The company is developing a clean sheet network switch and interconnect platform designed for generative AI scale. Fewer network tiers. Less latency. Less jitter. Single hop scale up domains with thousands of GPUs and scale out environments stretching toward millions. The design also aims to cut AI networking capex by up to 40% and reduce networking power by as much as 70%. That math gets the attention of hyperscalers, frontier labs, neoclouds, sovereign clouds, and enterprise AI operators watching every watt and microsecond.

There is serious engineering muscle behind the curtain. Eridu has tackled challenges across silicon, packaging, systems, and optics to build the architecture for the next wave of compute. TSMC is a key manufacturing partner, bringing advanced process and integration expertise into the mix. When the supply chain and architecture move in sync, the result is speed. In AI infrastructure, speed is currency.

The broader takeaway is simple. When a market the size of the $200B AI networking opportunity forms pressure points, the winners rebuild the foundation instead of polishing the surface. Eridu is betting the future of AI does not just need smarter models. It needs roads wide enough to move the traffic.