1 / 15

Selanet.ai

AI Agent Nodes as Web Infrastructure

2,700+ Nodes Settled

90+ Regions

30K TXs/day On Monad

Opportunity

AI Agents Are Exploding.
The Web Isn't Ready.

Every vertical is deploying AI agents — research, sales, support, logistics. But these agents need to access the real web: post-login data, dynamic UIs, multi-step actions.

Less than 10% of websites offer APIs. The other 90% is locked behind browsers built for humans.

$220B+ AI Agent Market by 2034

1.3B+ AI Agents by 2028 vs. millions in 2025 — Source: Microsoft

90% of web has no API

25–40% of agent cost is web access

Problem

Why AI Agents Fail
on the Real Web

AI agents access the web like bots — and that breaks at scale.

Web blocks AI traffic — Bans, CAPTCHAs, and costs explode at scale. GPTBot and ClaudeBot are frequently blocked.
Post-login blindness — Agents can't access data behind authentication — the most valuable data on the web.
Reliability collapses — Fragile automation, broken workflows, and data quality issues at scale.

Solution

Authenticated, Decentralized
Web Execution — Live Now

Selanet is a decentralized web execution network powered by real users' devices and logged-in sessions. Infrastructure is live across 90+ regions.

Authenticated & Decentralized — Real users' logged-in sessions. No centralized IP bottlenecks. Access post-login data that no other provider can reach.
Native to AI — Agent-native execution environment with structured data extraction. Open API/SDK for any agent framework.
Trustless & Verifiable — Cryptographic proof of execution via zk-TLS. Permissionless participation.

Architecture

How It Works

AI agents connect to the web through Selanet's distributed node network — real devices, real IPs, real logged-in sessions.

AI Agent / Client → Selanet Protocol → AI Agent Nodes → Web / Platforms

Real devices — Desktop, mobile, and data center nodes across 90+ regions
Full browser capabilities — Read, act, authenticate, and verify via zk-TLS
Sessions stay on-device — Credentials encrypted locally, per-action rate limiting

Use Cases

Same Question,
Different Result

ChatGPT stops at login walls, gets blocked by CAPTCHAs, returns shallow data. Selanet-powered agents execute behind authentication.

Upflux — Biz Lead Generation

"Find Korean companies that uses Salesforce"

AI Reasoning — Planning & decision-making
Real Web Access — LinkedIn via real user sessions
Verified Leads — Enriched from authenticated data

LinqAlpha — Investment Research

"Analyze trending cosmetics brands on Xiaohongshu"

Post-login Access — Xiaohongshu behind auth wall
Real-time Signals — Trend data for investment reports
Geo-specific Data — China-only platform, local sessions

AI can answer questions. Selanet lets agents execute behind login walls.

Traction

Live Infrastructure.
Real Pipeline.

8 LOIs $25K MRR Projected

25+ POC Pipeline

4.2s Median Latency

99%+ Success Rate

<1% Ban Rate

440 Developer Waitlist

Live Now

Browse API — $8/1K requests, production-ready
Node Client — 2,700+ nodes across 90+ regions
9,000+ social followers — growing developer community

Converting Now

8 LOI-signed clients — contracts in legal review
First invoices — scheduled this quarter
2 clients completing integration — live deployment in progress

Why We Win

Competitive Landscape

Others sell tools or access. Selanet builds the infrastructure layer AI depends on.

High (Large/Parallel)

Low (Local/Serial)

Centralized

Distributed
Human-like

AI Workload Scale

Execution Environment

Tavily

Browserbase

Browser Use

Playwright

Bright Data

Grass

Selanet

Tavily: '25 Seed $25m Browserbase: '25 Series-B $40m @ $300m Browser Use: '25 Seed $17m

Developer Tools (Playwright) → Don't scale parallel AI workloads

Centralized APIs (Tavily, Browserbase) → Centralized, failing at high scale workflows

Proxy/Bandwidth (Bright, Grass) → Not designed for agent native execution

Network Effect

The Flywheel

AI Agent Node breaks the DePIN growth ceiling — operators join for personal AI utility (social automation, work delegation), not just tokens.

Step 1

More Nodes

AI agent utility (social automation, work delegation) drives organic growth

Step 2

Richer Access

Broader geographic and platform coverage

Step 3

More Usage

Demand and transactions increase, BM diversifies

Step 4

Higher Revenue

B2B API fees compound with network scale

Step 5

Stronger Incentives

Network revenue reinvested into operator economics

Revenue

Business Model

Browse API — web execution as a service for AI agent builders.

