hermes-brain/projects/passepartout/strategy/game-theory.org

Game Theory

• Game 1: Enterprise Adoption (coordination with network effects)
• Game 2: Regulator Commitment
• Game 3: Hyperscaler Defense
• Game 4: Social Protocol Platform Competition
• Game 5: Geopolitical Competition
• The combined game: why the trajectory is stable
• Failure modes as parameter shifts

Why the adoption trajectory is structural, not just plausible. This document models the key strategic interactions as formal games with payoff matrices and Nash equilibria. Phase numbers refer to the Adoption document.

The central claim: the trajectory toward verified computing is the unique stable equilibrium of a multi-player game, not one of several possible outcomes. This document models the five core games — enterprise adoption, regulator commitment, hyperscaler defense, social protocol platform competition, and geopolitical competition between nation states — and shows that each has a unique Nash equilibrium that reinforces the others.

Game 1: Enterprise Adoption (coordination with network effects)

Two firms in a regulated industry choose whether to adopt a gate or maintain their current compliance process. Their payoffs depend on what the other firm does and on the regulator's stance (set independently in Game 2).

Strategy sets: {Adopt Gate, Maintain Current Audit}

Payoff parameters:

Let:

• C = annual compliance cost under current audit ($200K-$1M)
• G = annual gate subscription cost ($50K)
• R = risk of audit failure (probability × penalty, ~$500K/yr expected)
• D = competitive disadvantage if rival adopts gate and you don't (productivity gap, estimated 2-5% of compliance-adjacent revenue)
• N = network effect benefit if both adopt (shared gate rules, interoperable proofs, collective regression suite)

All values are per-firm per-year.

Payoff matrix (firm A row, firm B column):

	B adopts gate	B maintains audit
A adopts gate	(-G + N, -G + N)	(-G, -D)
A maintains audit	(-D, -G)	(-C - R, -C - R)

Nash equilibria:

1. Both adopt gate is a Nash equilibrium if G - N < C + R. Since G ≈ $50K, C ≈ $200K-$1M, R ≈ $500K, and N is positive, this inequality holds strongly — adopting the gate strictly dominates maintaining the audit when both firms adopt.
2. Both maintain audit is also a Nash equilibrium if D < C + R - G. If the competitive disadvantage of being the only non-adopter is small, both firms can get stuck in the current audit even though adoption would make both better off. This is a classic coordination failure.

What breaks the coordination failure:

The two equilibria exist because each firm's payoff from adopting depends on the other firm's choice. But the game changes when:

• The regulator encodes a rule as a gate. This removes the audit option for both firms — adoption becomes mandatory. The game collapses to a single outcome.
• Insurance differentiates on verification. The compliance cost C becomes C + I (insurance premium surcharge for unverified), where I can be 2-10x of C. This pushes D < C + R + I - G even for the first adopter, making unilateral adoption dominant.
• One firm has a large enough D that they adopt unilaterally. If the competitive disadvantage of being the only non-adopter exceeds the adoption cost, the first firm switches, and the second firm follows because -G + N > -D. This is how the flywheel starts.

Key insight: The coordination failure exists only in the absence of regulatory mandate or insurance differentiation. Both of these are themselves equilibrium outcomes of separate games (Game 2 and Game 3 interaction). The enterprise adoption game does not create a structural barrier — it creates a timing problem that the other games resolve.

Game 2: Regulator Commitment

A regulator chooses how to enforce a rule. An industry of N firms has already chosen their adoption status (partly determined by Game 1).

Strategy sets:

• Regulator: {Encode as gate, Maintain paper-based enforcement}
• Fate chooses the political environment: {Pro-gate, Anti-gate}

Payoff parameters:

Let:

• E = enforcement effectiveness (paper: 0.3, gate: 0.95 on a 0-1 scale)
• L = legitimacy cost of choosing the "wrong" approach (0 if aligned with political environment, k > 0 if misaligned)
• T = political turnover risk (probability that the next administration reverses the decision)
• B = benefit to regulator of career-defining legacy (gate: +M, paper: 0)
• S = surveillance-value loss to the state when data becomes invisible to bulk collection (paper: 0, gate: -S, where S varies by regime type)

Regulator's payoff function:

U(gate) = αE + βB - γL(gate|environment) - δT - εS U(paper) = αE(paper) - γL(paper|environment) + δT(paper)

Where α, β, γ, δ, ε are the regulator's preference weights.

