☿ Kairos War Room
← Back to master state
Gated source note · private

outside-in-xray

Scope and method

Phase 1 outside-in X-ray of Mysterium Network (mysterium.network) for a 30-day consulting engagement on shortest path to break-even. €0 budget, public sources only. Every claim below carries a source URL and a confidence mark: [solid] (read directly from a primary source), [estimate] (derived/third-party, directionally useful but not verbatim-primary), [unknown] (could not determine — flagged, not guessed).

1. Product surfaces — what they actually sell, to whom, at what price

Mysterium operates as an umbrella protocol with several branded surfaces, not one storefront:

Walking the B2B buyer path as a stranger: starting at mysterium.network, a company wanting to buy proxy bandwidth is routed off-domain to goproxies.com — a separately branded, seemingly independently-operated proxy business that happens to source supply from Mysterium's node network. There is no self-serve API signup, published enterprise pricing, or sales contact visible on mysterium.network itself. Finding: the B2B funnel is not owned by Mysterium's own domain — it's outsourced/white-labeled through a partner brand, which is a friction point and a branding fragmentation problem. [solid], inferred from the partner-page description and press release cited above.

Negative finding — ⚠ SUPERSEDED 2026-07-18: with GoProxies confirmed in-house, this is wrong as a portfolio claim — GoProxies has a live self-serve dashboard + Buy-now checkout (card + crypto) on every tier (public-surface-sweep-2026-07-18). The surviving gap is that the mothership never links the door. Original text kept below for the record: there is no visible self-serve "buy proxy bandwidth now" flow anywhere on a Mysterium-owned domain. The path a cold B2B prospect would have to take is genuinely unclear and multi-hop (mysterium.network → partners page → goproxies.com → unclear whether goproxies.com is even under the same operating company). [solid] — this is a real gap, not a guess.

2. Consumer dVPN — app store presence, pricing vs. peers

Finding: Mysterium's dVPN is a real, rated, live consumer product with a nontrivial review base (5.1K on iOS alone) — not vaporware — but sits in a brutally commoditized, low-price VPN market against well-capitalized flat-rate incumbents. Its differentiation (residential IPs, decentralization, crypto payment) is real but the review friction (bans, refunds, inconsistent speed) suggests execution/support gaps, not positioning gaps.

3. MystNodes supply side

Finding: $568K cumulative all-time network earnings, spread across 31.5K IPs over presumably several years, implies a very thin average per-node lifetime payout (order of magnitude: tens of dollars per node, not hundreds) — a red flag for node-runner economics and supply retention, though this is an [estimate] extrapolation, not a stated figure.

4. Engineering signals — GitHub activity

Org: github.com/mysteriumnetwork, 110 repositories total. [solid] (https://github.com/orgs/mysteriumnetwork/repositories)

Most recently updated repos (sorted by push date, pulled 2026-07-05), all dVPN/node/client infrastructure:

Repo Last updated Focus
nightly Jul 5, 2026 nightly builds
node Jul 3, 2026 core dVPN node — 1.2k stars, 345 forks, 76 open issues, 8,471 commits, actively maintained, latest release v1.38.3 (May 18, 2026)
discovery Jul 3, 2026 (undescribed)
myst-launcher Jul 2, 2026 exit node launcher
AVADO-DNP-Mysterium-Server Jun 29, 2026 Mysterium server package
node-js Jun 29, 2026 —
mysterium-vpn-design Jun 25, 2026 —
mysterium-vpn-desktop Jun 24, 2026 desktop VPN client (Win/Mac/Linux)
mysterium-vpn-release Jun 22, 2026 release builds
wireguard_dart Jun 16, 2026 WireGuard SDK
myst-launcher-release May 27, 2026 launcher releases
mysterium-vpn-api-client-dart May 21, 2026 —
mysterium-mobile-provider May 5, 2026 mobile node provider
mysterium-vpn-mobile Apr 30, 2026 mobile VPN client (237 stars)
dvpn-web Apr 28, 2026 —

[solid] (https://github.com/orgs/mysteriumnetwork/repositories?q=&type=&language=&sort=updated, https://github.com/mysteriumnetwork/node)

Finding: commit activity is 100% skewed toward dVPN core/client/launcher infrastructure. Nothing in the top-20-most-recently-updated list touches B2B/proxy business logic, an agent-access product, MCP, or x402-style payment tooling. A targeted search for agent/MCP/x402-related work in the org found zero matching Mysterium repositories — those terms only surfaced unrelated third-party projects. [solid] negative finding (https://github.com/orgs/mysteriumnetwork/repositories). If "agent access" is a strategic priority, it is not yet visible in the public engineering backlog.

