wave-map

Companion to the Kairos — Research Dossier, market-map, and outside-in-xray. Raw, source-tagged, not client-polished. Confidence marks: [confirmed] = primary-sourced, [claimed] = self-report unverified, [inference] = reasoning, not fact. Built to test Lee's read that early MN rode two converging mega-waves (crypto/ICO + privacy/freedom), and to scan the horizon wider without locking in prematurely. Origin frame: "waves rippling on the horizon" (Shane Snow, Smartcuts — catch a rising wave rather than generate all the force yourself; the biggest wins pair two converging waves).

Part 1 — What actually carried early MN (four waves, not two)

Lee's "crypto × privacy" is directionally right but collapses four distinct waves into two — and the dropped one matters most for the forward question.

Wave 1 — Ethereum / smart-contract enabling substrate [confirmed]. MN's whitepaper is an explicitly "Ethereum-based VPN" with its own micropayment system (CORE, later Hermes). ERC-20 (2017) made tokenized, per-GB, trustless bandwidth technically buildable. The rails, not sentiment.

Wave 2 — The ICO capital wave [confirmed]. 2017 peak: ~$5.6B raised across the year; MN took ~$14M in 45 minutes. Tall, fast, burst by mid-2018. This is the one Lee named "crypto/ICO."

Wave 3 — Privacy / surveillance backlash (demand) [confirmed]. Snowden (2013) → net-neutrality repeal (Dec 2017, VPN subs doubled in two weeks) → Cambridge Analytica + GDPR (May 2018). Slow, structural, still running. Lee's "privacy/freedom."

Wave 4 — the missed one: tokenize-idle-resources / decentralized sharing-economy — proto-DePIN [confirmed]. 2014 storage (Filecoin, Storj, Sia) → 2017 compute (Akash, Livepeer, Render). MN was the bandwidth instance of that wave. The term "DePIN" didn't exist yet (Messari/Multicoin coined it ~2022) but the wave did.

Why it matters: Wave 4 is a structural category wave, and it is cresting again, far bigger — DePIN ~$10B → projected $32–35B in 2026. MN was early to a wave that returned at ~3× scale.

The Smartcuts lesson in the timing [inference, well-grounded]: MN caught the two tall-fast waves (capital + tech) at their peak — perfect for raising $14M in 45 minutes. But the two waves that sustain a company — privacy demand and resource-sharing/DePIN — are slow-structural and didn't crest until ~2023–26. The fast waves broke in 2018; MN ran out of momentum in the trough before the slow waves arrived. Funded by the tall-short wave, but needing the slow-structural wave to mature. That mismatch — not bad execution — is why the early lead didn't compound.

The forward filter this hands us: prefer a convergence where at least one wave is slow-structural (so you survive to the crest), paired with one tall-fast wave (for capital/momentum). It's the pattern that built MN and the pattern that would have saved it.

Part 2 — The horizon, wider (2026 wave board)

Convergence zones (two rising waves crossing — the MN-origin pattern forward)

• CZ-1 — Agent access × Agent payments. AI-agents (tall-fast) × machine-payments (rising). MN touches both sides (residential egress + live micropayment rail). Weak slow-structural anchor, tightest grip-fit. Strongest commercial convergence.
• CZ-2 — Proof-of-personhood × Privacy/sovereignty. A network of consented real residential humans as the "verified-human / real-origin" layer as the internet drowns in AI slop. Crosses two rising waves and sits on MN's original slow-structural wave (privacy) — the one that never broke. Passes the Smartcuts filter cleanly. Build-heavy (attestation doesn't exist; consent unverified), most on-mission, closest to Lee's Sovereign-Creator brand gravity. Highest strategic coherence.
• CZ-3 — DeAI × DePIN-bandwidth. Decentralized web-data/access, picks-and-shovels for AI (the Grass lane). Real wave, weak position (crowded, ~100× subscale).