$8 per 1,000 Requests

~60% Gross Margin

$25K Projected MRR

Pricing Ladder

Free Tier → PAYG → Subscription → Enterprise

Revenue Expansion

Usage Growth

Clients scale from pilot to production volumes

Platform Upsell

Free tier to enterprise SLA conversion

Network Effect

More nodes = broader coverage = higher-value use cases

Go-to-Market

Win Long-tail,
Then Scale Up

Start with agent-first startups, expand to platforms serving thousands of enterprise clients.

Segment 1 — Now

Long-tail Agent Startups

Agent-first SMBs needing post-login access and reliable web execution. PLG acquisition.

Segment 2 — Q3 2026

Vertical AI Builders

Companies adopting AI agents for core business growth and cost reduction.

Segment 3 — Q4 2026

Platform Partners

Infrastructure and marketplaces serving thousands of downstream enterprise clients.

Financials

Financial Projection

$78M Cumulative Revenue (3Y)

47.5x Revenue Multiple

1,910 B2B Clients

$1.3M

Year 1

$15.0M

Year 2

$61.7M

Year 3

Y1: 8 LOI clients converting at ~$3K avg MRR + new pipeline growth. Y2–3: node scaling to 100K+, $8/1K requests.
Benchmarked against Bright Data ($300M ARR).

Roadmap

Roadmap 2026–2030

Phase 1 — Foundation

AI Agent Node launch (April 2026)
LOI → paid conversion
2,000+ active nodes
B2C premium rollout
SDK/API for AI agents

Phase 2 — 2027–28

100K+ nodes across 150 countries
Diversified web/platform access
Vertical AI agent partnerships
Dedicated node device

Phase 3 — 2029–30

1M+ nodes — default web layer for AI agents
Open protocol standard for agentic web access
Market leadership as vertical infrastructure

Team

Team — BC Labs

Backed by SBVA (SoftBank Ventures Asia) and KB Investment.

Stephen

CEO

Mirae Asset Financial · Ilshin Ventures VC · Hewlett-Packard PM

Yonsei Univ. M.S. EE · Georgia Tech M.S. MoT

Spock

CTO

Oasis CEX Exchange CTO · Traport CTO · Alpine Research CTO

Soongsil Univ. Computer Science

Jayden

Advisor

Samsung Electronics LLM RAG Architect · Humax Embedded SW Developer

Sungkyunkwan Univ. Ph.D in MoT

Web3 Strategy Expert

CSO

Dunamu Investment Head of BD · Samsung Securities Credit Analyst · Nielsen Data Scientist

Columbia Univ. M.A. Statistics

SW Expert

Dean

SW Engineer

BC Labs. SW Architect · Oasis Exchange SW Developer

Hanyang Univ. M.S. Engineering

SW Expert

Vincent

Blockchain Engineer

Probit Global Full Stack Dev · DXM (Dunamu) Full Stack Dev

Univ of Philippines. Computer Science

Product Expert

Mindshift Inc. Product Manager · Findex Chain Product Manager

Stony Brook University (SUNY) Business

Business and Partnership

Carl

SOOHOIO Business Development · Fingerlabs Head of Blockchain Strategy · Crominnovation PM

Backed by

SBVA KB Investment

Investment

The Ask

$1M Seed Current Round

$20M FDV Fully Diluted Valuation

Use of Funds

Engineering & Node Infra

60%

Ecosystem & Agent Integrations

20%

BD & Partnerships

15%

Legal & Compliance

Infrastructure is done. Revenue is next.

Backed by SBVA (SoftBank Ventures Asia) & KB Investment

Selanet.ai

Selanet.ai

AI Agents Are Exploding.The Web Isn't Ready.

Why AI Agents Failon the Real Web

Authenticated, DecentralizedWeb Execution — Live Now

How It Works

Same Question,Different Result

Upflux — Biz Lead Generation

LinqAlpha — Investment Research

Live Infrastructure.Real Pipeline.

Live Now

Converting Now

Competitive Landscape

The Flywheel

Business Model

Pricing Ladder

Revenue Expansion

Win Long-tail,Then Scale Up

Segment 1 — Now

Segment 2 — Q3 2026

Segment 3 — Q4 2026

Financial Projection

Roadmap 2026–2030

Phase 1 — Foundation

Phase 2 — 2027–28

Phase 3 — 2029–30

Team — BC Labs

The Ask

Use of Funds

AI Agents Are Exploding.
The Web Isn't Ready.

Why AI Agents Fail
on the Real Web

Authenticated, Decentralized
Web Execution — Live Now

Same Question,
Different Result

Live Infrastructure.
Real Pipeline.

Win Long-tail,
Then Scale Up