Kairos — Meeting Prep (2026-07-04)

Prep for tomorrow's meeting with Šaras and Ro (Robertas) re: the 2-person "SWAT team / special ops" venture hunting disproportionate, asymmetric ROI for Mysterium Network. Šaras is a trusted friend (passes the Naval filter). Context in 2026-07-03 — MOL — Post-Vitotas pitch, Šaras SWAT, Mantas fallback.

What Mysterium actually is (follow the money)

Decentralized VPN + residential-proxy / bandwidth marketplace. Two revenue engines, wildly different economics:

1. Consumer dVPN — commodity, low margin, competing in a ~$77B VPN market against NordVPN/Express. A race to the bottom. Probably NOT where the asymmetry is.
2. B2B residential-proxy / data-scraping supply — real people's residential IPs resold to businesses; residential IPs command a premium over datacenter; network takes a 20% service fee; region-based supply/demand pricing. This is the margin engine.

Plus a third dimension: MYST is a token — so tokenomics/market narrative is a lever the SWAT team could pull that a normal company can't.

Robertas' "margins low, profits trickling" almost certainly means the B2B demand side is under-sold, not that supply is short. In marketplaces, when profit trickles, the bottleneck is usually demand/GTM or pricing — not inventory.

Where the asymmetric ROI likely lives (angles to bring)

⚡ Freshest frame (2026, verified 07-03) — lead with this. The open web is closing to AI: Cloudflare blocks AI bots by default + Pay-Per-Crawl (402 paywall), and from Sept 15, 2026 blocks "Agent"/"Training" bots on ad pages. Tailwind: scarcer access makes a legitimate agent-access layer more valuable — Mysterium's problem gets bigger, not smaller. Trap: Cloudflare now re-flags raw residential IPs (JA4 + ASN signals; mobile survives better), so "sell more residential" is the losing pitch — the money is the consented full stack + a niche. One line for the room: "The web is closing to AI — raw residential is the commodity getting blocked; we build the access layer on top."

• Ride the AI-training-data wave. The single biggest new buyer of residential proxy right now is AI companies scraping the web for training/eval data at massive scale. If Mysterium isn't aggressively selling into that buyer, that's the disproportionate opportunity — a demand tsunami meeting an under-marketed supply.
• Narrative as a moat (Lee's lane) — but Grass already owns the obvious version. ⚠ The dossier is blunt: Grass (~$304M cap, ~100× Mysterium, AI-data-native) is already claiming the exact "ethical, consented supply" story and out-executed on the AI-data play for two years. So do NOT walk in pitching "let's own the ethical-supply narrative" — the sharp reply is "Grass owns it and they're 100× us." Bring a differentiated wedge instead: (a) Mysterium's 9-year VPN-privacy brand + Lithuania as the literal global capital of this industry (Oxylabs, IPRoyal, Decodo, Tesonet next door) as an enterprise-trust / local-talent angle Grass can't claim; or (b) a specific vertical — compliance-sensitive EU buyers who need GDPR-clean consented supply and won't touch a US crypto-bro brand. Name Grass first, before Ro does.
• Pricing / take-rate optimization — the 20% fee and regional algorithm are zero-cost levers. Repricing premium geographies (US/UK/DE residential) is pure margin if under-priced.
• Token narrative flywheel — tie B2B revenue growth to a visible MYST utility/burn story; the crypto-market re-rate can dwarf operational gains. High-variance, but that's the definition of asymmetric.

Lead bet — bring ONE, not the menu (this is for Ieva)

Ro is the warm room tomorrow, but the real judge is Ieva — a Coca-Cola / MV Group P&L operator. She will NOT approve €4-6k/mo cash for a "creative generalist" on a four-angle menu. She wants: which angle, what measurable 90-day outcome, what number, why you. So pick your strongest angle and walk in with it framed as one testable 90-day bet with a metric attached, e.g.:

"In 90 days I stand up an AI-data-buyer outbound motion (or a compliance-EU repositioning) and deliver N qualified enterprise conversations / €X pipeline — proving the demand side is the bottleneck, not supply."

The menu above is your reserve — the depth you draw on if they push. But you sell Ro one concrete bet he can carry upward to Ieva. Auditioning for her through him is the actual game.