Key parameter: S (surveillance-value loss)

For a democratic regulator in an EU-style system, S ≈ 0 or small — bulk surveillance is constrained by law. The gate's privacy properties are a feature, not a cost. The regulator's choice depends primarily on E, B, and L.

For an authoritarian regulator, S is large. The gate makes their regulated entities invisible to bulk collection. Even with high E and B, the payoff may be negative. The regulator's dominant strategy is to maintain paper enforcement or ban gates.

Nash equilibria by regulator type:

1. Democratic/pro-rule-of-law regulator: U(gate) > U(paper) when α(E_gate - E_paper) + βB > γ(L_gate - L_paper) + δ(T_gate - T_paper). Since E_gate >> E_paper, B > 0, and S ≈ 0, the gate dominates for any reasonable parameter range. Unique equilibrium: encode gate.
2. Authoritarian/high-surveillance regulator: U(gate) may be negative if εS > α(E_gate - E_paper) + βB. Unique equilibrium: maintain paper or ban gates.
3. Captured regulator: If incumbents (Game 3) successfully lobby, γ(L_gate|pro-incumbent) can be made large. The equilibrium depends on whether the capture is stronger than the enforcement benefit.

How Game 1 and Game 2 interact:

If Game 1 reaches "both adopt gate" without the regulator, the regulator's payoff from encoding increases further — there is now proven infrastructure, lowering L and T. This makes encoding more attractive even for a borderline regulator.

If Game 1 is stuck in coordination failure (both maintain), the regulator can break it by encoding — this is the most leveraged intervention. A forward-leaning regulator who encodes early creates the condition for Game 1 to tip.

What changes the equilibrium:

• A high-profile AI harm event increases B and decreases L for the regulator choosing gate (the public demands action).
• An incumbent lobbying campaign increases γ(L_gate) if the regulator is vulnerable to capture.
• A regime change can flip S from 0 to large or vice versa.

Game 3: Hyperscaler Defense

An incumbent hyperscaler (AWS, Azure, GCP) chooses how to respond to the gate architecture. Enterprise adoption level a ∈ [0,1] is a parameter.

Strategy sets:

• Hyperscaler: {Fight, Accommodate, Co-opt}

  • Fight: lobby against gate standards, bundle a fake "verified" wrapper, use procurement lock-in to deter adoption
  • Accommodate: support gate instances on their infrastructure, capture revenue from the compute that gate instances need
  • Co-opt: acquire or clone the gate provider, offer gate-as-a-service on their terms

Payoff parameters:

Let:

• R_h = hyperscaler's annual revenue from the enterprise segment that gates target ($B+ per hyperscaler in regulated markets)
• S_h = share of R_h that depends on unrestricted data access (surveillance advertising, training data extraction, platform lock-in)
• P_h = cost of fighting (lobbying spend, engineering cost of fake wrapper, reputation damage if exposed)
• A_h = revenue from accommodating (compute for gate instances, storage for proof logs — lower margin but insulated from competition)
• C_h = cost of co-opting (acquisition price or clone engineering cost)

Hyperscaler's payoff by strategy and adoption level a:

Strategy	a < 0.1 (Phase 0-1)	a > 0.1 (Phase 2+)
Fight	-P_h + (1-k)R_h (preserves most revenue but spends on lobbying)	-P_h + (1-k')R_h (lobbying less effective; revenue erosion accelerates)
Accommodate	A_h × a (small but growing revenue from gate compute)	A_h × a + residual non-gate R_h (gate revenue grows, non-gate revenue declines)
Co-opt	-C_h + R_h (full control but high cost)	-C_h + R_h (but harder to execute — gate ecosystem has network effects)

Where k, k' are revenue erosion rates from gate adoption (k < k').

Nash equilibria:

1. At low adoption (a < 0.1): Fight dominates if -P_h + (1-k)R_h > A_h × a and -P_h + (1-k)R_h > -C_h + R_h. The hyperscaler tolerates the lobbying cost because it preserves most of their business model. This is the equilibrium in Phases 0-1.
2. At high adoption (a > 0.1): If the gate ecosystem has enough momentum that A_h × a + residual > fight payoff, and co-opt is infeasible (network effects of the open regression suite, AGPL license, community), then Accommodate becomes dominant. The hyperscaler accepts the two-tier outcome and captures what they can from gate compute.
3. **Co-opt is only dominant if C_h < P_h + kR_h — the hyperscaler must believe they can acquire the gate provider cheaper than fighting or losing revenue. The AGPL license and the open nature of the regression suite make co-opt difficult (you can buy the code but you can't buy the community).

