The first large scale P2P application on the Internet was Napster (email/SMTP was designed to be P2P and self-hosted, but in practice ended up federated). The spiritual successor to Napster was the BitTorrent protocol, which at peak, represented ~25-35% of all internet traffic in the mid-2000s. Both were permissionless: anyone could host any file.
The common thread between Napster and BitTorrent is file sharing. This is intuitive. Anyone can host a file, and everyone has an upload and download connection. So naturally it makes sense to distribute in-demand files from people who already have the file nearby. File sharing naturally lends itself to a P2P architecture.
Except, this isn’t how the vast majority of files are distributed on the internet. Instead, files are primarily distributed through content delivery networks (CDNs), which are somewhat distributed (though not decentralized) and permissioned. The world’s largest CDNs have 200-400 points of presence (POPs), which are primarily concentrated in the USA and Europe.
A decentralized CDN (dCDN), built on crypto rails, has always been an obvious opportunity. Theoretically, a dCDN should be able to offer lower latency and more redundancy at lower cost due to a structural cost arbitrage on dramatically underutilized upload bandwidth. There were many early attempts in the 2016-2020 era, but none ultimately worked. We have been looking to invest in a dCDN since launching Multicoin in 2017.
Today, we’re excited to announce our $10M Series A investment in Permissionless Labs, the makers of Pipe Network, the first permissionless content delivery network (CDN) on Solana. The round also included Solana Ventures, Robot Ventures, and our friends Anatoly Yakovenko and Meltem Demirors. We believe Pipe Network has the potential to become the world’s largest and fastest CDN because it’s inherently permissionless, designed to radically reduce latency, and built on the most efficient crypto rails.
CDNs are naturally performance sensitive applications. With the right orchestration, a dCDN should be able to outcompete a centralized CDN (cCDN). However, building the orchestration layer to support a permissionless environment is hard and net new. Building a dCDN requires a team with deep domain knowledge. We believe we have found the right team, helmed by David Rhodus, founder and CEO. David boasts a 20-year professional career in streaming media and big data. Before founding Pipe Network, he served as the CTO of Volar Video, an early streaming company. Later, he was an early employee at Elemental Technologies, which built the earliest large-scale, HD broadcasting infrastructure for the NFL and other similar media properties. Elemental was ultimately acquired by AWS and is now an integral part of their video streaming platform. He then went on to be an early employee at Consensys where he gained an understanding of permissionless crypto networks from first principles. We believe David and his team have the perfect background to make the low-level optimizations needed to make a permissionless dCDN outperform a permissioned cCDN.
In addition to the pedigree of the team, we invested in Pipe because of their tight focus. There are many teams in the Decentralized Virtual Infrastructure Network (DeVIN) space that are trying to do many things. For example, there are many adjacent functions to a CDN, including archival storage, and computing over data. From the moment we met David, he was on a mission to build the world’s biggest and fastest CDN. As David likes to say, the north star for Pipe Network is 3ms + speed of light response times.
The Pipe Network will be launching its private beta soon and opening up testnet to Pipe POP Node Operators. The first public demo of Pipe Network will be at Breakpoint, Solana’s annual developer conference in Singapore, next week. In the meantime, you can learn more about it now here.
And naturally, Pipe is hiring. If you’re interested in building the future of high performance content distribution on crypto rails, get in touch.