Decision structure (who actually greenlights)

• Ro (Robertas Visinskis) — founder, Lee's friend, the champion. Invited Lee. Warm, but may not control day-to-day budget.
• Ieva Matulaitienė — operational CEO / vadovė (Šaras: "vadovė, Ieva"). Commercial background: The Coca-Cola Company, MV GROUP (big LT beverage group), Global Lithuanian Leaders; gave 2024 DPN interviews. She is the budget gate — a P&L / brand / go-to-market operator, NOT a crypto engineer.
• Šaras — internal advocate, joined Mysterium ~Dec 2025; brought Lee in.

The sharp read: tomorrow is Ro + Šaras (warm, wants it), but the initiative is NOT yet approved — Ieva can kill it on budget (ramble: "might be shot down by the CEO... or escalate to Robertas"). A great meeting with Ro doesn't secure it; the SWAT case must survive Ieva's business scrutiny. GOOD for Lee: Ieva speaks revenue / margin / market / brand — Lee's native tongue — not tokenomics (his weak spot). Frame the whole thing as a business case and win on her turf, not on crypto depth. Extra question for Šaras: "Where does Ieva stand — is the budget hers, is she already bought in, or do we still need to make the case to her?"

Questions to ask them (so you shape the room)

• What does "disproportionate ROI" mean concretely to you — revenue, MYST price, B2B contracts, user growth? Pin the target before hunting.
• Current revenue split: consumer dVPN vs. B2B proxy? Follow the money.
• Where do you think the bottleneck is — supply, demand, pricing, or narrative?
• Who are today's B2B buyers, and is anyone chasing AI-data / scraping buyers specifically?
• Is this approved, or still pending CEO sign-off? What decision are you making, by when?
• Structure: paid, equity, token, or rev-share? (See terms below.)

Your terms — decided: this is a runway play

Resolved (2026-07-03): Lee is at ~zero cash now — the reason to do this is runway, plain and simple. If he had money he'd go all-in on his own hubs; this is a forced, temporary detour. But one he genuinely enjoys — working with Šaras has historically been a creative delight. That sets the terms:

⚠ The paradox to resolve IN THE ROOM (highest-priority item): you're seeking runway security from a company that may have none. Per the dossier — $2.9M cap, token −60%/yr, zero disclosed revenue in 9 years, likely on ICO-treasury fumes — Mysterium is arguably the worst-positioned company to be a reliable monthly fiat payer. The retainer is only "runway" if it lands in your account on the 1st, in euros — not MYST, not "once we close the next deal." So before rate: establish they can actually pay. - First-month advance in fiat, in hand, before real work. Non-negotiable at zero liquidity. - Never more than ~2-4 weeks of unpaid exposure at any point (bi-weekly/milestone, not net-30). - Direct probe: "For a 3-month engagement — is this funded from operating cash or from treasury/token? I need the retainer stable before I commit hours away from my own build." Their answer tells you lifeline vs. trap. - Floor / walk-away: know your minimum monthly fiat + advance below which you don't start, BEFORE you sit down. Get them to name their budget first — but a warm room from a cash-strapped friend can't be allowed to drift you into working on spec. If they can't pay reliable fiat, the bet fails at both ends (no runway AND no hub progress) — walk.

• Secure the CASH BASE first — do NOT gamble runway on backend. At zero liquidity, the priority is a market-rate monthly retainer, paid regularly, with a first-month advance on signing. Results/ROI bonus layered on top as the upside (not greedy, but captures wins). Token/MYST/equity only as a sweetener, never instead of cash — illiquid upside doesn't feed you this month. (This corrects an earlier draft's "asymmetric-upside-not-hourly" take — that's a solvent person's move; at zero runway you secure the base first.)
• Price at market, unapologetically. Get what the market allows for a 3-month strategic/growth/BD engagement. Not greedy — worth it. Do not discount because it's a friend's company; undercharging helps no one and breeds quiet resentment.
• Cap time + scope. 3 months, defined deliverables, hours/week fixed so it can't cannibalize the €15k/mo-by-Oct hub build.
• Šaras = trusted (Naval filter passes) and a genuine creative delight to work with. Robertas is more of an unknown to you — read him in the room; the Vytautas beef is his, stay neutral.

Market rate (PRIVATE — never show them)

Raw comp (US fractional growth/BD leaders, 2026): $5–20k/mo retainer, avg ~$10–12k, or $200–500/hr. Climbs with proven track record of scaling comparable companies + hours committed.

Adjusted to reality: those are US SaaS rates; Mysterium is a European, cash-strapped crypto firm (crypto pays nearer global/dollar rates than local LT — good; but a struggling co can't pay top-of-band in cash — hence the bonus). Lee = strategic/creative/BD generalist with a conceptual edge, not a proven crypto-scaler → mid-band on pure comp.