Key insight: The hyperscaler's optimal strategy changes with adoption level. At low adoption they fight because the cost is low and the revenue preservation is high. At a threshold adoption level, the fight becomes more expensive than accommodation. This threshold is determined by S_h — the share of revenue dependent on unrestricted data access. A hyperscaler with high S_h (Google, whose advertising business depends on data) will fight harder and longer than a hyperscaler with lower S_h (AWS, whose revenue is more infrastructure-based).

What changes the equilibrium:

• A high S_h means the fight threshold is higher — the hyperscaler tolerates more revenue erosion before accommodating.
• AGPL license and community ownership raise C_h (cost of co-opting) and lower the probability of co-opt succeeding.
• If adoption jumps suddenly (regulator encode in Game 2), the hyperscaler may skip the fight phase entirely and move directly to accommodation.

Game 4: Social Protocol Platform Competition

The social protocol does not compete with any single platform. It offers an alternative to the entire paradigm of centralized internet services — a single protocol that replaces 20+ products across social graph, publishing, video, messaging, e-commerce, payments, contracts, identity, code hosting, collaboration, and freelancing. See the Social Protocol Competitive Landscape for the full platform map.

Why this is a different kind of game:

Unlike the institutional games (binary adopt/don't-adopt decisions by firms), the social game is a multi-front platform competition where:

1. The protocol competes against 20+ incumbents simultaneously, each with its own network effects and switching costs.
2. No single incumbent can copy the bundle — Meta cannot offer portable identity (it destroys their surveillance model), Google cannot offer private messaging (it destroys their ad model), Stripe cannot offer contracts and social, DocuSign cannot offer payments.
3. The protocol's value proposition is not "Twitter but better" — it is "one account replaces every platform you use."
4. Users do not switch all at once. They join for one use case and discover the rest. This is a sequential expansion game.

The entry vector game entry choice:

The protocol must choose which use case to lead with. Each candidate has different payoff parameters:

Entry vector	Cold start	ARPU	Bundle necessity	Competitive response	Failure mode
Organized communities (HOAs, clubs, PTAs)	Solved (groups exist as units)	Low ($20-100K fees)	Full	Negligible (no incumbent targets this segment)	No community finds PMF
Community refugees (banned subreddits, nuked discords)	Solved (arrive together)	Low-medium	Partial	Medium (Reddit/Discord defensive)	Migrations don't stick
Creators (OnlyFans, Patreon, adult content)	High (individual migration)	High ($1-5M fees)	Partial (identity + content + payments)	High (OnlyFans, Stripe, payment processors)	Creator acquisition cost too high
Freelancers (Upwork, Fiverr)	Low (scattered, no density)	Medium	Full (contracts + payments + reputation)	Low (Upwork network effects weak)	No dispute resolution trust
Developers (GitHub)	High (open-source communities exist)	Low	Partial (code + identity)	Very high (Microsoft/GitHub moat)	Developer habit inertia

The entry vector game payoff matrix (protocol chooses primary vector, fate chooses incumbent response strength):

Let:

• U_x = users acquired through vector x at Phase 0
• C_x = engineering cost to ship the features vector x requires
• R_x = revenue from vector x users
• I_x = incumbent response strength (0 = none, 1 = existential fight)
• S_x = stickiness (probability user stays for other capabilities after joining for vector x)

Expected protocol payoff for choosing vector x: P(x) = U_x × (R_x + S_x × future_value_of_expansion) - C_x - D_x

Where D_x is damage from incumbent response (lobbying, FUD, feature parity, price cuts) weighted by I_x.

Nash equilibria by entry vector:

1. Organized communities (Phase 0): I_x ≈ 0 (no existing platform serves this segment well). U_x is bounded but guaranteed (groups exist). S_x is high (full bundle forces exposure to all capabilities). Unique equilibrium: this is the dominant entry strategy.
2. Creators (Phase 0 parallel): I_x is high (OnlyFans, Stripe, payment processors have incentive to block). But R_x is also high (highest ARPU). U_x is unpredictable (depends on creator acquisition cost). Multiple equilibria: high-I/low-U leads to "do not enter," low-I/high-U leads to "enter." The outcome depends on whether payment discrimination against adult content is acute enough to overcome switching friction.
3. Freelancers (Phase 2): I_x is low (Upwork's network effects are weak — both sides multi-home). But S_x requires the reputation graph from Phase 0-1 to exist first. This is a sequential game — the payoff from enter in Phase 2 depends on having entered Phase 0-1 successfully.

