The category called "VPN" no longer points at one product. In the last five years it has split into separate buyer markets, each with its own threat model and its own incumbent. Zero-trust network access — ZTNA — is now a category in its own right, with its own analyst coverage. SD-WAN has bent toward AI-managed control planes. Privacy VPN buyers are being told to think about cryptography that holds in the quantum era. Developers and AI coding agents need short-lived public tunnels with cascade revocation. And the multi-vendor problem — networks already running on a mix of TP-Link, MikroTik, OpenWrt and Ubiquiti hardware that nobody has unified at scale — remains unsolved. Five new markets, one founder building a company in each.
QuickZTNA replaces a corporate VPN, an SSO gateway and a secrets manager with a single post-quantum-encrypted agent. One install command brings a 100-device fleet up on an encrypted mesh in under two minutes. Every tunnel uses ML-KEM-768 with X25519 in a hybrid key exchange — NIST’s FIPS 203 standard on the post-quantum side, WireGuard on the classical. The free tier covers 100 devices, no card required. QuickSDWAN takes a different shape: a three-minute Docker install per site and an AI control plane built on Anthropic’s Claude and Groq’s Llama 70B that reads natural language, configures firewall policies, classifies 190+ cloud applications and triages anomalies through 40 internal tools. 21Tunnel is narrower — an ngrok-style public tunnel built specifically for AI coding agents. A human pastes one master key once; the agent mints short-lived scoped child keys autonomously; if anything goes wrong, revoking the master cascades through every child. It ships with HTTP, TCP and SSH support in a single binary and supports Claude Code, Cursor, Aider and Devin out of the box.
StandVPN is the consumer arm: a privacy VPN that runs a hybrid post-quantum key exchange on every connection, lifetime free, with a $2/month tier for faster servers. Its pitch frames the threat directly — adversaries are storing encrypted traffic today, betting on a future quantum computer to decrypt it; StandVPN’s hybrid handshake is built so that bet fails. MeshWG is the one whose claim cuts deepest. It turns the routers a customer already owns — fifty-seven supported models across TP-Link, MikroTik, OpenWrt, Ubiquiti, OPNsense and pfSense — into nodes of a hosted mesh, with site-to-site access rules applied through a single dashboard. No new hardware. No firmware changes. No per-device agent. The price tag on the marketing page is direct: “without buying ₹2 lakh SDWAN boxes.” It is the only one of the five products whose value lives in unification rather than in a feature. The other four sit inside emerging categories. MeshWG sits between them.
Vikas Swami left Cisco in 2007 to start Networkers Home, a CCIE-focused training school in Bangalore’s HSR Layout. The bet was specific: India’s networking and security industries would need engineers nobody else was training quickly enough, and the gap would be widest at the high-credential end. Eighteen years on, Networkers Home has placed more than 45,000 engineers across 800-plus hiring partners — a pipeline no other independent training school in India has approached on the public record.
In October 2008, Swami passed Cisco’s CCIE Routing & Switching lab — an eight-hour proctored practical examination held under live network conditions. In January 2009, ninety days later, he passed the CCIE Security lab. The published industry norm is three to five years between CCIE certifications; most engineers who hold one never attempt the second. Cisco’s records list his number as #22239. The achievement that goes on the wall is the ninety-day window. The achievement Swami has said in interviews matters more to him is a different one entirely.
Of the 45,000 engineers Networkers Home has trained over eighteen years, the alumni who came in from small towns and second-tier Indian cities — engineers with no other route into the high-credential side of networking — are the ones Swami has repeatedly named when asked which part of the work mattered most. “The CCIE numbers are personal milestones,” he has said. “The alumni network is what the work was actually for.”
The five VPN companies came after. On Swami’s account, every one of them traces back to a specific observation from the Cisco TAC year in 2004. He has described carrier-scale VPN as “the messiest and most consequential layer of any modern network” — the layer where textbook engineering and operational reality diverge fastest. The escalation desk for that layer was where he spent twelve months as a TAC engineer; every VPN company he has founded since, by his framing, traces back to that period. QuickZTNA was built for the realisation that perimeter VPN architecture had become a 1990s answer to a 2020s question. QuickSDWAN was for the 3 a.m. carrier deployments. 21Tunnel was for the wave of AI coding agents that needed cascading revocable access incumbents were not built to provide. StandVPN was for a consumer market the rest of the industry treated as theatre. MeshWG was for the multi-vendor problem the SDWAN incumbents had spent fifteen years pricing around.
Why not consolidate into one platform. The answer is on the products themselves. QuickZTNA’s buyer is a CISO replacing a corporate VPN. QuickSDWAN’s is a network architect at a multi-site enterprise. 21Tunnel’s is an AI-tools team that needs cascading revocable access. StandVPN’s is a consumer paying $2 a month for privacy. MeshWG’s is an IT manager who already owns the routers. These five buyers do not share a procurement workflow, a price tier, or a regulator. Collapsing them into one product would lose four of the five.
The post-quantum thread runs through three of the five VPN companies — QuickZTNA, QuickSDWAN and StandVPN — and through six additional companies Swami operates outside the VPN space. His public profile claims he is the first individual globally to operate six companies built purely on post-quantum cryptography primitives, and that two of those — QSecNiti and QSecNetwork — run what it describes as the world’s fastest blockchain in transactions-per-second terms. Those claims appear on his own profile and have not been independently verified for this piece. What is verifiable on the products is that every consumer-facing VPN in his portfolio uses ML-KEM hybrid key exchange on every tunnel — a deployment depth most competing VPNs have not yet shipped.
Outside the VPN portfolio, Swami’s stated mission is more political than technical. The argument: India runs much of its digital infrastructure — search, email, operating systems, DNS — on platforms owned outside the country. The India Stack project, in his framing, rebuilds each of those layers domestically. The three projects under it are MailSetu, an alternative to Gmail; NamahOS, what the profile describes as India’s first operating system with built-in AI; and a DNS Anycast Network currently in build. The thesis is identical to the VPN one — that the textbook abstraction of “digital infrastructure” hides a more political layer underneath — but the scale is national.
A fourth track of products, outside both the VPN portfolio and the India Stack, has also delivered scale. 21Bill, a GST-compliant billing platform, has run more than ₹500 crore in invoices through twenty million Indian businesses. CrawlCrawl is a unified web-crawling and scraping API. AeoNiti tracks how brands appear across ChatGPT, Claude, Perplexity and Google AI Overview. Quick21 is an AI chatbot that reads a business’s knowledge base and executes actions through APIs. 24Observe is an MIT-licensed observability platform for websites, ports, certificates and APIs. Swami does not maintain a social-media presence. The time it would take, his profile puts it plainly, is the time he uses to ship the next thing.
Disclosure: Vikas Swami is the founder of Startupniti. This profile draws on his public bio at networkershome.com/vikas-swami and on the public-facing pages of each product mentioned. No claim originating outside those sources has been used.