Three scientists at the U.S. Naval Research Lab founded the dark web: David Goldschlag, Michael G. Reed, and Paul Syverson invented onion routing in 1995, the core technology that hides who is talking to whom online (1) (2). They first presented their work on May 31, 1996, and the code became publicly available in July 1996 (2).
The Three Scientists Who Invented Onion Routing in 1995
David Goldschlag, Michael G. Reed, and Paul Syverson at the U.S. Naval Research Lab (NRL) created the first research designs and prototypes of onion routing in 1995 (1). Their goal: build internet connections that don't reveal who is talking to whom, even to someone monitoring the network (1). This technology became the foundation for what we now call the dark web.
The Office of Naval Research funded the initial work (2). Goldschlag and Reed shared an office at NRL's Washington, DC campus (3). Syverson regularly carpooled with Goldschlag as they developed the system (3). This working arrangement meant the three scientists could discuss problems and test ideas daily, turning theoretical concepts into working code.
Why Three Scientists, Not One
Onion routing required expertise in cryptography, network protocols, and security analysis. Goldschlag, Reed, and Syverson each brought different skills to the project. No single researcher could have solved all the technical challenges alone. The 1995 invention emerged from their combined knowledge, not a lone genius.
The team designed onion routing to create internet connections that don't reveal who is talking to whom (1). Someone monitoring the network can see encrypted traffic moving between nodes, but cannot determine the original sender or final destination. This property—hiding communication patterns, not just message content—distinguishes onion routing from simple encryption.
From Research Lab to Working Prototype
By 1995, the three scientists had built working prototypes (1). These early systems proved the concept functioned: you could route traffic through multiple encrypted layers and hide the connection between sender and destination. The prototypes ran at NRL, tested by the scientists themselves before any public presentation.
The Naval Research Lab environment shaped the invention. Government funding meant the scientists could focus on long-term research without immediate commercial pressure. They could document vulnerabilities and design solutions before releasing anything publicly. This careful approach produced a system that addressed security threats from the start.
The First Onion Routing Prototype and Public Release (1996)
The NRL team presented their work publicly for the first time on May 31, 1996 (2). The paper, titled "Hiding Routing Information," appeared at the First Information Hiding Workshop (2). This presentation marked the moment onion routing moved from classified research to public knowledge.
The Proof of Concept System
The prototype they demonstrated was a 5-node system running on a single machine at NRL (2). It operated on Solaris 2.5.1/2.6 with proxies for web browsing (2). Running all five nodes on one machine proved the concept worked without requiring a distributed network. You could encrypt data in layers, route it through multiple nodes, and strip away each layer—all while hiding the connection between sender and destination.
This single-machine setup was practical for research but not for real-world use. A true onion routing network needs nodes on different machines, controlled by different operators, spread across different locations. The proof of concept showed the cryptography and routing logic functioned correctly. Scaling it to a public network would come later.
Public Code Release
Two months after the workshop presentation, NRL approved the generation 1 onion routing code for general public distribution in July 1996 (2). This decision transformed onion routing from a government research project into technology anyone could study, modify, and deploy.
The 1996 paper described features that wouldn't be deployed until later generations (2). Rendezvous points—the mechanism that allows hidden services to operate without revealing their server location—appeared in the paper as a theoretical concept (2). The paper also documented tagging attacks and other vulnerabilities the team had identified (2). These forward-looking sections gave future developers a roadmap for improving the system.
Public release in July 1996 meant researchers worldwide could examine the code, test it, and propose improvements (2). This openness accelerated development and identified weaknesses faster than any closed government project could. The decision to release the code publicly shaped everything that followed, including the second generation design that would emerge in the early 2000s.
From NRL Project to Tor: The Second Generation
In the early 2000s, Roger Dingledine, a recent Massachusetts Institute of Technology (MIT) graduate, began working on an NRL onion routing project with Paul Syverson (1). This collaboration marked the transition from the first generation prototype released in 1996 to a redesigned system that would become the modern dark web infrastructure.
Dingledine called the project Tor, which stood for The Onion Routing (1). The name served a specific purpose: to distinguish this original work at NRL from other onion routing efforts that were starting to pop up elsewhere (1). By the early 2000s, multiple research groups and developers were experimenting with anonymous routing systems. The Tor name clarified which project descended directly from the 1995 NRL invention.
The Second Generation Design Team
Roger Dingledine, Nick Mathewson, and Paul Syverson designed the second generation of onion routing (4). This three-person team combined Dingledine's fresh perspective from MIT, Mathewson's technical skills, and Syverson's experience from the original 1995 invention. The second generation addressed limitations in the 1996 prototype and implemented features the first paper had described but never deployed.
The 1996 paper mentioned rendezvous points and other advanced features that the single-machine proof of concept couldn't support (2). The second generation design made these features operational. Rendezvous points allow hidden services to exist without revealing their server location—the mechanism that enables .onion sites to operate on the dark web. Without this feature, onion routing would only hide client connections, not server locations.
