doxx.net

Cover Block

Public sources

Name	doxx.net
Tagline	A private, parallel internet for people and AI agents, with built-in security and no personal information required. [doxx.net]
Headquarters	Miami, United States [Business Wire, October 2026]
Stage	Series A [Business Wire, October 2026]
Business Model	SaaS
Industry	Security
Technology	Software (Non-AI)
Geography	North America
Growth Profile	Venture Scale
Founding Team	Repeat Founder
Funding Label	Series A (total disclosed ~$38,000,000) [Business Wire, October 2026]

Links

Public sources

• Website: https://www.doxx.net/
• LinkedIn: https://www.linkedin.com/company/doxxcare
• GitHub: https://github.com/doxxnet

Executive Summary

Public sources

Doxx.net is building a private, programmable network layer designed to give users and AI agents direct, secure connectivity without relying on traditional internet infrastructure or sharing personal data [Business Wire, October 2026]. The company’s $38 million Series A, led by Andreessen Horowitz, is a bet that the convergence of AI autonomy and growing privacy concerns creates a foundational need for a new kind of networking stack [a16z, October 2026].

The platform, termed Agentic Defined Networking (ADN), grew from a founder’s hobby project called Darkflare, a TCP tunnel built to bypass public detection and censorship [Network World, October 2026]. It aims to provide a “privacy-by-architecture” foundation where individuals and enterprises can create isolated networks, and where AI agents can be granted controlled, auditable access to resources. Founder and CEO Barrett Lyon brings a track record of building and exiting critical internet infrastructure companies, including Prolexic (acquired by Akamai) and Defense.net (acquired by F5 Networks) [Business Wire, October 2026]. His background in DDoS defense and large-scale network architecture directly informs the security-first design of the new platform.

With the capital, the company is moving from a closed beta that began in late 2025 into an open beta phase, pursuing a SaaS model with per-device and dedicated IP pricing [Superpower Daily, October 2026]. The next 12-18 months will test whether the technical vision of a parallel, agent-aware internet can translate into tangible enterprise adoption beyond early adopters, and if the platform’s control mechanisms for AI operations prove to be a compelling wedge into broader corporate security budgets.

Independently corroborated - Core company details, funding, and founder background are confirmed by multiple independent sources including Business Wire, Network World, and a16z.

Taxonomy Snapshot

Axis	Classification
Stage	Series A
Business Model	SaaS
Industry / Vertical	Security
Technology Type	Software (Non-AI)
Geography	North America
Growth Profile	Venture Scale
Founding Team	Repeat Founder
Funding	~$38,000,000

How the Company Got Here

Public sources

Doxx.net is a Miami-based networking infrastructure company, founded by serial entrepreneur and security expert Barrett Lyon. The company emerged into public view in October 2026 with a $38 million Series A financing led by Andreessen Horowitz and the launch of an open beta for its Agentic Defined Networking (ADN) platform [Business Wire, October 2026]. The founding year is not publicly disclosed.

The company's origin traces to a personal project called Darkflare, a TCP tunnel over Cloudflare that Lyon built to bypass public detection and censorship [Network World, October 2026]. This hobbyist effort evolved into the commercial infrastructure now branded as doxx.net. A closed beta for the platform reportedly began in December 2025, preceding the public open beta by approximately ten months [Superpower Daily, October 2026].

Lightly corroborated -- Key milestones corroborated by multiple news reports; founding date and early corporate history not independently verified.

Product and Technology

Sources and analysis The core product is Agentic Defined Networking (ADN), a platform that creates private, parallel networks for users and AI agents with built-in security controls [Business Wire, October 2026]. The public positioning frames this as "privacy-by-architecture" infrastructure, designed to operate without requiring personal information from users and to give them direct control over network connections [doxx.net About]. The platform's genesis is traced to Darkflare, a hobby project by founder Barrett Lyon that created a TCP tunnel over Cloudflare to bypass public detection and censorship [Network World, October 2026].