5. Team signals — LinkedIn, headcount, hiring

Finding — this is a strong negative signal for the engagement's B2B thesis: the only visible open req is another engineering hire, not sales/BD/marketing. Combined with a team page that publicly names zero salespeople, this is consistent evidence that Mysterium currently has no visible dedicated commercial/sales function — the org, at least in its public face, is still an engineering-led, product-led shop with no B2B go-to-market bench. [solid], directly observed.

6. Token/financial state — MYST

Finding: MYST is a micro-cap, thinly-traded, majority-off-tier-1-exchange token, down ~98% from its 2018 ATH, with no Tier-1 CEX listing (no Binance/KuCoin) — meaning the token itself carries very little liquidity or brand-recognition leverage for any go-to-market or fundraising strategy the engagement might consider.

7. Narrative surface — blog cadence, founder presence, conferences

⚠ SUPERSEDED (2026-07-18) — the "dormant narrative engine" finding below is FALSE as a portfolio claim. This section only checked mysterium.network/blog (genuinely dead) and missed the live engines: mysteriumvpn.com/blog/news publishes near-daily through July 2026, goproxies.com/blog runs original research with press pickup (Apr 2026), mystnodes.com/blog is thin-but-alive (Mar 2026). True problem = fragmentation, not silence. Corrected ground truth: public-surface-sweep-2026-07-18.

Finding: the company's owned-content narrative engine looks largely dormant. A ~3-year-stale primary blog, no confirmed 2026 founder interview, and no conference presence together suggest whatever public-facing storytelling is happening (if any) is not happening on channels this search could find — either it's genuinely quiet, or it's concentrated somewhere unindexed (e.g., private Discord/Telegram, which do show real member counts: 10.3K Discord, 5.9K Telegram).

8. The agent-access angle — public evidence of "access for AI agents"

Finding: the founder's stated priority ("we're implementing agent access") has zero corroborating public footprint as of 2026-07-05. This is either (a) genuinely unbuilt/pre-announcement — consistent with the sole open job req being another generalist fullstack engineer, not an AI/agent specialist — or (b) being built in a private repo/unlaunched branch not yet public. Either way, the consulting engagement should not assume any existing agent-access product, code, or GTM motion exists publicly — this would need to be verified internally on day one.

What only internal data can answer

Public sources could not resolve the following — these are exactly the questions the engagement should ask Mysterium directly in week one:

  1. Actual revenue and its composition — how much (if any) comes from goproxies.com/B2B proxy sales vs. consumer VPN subscriptions vs. node-network transaction fees vs. treasury/token operations. No revenue figure of any kind is public.
  2. Whether goproxies.com is majority-owned/controlled by Mysterium or a fully independent company that merely sources supply from Mysterium's nodes — the public partner language ("partnership") suggests independence, but the real equity/commercial relationship, revenue share, and who actually owns the B2B customer relationship is unknown.
  3. Treasury balance, burn rate, and runway — no published treasury report found. Given the ~25-person headcount and a token FDV of ~$2.9M, understanding actual operating cash (likely fiat/stablecoin reserves from past raises, not MYST-denominated) is essential to any break-even plan and is entirely opaque from outside.
  4. True node-runner unit economics — average earnings per active node per month, node churn/retention rate, and what fraction of the 31,547 "participating IPs" are actually active vs. dormant/inactive. The cumulative $568K all-time figure cannot be decomposed into a current run-rate from outside.
  5. Actual B2B pipeline — whether any sales conversations, enterprise contracts, or pilot customers exist today for proxy/agent-access data, since no sales function or funnel is visible publicly. Is there truly zero B2B revenue, or is it happening off-site (direct relationships, not through goproxies.com)?
  6. The real status of "AI agent access" — is there unreleased code, a signed pilot customer, or is this purely a founder narrative/roadmap item with no engineering investment yet? Public GitHub shows nothing; internal repos may differ.
  7. Team composition beyond the 2 named individuals — actual current headcount split by function (engineering vs. ops vs. support vs. any commercial staff), given the public team page names only Founder + Director and LinkedIn only gives a size band (11-50) plus a headcount estimate (25-28).
  8. Historical CAC/LTV and churn for the consumer VPN app — install base, paying-subscriber count, and retention are not disclosed by app stores beyond rating/review counts.
  9. MYST holder concentration — genuine top-holder/whale concentration (vs. the raw, likely airdrop-inflated holder counts found on Etherscan/Polygonscan) would need a wallet-cluster analysis not completed here.
  10. Why the primary blog has been dormant since ~2023 — deliberate strategic shift to Discord/Telegram/X, resourcing gap, or something else — worth asking directly since it materially affects how "known" this company already is to its own market.

Corrections & additions from Lee (2026-07-05, post-read)

Browser rendering of the War Room source library. The vault remains canonical.