What Changed Between Generations
The first generation proved onion routing could hide communication patterns (1) (2). The second generation made it practical for a distributed public network. The redesign addressed scalability, performance, and security issues that emerged after the 1996 code release (2). The team built a system that could support thousands of nodes operated by volunteers worldwide, not just five nodes on a single NRL machine.
The second generation architecture became the foundation for the modern dark web (4). When people access .onion sites today, they use the routing protocol and hidden service design that Dingledine, Mathewson, and Syverson created in the early 2000s. The technology that enables onion sites and how to access them traces directly to this second generation work.
The transition from NRL project to Tor shows how the dark web evolved through documented stages: military research in 1995, public code release in 1996, and practical redesign in the early 2000s (1) (2) (1) (4). No single moment or person "founded" the dark web—it emerged from collaborative work across a decade.
Dark Web Technology Timeline: From Naval Research to Public Network
| Year | Milestone | Key People | What Happened | Source |
|---|---|---|---|---|
| 1995 | Onion routing invented | Goldschlag, Reed, Syverson (NRL) | First designs and prototypes created | Tor Project History |
| May 1996 | First public presentation | Goldschlag, Reed, Syverson | 'Hiding Routing Information' paper presented | Onion Routing History |
| July 1996 | Code released publicly | NRL team | Generation 1 code approved for distribution | Onion Routing History |
| Early 2000s | Second generation begins | Dingledine, Syverson | MIT graduate joins NRL project | Tor Project History |
| Early 2000s | Tor project named | Dingledine, Mathewson, Syverson | Second generation design completed | NRL ACM Fellow |
| 2000s onward | Term 'dark web' emerges | No specific inventor | Public term for hidden services | No documented origin date |
Q&A
- Who found dark web in usa
David Goldschlag, Michael G. Reed, and Paul Syverson at the U.S. Naval Research Lab in Washington, DC invented onion routing in 1995, the technology that enables the dark web (1) (2). Goldschlag and Reed shared an office at NRL's DC campus, while Syverson carpooled with Goldschlag as they developed the system (3). The Office of Naval Research funded their initial work (2).
- How did the dark web get started?
The dark web started when NRL scientists released their onion routing code for public distribution in July 1996, two months after presenting it at the First Information Hiding Workshop (2). Roger Dingledine, Nick Mathewson, and Paul Syverson then designed the second generation in the early 2000s, which implemented rendezvous points that allow hidden .onion sites to operate (4). This second generation transformed the research prototype into a network that could support the hidden services we now call the dark web.
- Why was the dark web created?
The Naval Research Lab created onion routing to build internet connections that don't reveal who is talking to whom, even to someone monitoring the network (1). The military needed to protect intelligence communications, but the system had to work for everyone—if only government users accessed it, their traffic would stand out. The scientists designed it to hide communication patterns for any user, which is why they released the code publicly in 1996 (2).
- Does the dark web still exist in 2026?
The provided facts document the dark web's creation in 1995 and development through the early 2000s, but contain no information about its current operational status. The second generation architecture designed by Dingledine, Mathewson, and Syverson became the foundation for modern Tor (4), though these facts don't confirm whether the network operates today. You can verify current status by checking the Tor Project's official documentation.
- Can I join the dark web?
The provided facts explain how scientists created the technology and released the code publicly in July 1996 (2), but don't include instructions for accessing the network. The second generation design implemented features like rendezvous points that allow hidden services to operate (4), though these historical facts don't describe how to connect. Check current Tor Project documentation for access methods and security considerations.
- What are the top 5 dark web sites to explore?
The provided facts document the invention and technical development of onion routing from 1995 through the early 2000s (1) (4), but contain no information about specific sites or services. The 1996 paper described rendezvous points that would eventually enable hidden services (2), though these historical sources don't list or recommend particular destinations. Site recommendations fall outside the scope of this historical and technical documentation.
Key Takeaways
Three Naval Research Lab scientists—David Goldschlag, Michael G. Reed, and Paul Syverson—invented the technology that enables the dark web in 1995, not a single mysterious founder (1) (2). Here's what matters:
The inventors were government researchers, funded by the Office of Naval Research to solve a military communication problem (2).
Public release happened quickly: NRL approved the code for general distribution in July 1996, just two months after the first presentation (2).
The second generation team made it practical: Roger Dingledine, Nick Mathewson, and Paul Syverson redesigned onion routing in the early 2000s, implementing features like rendezvous points that allow hidden services to exist (4).
No single person "found" the dark web: it emerged from collaborative research, public code releases, and iterative improvements over years.
The technology was designed for everyone: the scientists released it publicly because a government-only network would defeat the purpose of hiding communication patterns (1) (2).
If you want to understand how this technology works in practice today, read our guide on onion sites and how to access them.
Where to look next
These services are useful starting points for further research. Tools & services.