Series A Funding Explodes to $200M for AI Startup Eridu

AI networking startup Eridu raised $200 million Tuesday. The series a funding marked one of the largest early-stage rounds in infrastructure this year. Oversubscribed from day one.

Socratic Partners led. So did legendary VC John Doerr, Matter Venture Partners, and others. Total raised: $230 million. Valuation wasn’t disclosed, but founder Drew Perkins said it’s “comparable to others raising this much” at Series A. Translation: probably unicorn territory. Maybe not. He won’t say.

Perkins knows networks. Started in the 1980s. Helped create Point-to-Point Protocol that became part of TCP/IP—the protocol the internet runs on. In 1999, his optical switch company Lightera Networks sold to Ciena for over $500 million. Next: Infinera, which IPO’d and sold to Nokia for $2.3 billion in 2025. Also co-founded Gainspeed (sold to Nokia) and AR startup Mojo Vision.

Then ChatGPT happened.

February 2023: Perkins and OpenAI CEO Sam Altman spoke at a small conference. They got to chatting. “Sam told me that what enabled AI and ChatGPT was just enormous amounts of compute. At which time, I think he meant 4,000 GPUs, but now we’re talking about millions of GPUs,” Perkins said.

That’s when it clicked. The bottleneck isn’t just chips. It’s how chips talk to each other. “What we needed to do in the networking sector was come up with a brand-new way of thinking about how you build networks and build network equipment, network chips, and the entire thing.”

By late 2023, Perkins met co-founder Omar Hassen. Hassen’s background: networking chip design for Broadcom and Marvell. In 2024, they founded Eridu.

**What the Series A Funding Means for AI Infrastructure**

Eridu is rethinking computer networking from scratch. Starting with silicon. New chips designed for AI that integrate more networking functionality on-chip. The goal: sell complete systems that replace what Arista Networks sells in traditional data centers.

How it works: Replace tiered optical connections with on-chip communications. Today’s approach adds boxes when you need more networking. More boxes = more hops = more latency. That delay between typing a prompt and getting a response? Partly that.

Eridu’s building a switch that puts more functions on the chip itself. “So now I’m saving a ton of power, I’m saving a ton of cost, and then my network is much more reliable because the optics are the least reliable part of the network,” Perkins explained.

The gap is widening. “GPU compute and memory bandwidth are improving by roughly 10x per year, while data center switches from Broadcom, Marvell, Cisco, etc. are still only improving 2–3x every 2–3 years,” Perkins said.

That mismatch creates opportunity. Huge opportunity.

**How Eridu Landed the Series A Funding**

Perkins made a few calls to VCs he’s known for years. Landed Wen Hsieh as a key investor. Hsieh is founding managing partner of Matter Venture Partners and previously led Kleiner Perkins’ China investing group.

Hsieh told Doerr about Eridu. Doerr—the legendary former Kleiner Perkins investor who backed one of Perkins’ previous startups—wanted in immediately.

That set off a frenzy. “My phone has been ringing off the hook,” Perkins said. “It’s been a fun time raising money for this venture… we’re very oversubscribed.”

Other leads in the financing include Hudson River Trading and Capricorn Investment Group. Participants: SBVA, MediaTek, Bosch Ventures, TDK Ventures, Eclipse, and VentureTech Alliance (an investing vehicle of chip fab giant TSMC).

Perkins wouldn’t comment on valuation beyond saying it’s “not too low, nor too high.” He wants his roughly 100 employees to do well on their stock options. He also declined to say whether Eridu hit unicorn status.

Smart move. Keeps expectations manageable.

**Why This Round Matters**

If Eridu delivers, it lands in the middle of the largest data center build-out in history. AI infrastructure spending is exploding. Every hyperscaler needs faster, cheaper, more reliable networking to connect millions of GPUs.

Most SaaS startups raising $200 million at Series A are 20-something college dropouts with a vibe and a pitch deck. Eridu’s different. Perkins and Hassen have deep experience. Perkins has built and exited networking companies for decades. Hassen designed chips for the biggest players in the industry.

That matters. Building networking silicon isn’t a side project. It takes years, deep technical expertise, and relationships with fabs like TSMC. Which is why TSMC’s investment vehicle participated in the round.

The market knows the difference between founders who’ve shipped hardware and founders who’ve shipped PowerPoints. This series a funding reflects that.

**What’s Next for Eridu**

Eridu has capital. Now comes execution. The company needs to:

1. Finish chip design and tape out silicon
2. Validate performance claims in real AI workloads
3. Build manufacturing partnerships (TSMC investment helps here)
4. Convince hyperscalers to rip out Arista/Cisco and replace with unproven startup gear

That last one is the hard part. Data center operators are conservative. They don’t swap out networking infrastructure on a hunch. Eridu will need to prove 10x better performance, cost, or reliability. Preferably all three.

But if anyone can do it, it’s Perkins. He’s been here before. Multiple times. Sold companies to Ciena and Nokia. Took Infinera public. Built the protocols the internet runs on.

Most startup advice is bullshit from people who never built anything. Perkins built plenty.

Question is whether the series a funding capital accelerates product development fast enough to beat Broadcom, Marvell, and Cisco to market. Those incumbents aren’t sitting still. They see the same GPU networking gap Eridu sees.

First mover advantage matters in infrastructure. So does execution speed.

Eridu has the team. Has the capital. Has the vision. Now they build.