The asymmetric bundle advantage:

The protocol's structural advantage is that it competes with 20+ incumbents, but no incumbent competes with the protocol. Each incumbent faces a different threat:

Incumbent	Threat level	Can adapt?	Adaptation would require
Meta (Facebook, Instagram)	Existential	No	Abandon surveillance advertising and portable user data — their entire business model
Google (YouTube, Gmail, Google Docs)	High	No	Abandon data mining for ads — their entire business model
Microsoft (GitHub, LinkedIn, Office)	Moderate	Partial	Accept decentralized identity — but Windows/Office lock-in is different from data mining
Twitter/X	Medium	No	Algorithmic feed is their product — giving up curation control destroys their ad model
Stripe	Low	No	Cannot offer social, contracts, and identity — outside their competence
Discord	Medium	Possible	Can add more features but cannot offer portable identity or zero-fee payments
Substack/Medium	Medium	Possible	Can improve creator tools but cannot match zero-fee + censorship resistance
OnlyFans/Patreon	Low	No	Payment discrimination is structural (banking regulations) — not a choice they can undo
DocuSign	Low	No	Contracts + payments + social is outside their competence
Reddit	Low	Possible	Decentralized moderation would destroy their ad business

Key insight: of the 20+ incumbents, most (Meta, Google, Twitter, Stripe, DocuSign) cannot adapt because adaptation requires abandoning their business model. A few (Discord, Substack, Reddit) can make marginal improvements but cannot match the bundle because they don't control the other layers. Zero incumbents can match the full protocol.

Geopolitical dimension: free speech and association infrastructure:

The social protocol is not just a consumer product. Its architecture makes it naturally censorship-resistant and surveillance-resistant:

• Speech on the protocol cannot be removed by any government because there is no central server to issue takedown orders to. Each user controls their own PDS; content is addressed by CID, not hosted on a platform URL.
• Association on the protocol cannot be surveilled because the relay network routes by DID, not IP. The state can see that a message was sent but cannot determine who sent it to whom without controlling the entire relay graph.
• Organisation on the protocol cannot be disrupted because Collective Personas exist cryptographically — there is no office to raid and no server to seize.

This makes the protocol a direct geopolitical tool. For citizens in restrictive regimes (China, Russia, Iran, Belarus, Myanmar, Eritrea, and dozens of others), the protocol offers the first infrastructure that enables free speech and association that their government cannot control, surveil, or shut down.

This changes the entry vector analysis:

Entry vector	Previously framed as	Geopolitical frame adds
Community refugees	Banned subreddits, nuked discords	Political dissidents, exiled journalists, banned NGOs — users fleeing state censorship, not just platform moderation
Creators	OnlyFans/Patreon refugees	Journalists and writers under threat in authoritarian regimes — their incentive is not just revenue but survival
Organized communities	HOAs, clubs, PTAs	Exile communities, diaspora groups, cross-border activist networks — users with an organizational structure that cannot exist on any centralized platform

The geopolitical entry vectors have different payoff parameters than the consumer ones. Their cold start is solved not by group density but by necessity (the alternative is censorship or imprisonment). Their stickiness S_x is near-maximal because the protocol is not a convenience but a lifeline. Their ARPU may be low but their strategic value is high — a single dissident community successfully using the protocol demonstrates censorship resistance to millions.

How this changes Game 5 (geopolitical competition):

The legitimacy cost L_i for a state that bans the protocol was modeled in Game 5 as a general cost. But the social protocol's free speech dimension makes L_i much larger than a standard trade restriction:

• A state that bans the protocol is not just blocking a technology — it is telling its citizens "we are afraid of you speaking freely."
• The protocol makes the ban visibly ineffective — citizens who can access the relay network retain their speech. The state either accepts visible defiance or invests in internet shutdowns, both of which carry high legitimacy costs.
• The protocol's verification layer means that a banned instance can still prove it is operating correctly. A state that claims to have shut down gates cannot fake compliance.