Based on the company's website and launch materials, the platform's advertised capabilities are extensive but remain largely self-reported. It claims to enable private networks where users and AI agents can browse, communicate via encrypted calls and messages, and transfer files directly between devices [doxx.net]. The architecture is described as a zero-log infrastructure using its own IP space, DNS root, and hardware, with features like native IPv6 privacy and sub-millisecond threat detection [doxx.net]. A key differentiator is the ability for AI agents to configure network controls via API, ostensibly giving developers fine-grained oversight over where and how autonomous agents connect [a16z, October 2026]. The software is offered under a dual license model: MIT with a Commons Clause for commercial use [docs.doxx.net].

Pricing, as listed on the company's site, follows a simple per-component model: adding a device or agent costs $1 per month, a dedicated IP is $10 per month, and a doxx.net domain is $1 per month [doxx.net Pricing]. The company announced an open beta for the platform on October 1, 2026, following a closed beta that reportedly began in December 2025 [Superpower Daily, October 2026]. Operational scale is described in one source as a mesh network run by 12 people across 26 global sites [BigGo Finance].

Public sources

The market for doxx.net's offering is defined by a growing tension between the expanding autonomy of AI agents and the perceived inadequacy of existing network infrastructure to securely manage them. The company's launch coincides with a period where AI agents are moving from controlled, sandboxed environments into more open, operational roles, raising new questions about data provenance, access control, and network-level security that traditional cloud and VPN architectures were not designed to answer.

A formal, third-party market sizing for "Agentic Defined Networking" is not yet available, as the category is nascent. The company's stated customer scope,individuals, enterprises, and AI agents [Business Wire, October 2026],suggests it is targeting a convergence of several large, adjacent markets. For context, the global market for Zero Trust Network Access (ZTNA), a related security architecture, was valued at approximately $3.5 billion in 2025 and is projected to grow at a compound annual rate of over 17% through 2030 (analogous market, Gartner). The broader secure access service edge (SASE) market, which combines networking and security, is forecast to exceed $25 billion by 2028 (analogous market, IDC). These figures illustrate the scale of spending on modernizing secure network access, a core problem doxx.net aims to address with a different architectural premise.

Demand drivers cited in coverage focus on the limitations of current internet protocols and application designs. Analysis suggests the internet's foundational protocol layer has stagnated, leading to widespread carrier-grade network address translation (NAT), identity-based access restrictions, and a reliance on server-mediated applications like WhatsApp and Signal [BigGo Finance]. This architecture creates chokepoints and potential surveillance vectors. The specific tailwind for doxx.net is the rise of AI agents that operate autonomously across networks. As Andreessen Horowitz noted in its investment announcement, the platform provides control over "where AI agents connect and what they can access," moving beyond rented or unverifiable intermediary infrastructure [a16z, October 2026]. This positions the product at the intersection of networking, AI operationalization, and privacy.

Key adjacent and substitute markets include enterprise VPNs, SD-WAN, ZTNA platforms, and consumer privacy tools like encrypted messaging apps. The regulatory and macro environment presents a mixed picture. Increasing global data sovereignty laws (e.g., GDPR, various national data localization rules) and scrutiny of large platform power could create demand for decentralized, user-controlled networking alternatives. Conversely, the same regulatory focus on combating illicit online activity may view fully private, zero-log networks with suspicion, potentially limiting adoption in regulated industries or certain jurisdictions until compliance frameworks are established.

Lightly corroborated -- Market sizing is based on analogous, established categories (ZTNA, SASE) from third-party analysts. Demand drivers and product-market fit claims are supported by investor and media commentary, but the specific addressable market for ADN remains unquantified by independent research.

Competitive Landscape

Sources and analysis Doxx.net enters a market where its core proposition,a private, agent-controlled network,sits at the intersection of several established and emerging categories, none of which offer a direct, like-for-like substitute. The company's positioning is not against a single incumbent but against a fragmented set of alternatives that each solve part of the problem for different user segments.

The competitive analysis must therefore map the landscape through its functional segments.