Defensible anchor: part-time ~15–25 hrs/wk → base €4–6k/mo, scaling with hours, plus results bonus on defined ROI. Anchor at the top, expect to settle mid. Cash base; token/MYST only as sweetener.

Biggest swing factor = hours/week — undefined. Pin it in the meeting first. Tactic: get them to name their allocated budget range first; use this data as your check against a lowball, not the number you blurt. (Caveat: only hard data is US fractional-CMO comps — a reasoned triangulation, not a quote.)

The honest flag (radical truth)

Named and accepted: this is 3 months hunting ROI for someone else's struggling crypto company while the north star is €15k/mo net by October on the three hubs. The distraction risk is real — the mitigation is the time/scope cap above and treating it strictly as runway, not a new identity. It clears the bar because it serves (a) immediate cash/runway and (d) it energizes rather than drains (Šaras). Forced detour, genuinely enjoyable, capped. All three true at once.

Supply-side thesis: consented mobile / FWA (⚠ RESEARCHED 2026-07-03 — LARGELY DEFLATED)

Verdict after research: probably NOT the play. Do not lean on it in the room — carry it only as a rigorous question. Three hits from the data: 1. The trust premium may not apply to FWA at all. Mobile's value = carrier-grade NAT (1000s of subscribers per IP = collateral-damage shield). FWA is a single-household router box → likely classifies as ordinary residential, giving zero premium. No vendor has published data either way. Payoff = 0% confirmed. 2. "Consented" ≠ "compliant." Every carrier checked (T-Mobile, Verizon, AT&T, Vodafone, EE) categorically bans proxy/server/bandwidth-sharing use. Low-volume decentralised traffic = detection-evasion, not compliance. This guts the ethical/compliant-supply moat — user consents, carrier forbids. 3. Market voted: Bright Data (category leader) sunset its mobile-proxy product for new customers (April 2026). The biggest player retreated from this exact gap.

What survives: (a) geography is fine — FWA is now ~70% metro/suburban US, overlapping premium demand; (b) the killer test is cheap and Mysterium can run it near-free on its own existing FWA nodes — classify them via IPinfo/Spur/MaxMind + a live Cloudflare target; do FWA IPs read as mobile or plain residential? Carry as a question: "Have we measured whether our FWA/mobile nodes get mobile-grade trust or read as residential? And how do we square 'consented' supply with carrier ToS that forbid it?"

Original hypothesis (kept for the reasoning, now known to be mostly wrong): carry as questions, not claims (two ears, one mouth).

Integrated research read (2026-07-03) — demand pass, the important one

The money is UP the stack, not in supply. Raw IPs are commodity (residential $1-10/GB, mobile $15-30/GB, price-shopped) — Mysterium's current low-margin position. The 10-30× markup lives in managed unblocking (Bright Data Web Unlocker bills per successful page, not per GB) and the emerging agent trust/identity layer. Selling better supply = competing in the basement.

Starving crowd, ranked: 1. Agent trust/identity layer — max desperation + margin (agent traffic >57% of web; Amazon won an order blocking Perplexity's agent; Cloudflare Sept-15-2026 default-block; Air Canada agent misrebooked 1,247). But Cloudflare/Visa/Mastercard/Google already racing it → lowest reachability, not a green field. 2. Compliant, audited, SLA-backed access for a regulated vertical (EU finance/healthcare) — most reachable challenger wedge. Pain: sub-$50K-ACV buyers can't justify building compliant pipelines → must outsource. Bright Data owns enterprise trust broadly but no focused vertical. 3. Bulk scraping — confirmed dead (thin, price-shopped).

Implication for Mysterium (straight): its asset is the commodity bottom; the money is a layer up it would have to build (managed/compliant vertical product) — a real lift for a flat, 25-person, micro-cap, supply-first company. So the asymmetric-ROI question = "move up-stack into managed compliant vertical access — and can we execute?" A positioning/GTM problem (Lee's lane, Ieva's language), not crypto engineering.

Two killed assumptions (don't carry them in): (a) "consented sourcing = price premium" — WRONG, it's a gate not a premium; value = bundling compliance+reliability+audit into one SKU. (b) "US-crypto-bro brand won't fly in EU finance" — UNSUPPORTED; Bright Data (US/Israeli) IS the enterprise default. Drop it.