This creates a new dynamic in Game 5: states that ban the protocol face a legitimacy cost that grows with the protocol's adoption in free-world jurisdictions. The more citizens in free countries use the protocol, the more citizens in restricted countries know what they are missing, and the higher the legitimacy cost of the ban.

The interaction between Game 4 (social) and Game 5 (geopolitical):

Direction	Effect
Game 4 → Game 5	Social protocol adoption in free-world jurisdictions increases L_i for restrictive states (citizens in restrictive regimes see what they're missing; free-world users amplify dissident content).
Game 4 ← Game 5	A restrictive stance in Game 5 (China bans gates, Russia blocks relays) makes the social protocol's censorship resistance the only option for citizens in those countries, accelerating adoption as a human rights tool.
Game 4 → Game 2	Social protocol's demonstrated censorship resistance gives democratic regulators evidence that verification is compatible with free speech, lowering their L(gate	political) in Game 2.

The net effect: the social protocol's geopolitical dimension does not just make it more resilient — it creates a positive feedback loop with the geopolitical game that no purely consumer platform can generate.

Tech industry alignment: natural allies:

The protocol creates winners and losers in the tech industry. Companies whose business model depends on surveillance advertising (Meta, Google, Twitter/X, TikTok) are structural losers — the protocol's verification layer blocks their data extraction. But companies whose business model is compatible with or enhanced by verification are structural winners. These companies have an incentive to embrace the protocol as a competitive weapon against their surveillance-dependent rivals.

Company	Business model	Gate compatibility	Incentive to embrace	What they gain
Apple	Hardware (iPhone, Mac) + services (iCloud, App Store)	High — privacy is their brand, Secure Enclave is already a proto-gate	Strong — verification structurally disadvantages Meta and Google, Apple's main competitors in services and AR	New hardware differentiation ("the verified iPhone"), new services (iCloud PDS, gate appliance), brand reinforcement
Microsoft	Enterprise software (Office, Azure) + OS (Windows)	Medium — Azure can host gates, but Windows telemetry conflicts	Moderate — enterprise clients want verification, but Windows surveillance revenue is small	Azure differentiation for regulated industries, Office integration with gate proofs
Cloudflare	Edge infrastructure (CDN, DNS, security)	High — already privacy-forward, Workers platform could host gate logic	Strong — gate-compatible edge services are a new product category	New revenue from gate relay and verification edge services
Samsung	Hardware (phones, TVs, appliances)	Medium — Android allows gate installation, Knox security platform	Moderate — can differentiate Galaxy as "gate-compatible Android"	Hardware differentiation, B2B enterprise sales
IBM	Enterprise services + consulting	High — compliance is their core market	Strong — gate rule consulting is a natural service line	New consulting practice, mainframe integration
Amazon/AWS	Cloud infrastructure + advertising + retail	Mixed — AWS can host gates, but Amazon advertising depends on data	Conflicted — AWS division may embrace, advertising division fights	AWS can capture gate compute revenue but loses advertising data access

The Apple scenario as a case study:

Apple's incentive to embrace the protocol is the strongest of any major tech company because:

1. Apple's revenue does not depend on surveillance — 80%+ comes from hardware and services that work identically with or without data extraction. The protocol's privacy properties cost Apple nothing.
2. Apple's brand is built on privacy ("what happens on your iPhone stays on your iPhone"). Integrating a DID-based identity system and gate-compatible Secure Enclave makes this claim verifiable rather than aspirational.
3. Apple's main growth competitors (Meta in AR/VR, Google in services and AI) are structurally threatened by verification. Meta's entire business model collapses if advertising cannot extract user data. Google's AI advantage depends on unrestricted data access. The protocol harms them more than it harms Apple.
4. Apple already has the hardware foundation: the Secure Enclave, M-series chips with隔离 zones, and iCloud Keychain are all proto-gate infrastructure. Shipping a gate-compatible iPhone would be a feature update, not a new product.

Apple's optimal strategy: embed the protocol's DID system into Apple ID, ship gate-compatible hardware in the next iPhone generation, position iCloud as a PDS hosting service, and market the entire stack as "privacy you can prove." This would instantly put gate capability in 1B+ existing devices and create a distribution channel no other protocol project has ever achieved.