• Traditional VPNs & Zero-Trust Networks. Incumbents like Cloudflare (Cloudflare One), Zscaler, and Palo Alto Networks (Prisma Access) dominate the enterprise secure access market. Their models are built around user-to-application tunnels with centralized policy control, not peer-to-peer agent networking. Doxx.net's wedge is architectural: it proposes eliminating the intermediary server entirely for certain communications, a shift from a hub-and-spoke to a mesh model [Business Wire, October 2026]. This edge is perishable if major players decide to incorporate similar P2P agentic controls into their existing platforms, leveraging their massive distribution.
• Consumer Privacy & Messaging Apps. Services like Signal, WhatsApp (with its optional end-to-end encrypted chats), and Tailscale offer encrypted communication and networking. However, these are either application-specific (messaging) or still rely on mediating servers for coordination. Doxx.net's claim of "privacy-by-architecture" through its own IP space and zero-log infrastructure targets users for whom even provider-held metadata is a concern [doxx.net]. The exposure here is that mainstream adoption of privacy tools has historically been limited by convenience; doxx.net must prove its agent-centric use case is compelling enough to change behavior.
• AI Agent Infrastructure & Orchestration. Startups like LangChain and SmythOS provide frameworks for building and managing AI agents, but they typically abstract away the underlying network layer, relying on standard internet connectivity. Doxx.net's differentiator is giving those agents a first-class, configurable network identity and private channels [a16z, October 2026]. This is a defensible edge if AI agent deployment scales to a point where network behavior and data provenance become critical operational issues, a need not yet fully addressed by existing platforms.
• Peer-to-Peer Networking Protocols. The technology underpinning doxx.net has conceptual parallels with projects like libp2p or older P2P networks. The company's commercial edge is packaging this capability with a SaaS business model, built-in security filtering, and a focus on AI agents as a primary user [Network World, October 2026]. This edge is durable if they can maintain performance and simplicity superior to open-source alternatives, but it is exposed to the risk of a well-funded open-source project or a cloud provider offering a managed version of similar P2P tech.

The most plausible 18-month scenario hinges on the adoption curve for autonomous AI agents. If enterprise AI agents move beyond sandboxed APIs and begin performing actions across distributed resources, the need for a secure, auditable agent network becomes acute. In that scenario, Cloudflare could be the winner if it rapidly extends its existing developer-focused Workers platform and Durable Objects to natively support agent-to-agent P2P connections, leveraging its ubiquitous edge network. Conversely, doxx.net could be the loser if the market for standalone agent networking develops more slowly than anticipated, leaving the company to compete for traditional VPN use-cases where incumbents have deep sales relationships and feature suites. The Series A capital provides a runway to develop the market, but the clock is ticking to demonstrate that Agentic Defined Networking is a category, not just a feature.

Lightly corroborated -- Competitive mapping is inferred from product claims and adjacent market analysis; no direct competitor comparisons from independent sources were captured.

Opportunity

Public sources The size of the prize for doxx.net is the potential to become the foundational, privacy-native network layer for a future where AI agents are as common as web browsers, creating a new category of infrastructure valued in the tens of billions.

The headline opportunity is that doxx.net could define and own the category of Agentic Defined Networking (ADN), becoming the default secure transport for AI-to-human and AI-to-AI communication. This outcome is reachable because the company is building from the protocol layer up, a historically defensible position, and is targeting a wedge,controlling where AI agents connect,that is not yet directly addressed by existing networking or security incumbents [Business Wire, October 2026]. The founder’s track record of building and selling infrastructure companies to major acquirers like Akamai and Fortinet provides a credible playbook for scaling and eventual exit [Business Wire, October 2026]. The bet is that as AI agents proliferate, the need for a private, verifiable, and agent-configurable network will become non-negotiable for enterprises, creating a platform dependency.

Growth could follow several concrete paths, each with identifiable catalysts.

Scenario	What happens	Catalyst	Why it's plausible
The Enterprise Security Standard	doxx.net becomes the mandated internal network for AI agent deployments within regulated industries (finance, healthcare).	A major financial institution publicly adopts the platform for its AI trading or compliance agents.	The platform’s stated “privacy-by-architecture” and zero-log claims directly address regulatory concerns about data sovereignty and audit trails [doxx.net]. Founder Barrett Lyon’s background in DDoS defense for enterprises establishes credibility with security buyers [TechCrunch, August 2013].
The Embedded Infrastructure	The ADN protocol is licensed and embedded by major cloud providers or AI model developers as the default secure channel for their agents.	A partnership announcement with a cloud hyperscaler or a leading AI lab (e.g., OpenAI, Anthropic).	The platform’s origin as Darkflare, a tool to tunnel through existing networks, demonstrates a pragmatic approach to integration rather than displacement [Network World, October 2026]. The dual-license model (MIT/Commons Clause) is structured for this kind of commercial adoption [docs.doxx.net].
The Consumer Privacy Platform	The product achieves viral adoption among privacy-conscious individuals and developers, creating a massive, decentralized peer-to-peer network.	A high-profile data breach or privacy scandal drives consumer demand for serverless, end-to-end encrypted alternatives to apps like WhatsApp or iMessage.	The platform’s published pricing suggests a focus on individual users and small devices ($1/month per agent), and its feature set includes consumer-friendly P2P comms and file transfer [doxx.net Pricing].