Wildcards (low-probability, high-payoff, non-obvious)

• WC-1 — Federated-learning / distributed-private-compute substrate. MN's nodes aren't only IPs/bandwidth — they're consented edge devices with a rail to pay them = a federated-learning / edge-inference substrate (train without centralizing data; rides the EU-AI-Act mandate). Reframes the asset from egress to compute fabric. Caveat: light-compute nodes, no ML orchestration — shape fits, capability must be built. [inference]
• WC-2 — Grant-funded censorship-resistance (a different money model). MN's original mission is fundable by grants, not customers. The Open Technology Fund's Surge & Sustain Fund bankrolls circumvention tools reaching 46M users/mo in China, Iran, Russia; awards $10k–$900k; for-profits eligible. MN is a decentralized censorship-resistant network — monetizes (partially) what it already is, a buyer nobody in the dossier considered, and the splinternet wave grows the pool. Caveats: USAGM-adjacent (geopolitical strings, volatility), sub-venture scale ($900k cap). Best as a non-dilutive, mission-pure runway line, not a business — while choosing the real commercial swing. [confirmed fund + terms]

The transport-layer gap (HAPP), corrected. MN can't credibly serve this niche as-is: its CEO (Ro) reaches the western internet from Russia via HAPP, not Mysterium, because MN's OpenVPN/WireGuard-class transport is fingerprinted and killed by Russia's TSPU in seconds, while HAPP's VLESS+Reality protocol impersonates real TLS (microsoft.com/Cloudflare) and gets ~98–99% through [confirmed mechanism; Ro's case = strong signal, n=1]. So entering WC-2 requires a transport build (integrate Reality/XTLS into the node protocol), not just a go-to-market. But do not over-read this — HAPP and the entire Xray/Reality client layer are free, donation-funded, explicitly non-monetized [confirmed]. The layer that wins censorship is the layer that's given away; the money is in servers/networks/distribution — MN's existing layer. Reality transport is therefore a niche-unlock feature + credibility/PR play, never a business model. The seductive move ("pivot to the tech that actually works") is a trap: the tech that works can't be sold. It also does not explain why MN failed to compound — that's a commercialization failure on a proven-lucrative layer (Nord/Surfshark/Bright Data/Oxylabs won the same layer on GTM/brand/capital, several in Vilnius), not a wrong-layer or wrong-thesis problem. The one payoff worth the transport build: it would let MN offer what HAPP can't — Reality stealth + decentralized residential-exit IPs + token-paid nodes, a combination nobody sells. - WC-3 — Private egress for personal AI agents (consumer, not enterprise). The agent-infra market chases enterprise scraping (Browserbase, Anchor, Firecrawl). Unserved: personal agents needing a private, residential, non-datacenter way to act on the web for their user. Privacy-wave × agent-wave on the B2C side — MN's actual brand DNA (consumer VPN). Inverts who everyone sells to. [inference] - WC-4 — Real-world provenance oracles (furthest out). Nodes as distributed real-world witnesses attesting "observed from a real place at a real time" for the AI-provenance stack. Tallest AI-integrity wave; capability doesn't exist. Flagged for direction, not reachability. [inference]

Synthesis

Tightest commercial bar: CZ-1 (agents, both sides). Highest coherence, and the one that passes MN's own historical filter (fast wave for momentum + slow-structural wave to survive on): CZ-2 (verified-human / provenance on the privacy wave that never broke). Sleeper: WC-2 (a funder that pays MN to be what it already is). Deliberately excluded: GPU/compute DePIN (closed, no supply) and wholesale bandwidth resale (subscale trap).

Sources

newsbtc (MN whitepaper), Medium 2017-ICO retrospective, CoinMarketCap DePIN lineage, Recast privacy-backlash timeline, Digiday net-neutrality/VPN boom, Orochi & BlockEden DePIN 2026, Laika Labs & KuCoin 2026 crypto narratives, Biometric Update proof-of-personhood, CircleID internet-governance 2026, Open Technology Fund Surge & Sustain Fund. (URLs in session transcript 2026-07-17.)