The strategic consequence for the five games:

If a major compatible company (Apple, Cloudflare, or IBM) embraces the protocol:

• Game 1 (enterprise adoption): Apple's endorsement massively lowers G (gate cost) because enterprise buyers already have Apple hardware. The coordination failure breaks faster.
• Game 3 (hyperscaler defense): A compatible hyperscaler (Cloudflare, Azure) providing gate-compatible infrastructure creates an accommodation path for enterprise users, reducing the fight incentive.
• Game 4 (social protocol): Apple's distribution gives the social protocol a built-in user base of 1B+ devices. The cold start problem in every entry vector is solved by the installed base.
• Game 5 (geopolitical): A US-based company (Apple) embracing the protocol makes it harder for the US to restrict it. Apple's lobbying power counterbalances Meta and Google's anti-gate lobbying.

The protocol does not need any major company to embrace it — the trajectory is viable without them. But if even one does, the timeline shortens from decades to years.

Sequential expansion game:

The protocol expands through categories sequentially, not simultaneously. Each phase builds the conditions for the next:

Phase	Vector	Capabilities shipped	Enables next phase by
0	Organized communities	Full bundle (identity + content + payments + contracts + governance)	Proving the bundle works for real coordination
1	Refugees + creators	LSAT (paywalled content), Lightning subscriptions	Building reputation graph and creator audience
2	Freelancers	SCAL stack, arbitration guilds	Reaching contract marketplace critical mass
3	Institution crossover	Enterprise-grade verification	Leveraging network for institutional products

This is not a choice between vectors — it is a sequence where each phase unlocks the next. The protocol can fail at any stage if the current vector never reaches critical mass.

How the social game interacts with the institutional games:

The social and institutional games are independent in their causal logic but connected through resource flows:

• Institutional → Social: Compliance revenue would accelerate the social protocol's development timeline, but the architecture is self-developing after Stage 2 — the bootstrap loop means engineering does not depend on external funding. Revenue is amplificatory, not necessary.
• Social → Institutional (late stage): At Phase 3+, the social network's users become the distribution channel for institutional products. Institutional verification tools sell as fulfillment orders to a network that already exists.
• Social → Institutional (ongoing): Social contract disputes generate real-world edge cases for the collective regression suite, making the certification more valuable for institutional buyers.

Asymmetric coupling: The two sides are largely independent in both directions. Institutional success can exist without social success (compliance sales work with zero social users). Social success can exist without institutional success — the social protocol reaches critical mass through pure platform competition, and its development is not blocked by lack of institutional revenue because the architecture is self-developing after Stage 2 (the agent writes its own code; the bootstrap loop makes engineering self-sustaining). The coupling is amplificatory, not foundational: each side's success makes the other more valuable, but neither depends on the other to survive.

Game 5: Geopolitical Competition

Nation states do not act in isolation. Each regulator's choice (Game 2) is shaped by what other states are doing, creating a strategic game between major geopolitical blocs. This game determines whether the protocol remains unified or fragments into incompatible networks.

The players:

Player	Surveillance dependence	Regulatory sophistication	Gate incentive	Key constraint
EU	Low (GDPR limits bulk collection)	High (AI Act, NIS2, eIDAS)	Strong — gates solve enforcement at scale	Must maintain coherence across 27 member states
US	High (NSA, FBI surveillance apparatus)	Fragmented (sectoral agencies, no unified AI law)	Conflicted — financial regulators want verification, intelligence community wants visibility	Tech incumbents (Meta, Google) lobby against; surveillance state fights gates structurally
China	Total (social credit system, mass surveillance)	Top-down (state dictates standards)	Negative — gates block surveillance entirely	Would develop state-controlled "verified" alternative with backdoors
UK	Medium (Investigatory Powers Act)	Medium (emerging AI safety framework)	Positive — post-Brexit wants regulatory leadership	Must balance US alliance with EU regulatory alignment
India	Medium (growing surveillance infrastructure)	Developing (DPDP Act 2023)	Positive — leapfrog to verification without legacy audit industry	Wants sovereignty; may resist a standard defined by EU or US

Strategy sets:

Each state chooses one of:

• Promote: Encode gate rules, subsidize adoption, advocate as international standard
• Tolerate: Allow gate instances but don't mandate; let the market decide
• Restrict: Ban or backdoor gate instances; require state-approved verification
• Compete: Develop an alternative state-controlled verification standard