Compounding for doxx.net would manifest as a classic network effect layered with a technical moat. Each new user or agent joining the network increases the value of the private address space and the utility of its P2P communication tools. More critically, as the network grows, the company accrues operational data on connection patterns and threat models across its proprietary infrastructure,its own IP space, DNS root, and hardware [doxx.net]. This data could fuel the “AI-driven operations” mentioned in coverage, creating a feedback loop where the network becomes more efficient and secure at scale, raising barriers for any new entrant trying to replicate the system [BigGo Finance]. The architecture itself, designed to bypass carrier-grade NAT and centralized servers, becomes the moat [BigGo Finance].

The size of the win, in a successful scenario, can be framed by looking at the valuation of companies that own foundational internet layers. Cloudflare, a company that built a massive network for security and performance, currently holds a market capitalization of approximately $30 billion. A more direct, though smaller, comparable is Tailscale, a zero-config VPN and mesh networking provider, which was valued at over $1 billion in its 2024 Series B. If doxx.net successfully defines the ADN category and captures a material portion of the future AI-agent connectivity market, a multi-billion dollar outcome is a plausible scenario (scenario, not a forecast). The $38 million Series A provides the capital to attempt this category creation, with backing from a firm, Andreessen Horowitz, known for funding infrastructure bets with network effects [Business Wire, October 2026].

Lightly corroborated -- The core opportunity thesis is built on public product claims and founder history, but evidence of market traction or specific enterprise adoption is not yet available.

Sources

Public sources

1. [Business Wire, October 2026] doxx.net Raises $38 Million, Led by Andreessen Horowitz, to Launch Agentic Defined Networking | https://www.businesswire.com/news/home/20261001294246/en/doxx.net-Raises-$38-Million-Led-by-Andreessen-Horowitz-to-Launch-Agentic-Defined-Networking

2. [doxx.net] Agentic Defined Network - doxx.net | https://www.doxx.net/

3. [Network World, October 2026] Startup doxx.net hands the network controls to AI | https://www.networkworld.com/article/4230194/startup-doxx-net-hands-the-network-controls-to-ai.html

4. [a16z, October 2026] Investing in doxxnet | https://a16z.com/announcement/investing-in-doxxnet/

5. [Superpower Daily, October 2026] doxx.net Raises $38 Million and Opens Private Networks to ... | https://superpowerdaily.com/posts/doxx-net-raises-38-million-and-opens-private-networks-to-people-and-ai-agents

6. [Startup Researcher, October 2026] doxx.net Raises US$38 Million in Series A from Andreessen | https://www.startupresearcher.com/news/doxx-net-raises-usd38-million-series-a-led-by-andreessen-horowitz

7. [BigGo Finance] Why Online Privacy Is Harder Than You Think｜a16z | https://finance.biggo.com/podcast/fe719256c1dc9a34

8. [doxx.net About] Agentic Defined Network - doxx.net | https://www.doxx.net/about

9. [docs.doxx.net] doxx.net Documentation | https://docs.doxx.net/

10. [doxx.net Pricing] doxx.net Pricing | https://www.doxx.net/pricing

11. [TechCrunch, August 2013] Cyber Security Expert Barrett Lyon Brings Defense.net Out Of Stealth To Battle A New Wave Of DDoS Attacks | https://techcrunch.com/2013/08/06/defense-net/

12. [Refresh Miami, October 2026] doxx.net raises $38M to put AI agents on a shorter leash | https://refreshmiami.com/news/doxx-net-raises-38m-to-put-ai-agents-on-a-shorter-leash/