Tension tying both passes: to sell "compliant access" the supply must actually BE compliant — mobile/FWA breaches carrier ToS. Compliance must be real at the node level (cleaner on home-broadband nodes), not marketed.

Carry as questions: "Where do we play on the stack — supply, unblocking, or trust layer? Move up into managed compliant vertical access? What does Ieva count as 'asymmetric ROI'?" Then listen.

The convergence (the actually-interesting bit): the premium demand and the viable supply are the same thing. - The desperate, high-margin buyer is NOT the bulk scraper (commodity, thin margin). It's the AI-agent company whose product dies when blocked — agents logging in, booking, transacting. Blocked = product broken = burning platform = flush and willing to pay. (This IS Šaras's "tools for AI agents" mandate — demand-side, as briefed.) - That traffic is low-volume, human-paced — which is exactly the traffic that does NOT trip carrier fair-use throttling or "no-server" ToS. Bulk scraping breaks economics + ToS; agent access threads both. - So the "engineering-around" for carrier terms isn't a hack — it's the decentralised structure itself: thousands of consented nodes each carrying a trickle of gentle agent traffic stays under every threshold and never looks like a proxy. It self-selects for the premium use case.

The supply idea: mobile/cellular IPs are the top trust tier (carrier-grade NAT = thousands of real users behind one IP = blocking = mass collateral damage). Hard to source because phones are bad nodes (battery/OS) and consumer mobile has ToS/throttling limits. Fixed Wireless Access (5G home internet) may thread it: an always-on, mains-powered router on a carrier CGNAT IP — mobile-grade trust, residential-grade reliability.

Four load-bearing questions (being researched now): 1. FWA scale & geography — and does it overlap premium (US) IP demand? 2. Carrier ToS — how universally is proxying/bandwidth-sharing forbidden; does any plan permit it; does volume matter to enforcement? 3. IP classification — do FWA IPs get mobile-grade anti-bot trust, or are they flagged as their own category? (empirical crux) 4. Does anyone run consented mobile supply at scale — and if not, why not?

Caveats (my errors — don't repeat them in the room): (a) NOT a per-GB-cost story — unlimited mobile is common (Lee owns one); the real barrier is ToS/throttling. (b) NOT a local/Baltics play — the Baltics have world-class cheap fibre, so FWA is thin there; FWA lives in US-rural + emerging markets. (c) Mysterium's perennial failure is supply-recruitment execution, not ideas — a thesis they can't execute is worth little.

How to wield it: as a question — "What's our supply mix — any fixed-wireless / 5G-home nodes? And who's the most desperate premium buyer for un-blockable agent access?" Listen. (Findings will update this section.)

Next session — resume here (handoff, 2026-07-03)

Goal: Mysterium SWAT-team meeting with Ro (Robertas Visinskis, founder) + Šaras, ~2026-07-04. Lee goes in to ask and listen — two ears, one mouth — NOT to pitch. Confirmed: Lee's "Ro" = the founder profiled in the dossier; Ieva Matulaitienė (ex-Coca-Cola/MV Group) is the operational CEO and budget gate.

State — fully prepped. This note + Kairos — Research Dossier are complete. Key conclusions to carry: - The FWA/consented-mobile supply thesis is largely deflated (see that section) — don't lean on it; carry only as a cheap-to-test question. - The demand "starving crowd" read: the money is UP the stack (managed unblocking + agent trust/identity layer), not in raw supply. Most reachable challenger wedge = compliant, audited, SLA-backed access for a regulated vertical (EU finance/healthcare). The asymmetric-ROI question is "move up-stack — and can Mysterium execute?" — a GTM/positioning problem (Lee's lane, Ieva's language). - Terms: runway play, cash-first. Establish they can pay (fiat, advance on signing) before rate. Mysterium is cash-fragile (micro-cap, no disclosed revenue).

Next steps (post-meeting): 1. Capture what Ro/Šaras/Ieva actually said — answers to the 6 questions, their ROI definition, budget & who signs off, where they want to play on the stack, whether the agent-access mandate is product or supply. 2. Decide go/no-go against the terms (reliable fiat? energizing? capped?). 3. If go: build the ONE testable 90-day bet with a metric, framed for Ieva (revenue/pipeline, not crypto).

Open thread (unrelated): Lee's dictation setup still needs fixing — FreeFlow's mic is recording dead silence; recommended path is record-on-device (RØDE on-board or Voice Memos) then Claude transcribes via Groq.