Payoff parameters for state i:

Let:

• S_i = surveillance-value loss from gate adoption (0 for EU, high for China)
• E_i = enforcement benefit from automated regulatory compliance
• T_i = trade dependency on states with incompatible stances
• A_i = alliance alignment pressure (cost of diverging from key allies)
• L_i = legitimacy gain/loss from gate stance — includes not just public opinion and human rights reputation, but the specific cost of being seen to block free speech and association infrastructure. This parameter grows with the social protocol's adoption in free-world jurisdictions, because each new user in a free country makes the ban's censorship intent more visible to citizens in restricted countries.
• K_i = first-mover advantage if state defines the standard (tech industry attracted, standard-setting fees, geopolitical influence)

Payoff function:

U_i(stance) = αE_i - βS_i(stance) - γT_i(stance) + δA_i(stance) + εL_i + φK_i(stance)

Key equilibria by scenario:

1. EU promotes, US tolerates, China restricts — the fragmentation equilibrium. EU encodes gates as the AI Act enforcement mechanism. US allows gate instances but does not mandate them (compromise between financial regulators wanting verification and intelligence community opposing it). China bans all gate instances that it cannot surveil and develops a backdoored "verified" alternative. Global enterprises operate dual configurations: EU-compliant gate instances for European operations, non-verified for China, mixed for US. The protocol fragments into two tiers: "free world" protocol and "China-controlled" alternative, but the free-world protocol remains functionally unified because the EU and US standards are technically compatible (even if US doesn't mandate enforcement).
2. The EU also restricts — the worst case. If the EU, under pressure from member states with surveillance priorities, mandates data localization or backdoors for gate instances, the protocol's core value proposition (verification cannot be bypassed by the state) is broken. The EU and US would converge on a weakened standard that preserves some surveillance capability, and the protocol loses its structural advantage. This scenario requires a significant shift in EU governance toward surveillance — possible but unlikely given GDPR and current AI Act trajectory.
3. EU promotes, US promotes, China isolates — the best case. The US resolves its internal conflict in favor of financial regulators and AI safety advocates (triggered by a high-profile AI harm event that makes gates politically necessary). US and EU jointly promote gates as the international standard. China isolates itself with an incompatible alternative. Global enterprises standardize on the EU-US protocol. This is the scenario that maximizes the protocol's growth trajectory.
4. Cold War standoff — US restricts, EU tolerates. The US intelligence community wins the internal debate and bans unrestricted gate instances (requiring backdoors for law enforcement). The EU tolerates but does not mandate. The protocol survives in the EU but loses the largest enterprise market (US-based global firms). Growth slows significantly but does not stop — the EU market alone (450M people, major regulated industries) is enough to sustain the trajectory.

The race dynamic:

The first major jurisdiction to encode gates as a regulatory standard gains first-mover advantage (K_i). The standard they define — rule format, proof requirements, certification process — becomes the default. Late adopters must either adopt the existing standard (ceding some sovereignty) or fragment the market. This creates a race:

• EU is best positioned. AI Act enforcement begins 2026-2027. The EU can encode gate rules as the compliance mechanism for high-risk AI systems. If they do this before the US resolves its internal conflict, the EU defines the standard and the US becomes a standard-taker.
• US could catch up. A single AI harm event that causes measurable financial damage ($1B+) could shift US political will. If the US then moves faster than the EU through executive action, they could define the standard instead.
• China cannot win the race but can fragment it. China cannot adopt the protocol without losing surveillance. But they can fund an alternative standard, pay developing countries to adopt it, and create a bifurcated global market.

How Game 5 interacts with Games 1-4:

• Game 5 determines the parameter S (surveillance-value loss) for each regulator in Game 2. A state that chooses "restrict" in Game 5 has high S in Game 2, shifting the regulator's equilibrium.
• Game 5 determines whether enterprises face compatible or incompatible requirements across jurisdictions, which affects their Game 1 payoff (N = network effect benefit is lower if the protocol fragments).
• Game 5 outcome affects the hyperscaler's S_h (if gates are legal in the US, AWS can accommodate; if banned, they must fight).
• Game 4 (social) is the most resilient to geopolitical fragmentation — individual users can still participate in the protocol regardless of their state's stance, as long as the relay network operates across borders.

The stability claim for Game 5:

The protocol has a structural defense against geopolitical fragmentation that no previous decentralized technology had: the gate itself. Verification is self-authenticating — a gate instance in a non-EU jurisdiction can still prove compliance with EU gate standards. The protocol does not need states to agree; it only needs each state to allow at least one path for verification to exist. A state that bans gates loses the economic benefits of verification (lower compliance costs, AI safety, efficient contracts) without gaining enforcement capability (because banned gates will exist in other jurisdictions anyway). This creates a ratchet: once verification is adopted by enough economically significant jurisdictions, non-adopting states face a growing competitive disadvantage that eventually exceeds their surveillance benefit.

The combined game: why the trajectory is stable

The five games interact through their parameters:

1. Game 1 (enterprise adoption) can have two equilibria, but Game 2 (regulator commitment) and the insurance loop (Game 3 consequence) both break the coordination failure by changing the dominant strategy for the first mover.
2. Game 2 (regulator commitment) has a unique equilibrium for democratic regulators: encode gates. The only failure mode is authoritarian suppression or capture, both of which are jurisdiction-specific — they fragment the protocol but don't stop it globally.
3. Game 3 (hyperscaler defense) has a threshold that depends on adoption level. Since Games 1 and 2 push adoption past that threshold, the hyperscaler's optimal strategy shifts from fight to accommodate. They do not defect back to fight once adoption is high because the sunk cost of fighting has already been incurred and the revenue base has already shifted.
4. Game 4 (social protocol) is structurally independent of the institutional games — it does not depend on any enterprise outcome. Its trajectory is driven by the asymmetric bundle advantage: 20+ incumbents, zero of whom can match the full protocol. The binding constraint is not competition but status quo inertia — the protocol must demonstrate one use case that is dramatically better, then expand from there.
5. Game 5 (geopolitical competition) determines whether the protocol fragments or remains unified. The default equilibrium is fragmentation between surveillance and rule-of-law blocs, but the free-world protocol remains functionally unified because verification is self-authenticating — states don't need to agree. The ratchet dynamic means non-adopting states face growing competitive disadvantage over time.

The stability claim:

The only way the trajectory fails is if one of the following parameter conditions holds:

• Game 1 parameter failure: G + N > C + R even with insurance differentiation. This would require gate costs to be higher than compliance savings — unlikely given the 4-20x cost ratio, but possible if the gate implementation is expensive.
• Game 2 parameter failure: εS > α(E_gate - E_paper) + βB for all major regulators. This requires a coordinated shift toward high- surveillance governance across all regulated markets — possible but requires a global authoritarian turn.
• Game 3 parameter failure: S_h ≈ 1 and C_h < P_h + kR_h — the hyperscaler both depends entirely on unrestricted data and can successfully co-opt the gate provider before the ecosystem reaches critical mass. The AGPL license and open community are the defense against this.
• Game 4 parameter failure: every entry vector fails to demonstrate a use case dramatically better than the status quo, OR the social protocol never achieves critical mass in any vector. This is possible if the bundle is too complex for even organized communities (the lowest-friction vector).
• Game 5 parameter failure: the EU shifts toward surveillance governance, adopting data localization or backdoor mandates for gate instances. This breaks the protocol's core value proposition in its most natural market. This scenario requires a significant shift in EU political trajectory — possible but unlikely given GDPR and current AI Act direction.

If none of these parameter conditions hold, the unique equilibrium across all five games is: gates win on the institutional side, and the protocol has a viable trajectory on the social side via at least one entry vector. The two sides reinforce each other but do not depend on each other for survival — each can succeed independently, and together they compound.

Failure modes as parameter shifts

The failure modes from the Adoption document can now be re-expressed as parameter changes:

• Agent quality insufficient: G increases (gate requires too much human correction, raising effective cost). If G + N > C + R, Game 1 reaches "both maintain audit" and the regulator's encoding in Game 2 becomes less attractive (E_gate drops because gates are unreliable).
• State suppression: εS dominates for all major regulators. Game 2 equilibrium shifts to "maintain paper/ban gates" in significant jurisdictions. Protocol fragments.
• Bootstrap loop stalls: Same as agent quality — G stays high for too long. Game 1 fails to tip.
• Neither adoption reaches critical mass: Games 1 and 2 both fail — enterprise adoption stalls at Phase 1 (Game 1 coordination failure unresolved) and social protocol never achieves network effects (different game, not modeled here).