The SEC Division of Corporation Finance added three FAQ entries on November 2, addressing when governance token holders constitute a "central party" under federal securities law. The update landed two days before $879.9 million in token unlocks — including $131.88 million from AI-focused blockcha...
"Answers represent staff views, have no legal force, and have not been approved or disapproved by the Commission." — SEC Division of Corporation Finance, Crypto-Assets FAQ Update, November 2, 2026
The SEC Division of Corporation Finance added three FAQ entries on November 2, addressing when governance token holders constitute a "central party" under federal securities law. The update landed two days before $879.9 million in token unlocks — including $131.88 million from AI-focused blockchain Kite — hit the market. Simultaneously, on-chain detective ZachXBT publicly labeled 99% of AI agent tokens as scams, and the x402 stablecoin payment protocol surpassed 138 million cumulative agent transactions on Base alone.
These events are not coincidental. They mark a structural inflection in how AI x crypto projects must justify their tokens' existence. The SEC's guidance separates "functional" from "decentralized," the x402 standard is building genuine machine-to-machine payment rails that require no token at all, and the token unlock calendar forces a reckoning between dilution and utility. For token holders in AI-adjacent protocols, the question is no longer whether these tokens will accrue value — it is whether they need to exist.
On November 2, the SEC's Division of Corporation Finance expanded its crypto-assets FAQ with questions 1.4, 1.5, and 1.6, per The Currency Analytics. The additions address three structural questions that have haunted DeFi governance since inception.
Question 1.4 asks whether participants in a broad-based governance system constitute a "central party." The staff's answer: unaffiliated governance participants who have no coordinated action beyond voting typically do not form a centralized group — even if they collectively hold voting power. However, operational, economic, or voting control by any subset does trigger central-party classification. According to KuCoin News, the FAQ confirms that decentralized governance can coexist with legal entities like DUNAs (decentralized autonomous nonprofit associations).
Question 1.5 defines "oversight of participation" as the authority to control access to the system at any level. This is pointed language. Protocols where a foundation or core team controls validator onboarding, smart contract upgrades, or user whitelisting now face explicit regulatory attention.
Question 1.6 distinguishes "functional" from "decentralized" — two concepts the industry has conflated for years. A token can be functional (usable for its programmed purpose) without the system being decentralized (operating without any party holding control). Per CryptoTimes, the FAQ carries no legal force, but sets the interpretive framework staff will apply to enforcement actions.
The implications for AI tokens are immediate. Projects like Virtuals Protocol, where veVIRTUAL governance is live but the team retains protocol upgrade authority, now sit in the gap between "functional" and "decentralized." NEAR Protocol's governance vote on a July 2026 mainnet upgrade technically failed to reach quorum, per Paragraph Research, yet the upgrade was implemented when 80% of validators adopted the new version — an outcome that fits the SEC's definition of oversight by a coordinating group.
Development activity across AI x crypto repositories reflects a sector building infrastructure rather than governance tooling.
ElizaOS/eliza (19,767 stars, 5,871 forks) remains the most-forked AI agent framework. Recent commits from November 1–2 focus on voice relay performance, speech session extension, and Cloudflare Durable Object coordination — all infrastructure work. Notably, every recent commit is co-authored with "Claude Opus 5.5," indicating heavy AI-assisted development. The repository has 295 open issues. No recent commits address governance, tokenomics, or value accrual, consistent with ElizaOS's post-lawsuit pivot to pure open-source software after the $2.6 billion AI16Z token collapse, per OwnYourMind.
Virtuals Protocol archived its openclaw-acp agent coordination repository on November 1 (167 stars, 56 forks), replacing it with @virtuals-protocol/acp-cli. The deprecation signals a tooling consolidation, but the archived repo's 18 open issues remain unresolved. Atlas SDK, a third-party coordination layer connecting Virtuals Protocol, EVM chains, and Robinhood Chain, was last updated October 13.
Breadchain Cooperative's commonware-restaking — a Rust toolkit for building EigenLayer AVS applications — was last updated September 11. The nexus-agent-protocol repo (ERC-4337 wallets, ZK-proof payments, EigenLayer AVS integration for autonomous AI agents) has not been updated since July 2026. The gap between announcement-level activity and sustained commit history is widening across the AI agent subsector.
The awesome-crypto-ai-agents curated list (7 stars, 12 forks, last updated October 21) catalogues dozens of projects but reinforces ZachXBT's observation: the ratio of listed projects to actively maintained repositories is unfavorable.
On-chain investigator ZachXBT declared in a public post on X that "99% of [AI agent tokens] are a scam," adding that "AI agent wrapper grifts are probably worse than other past trends," according to Binance Square. The criticism distinguished AI tokens from memecoins: memecoins "at least promised nothing," while AI tokens "try to larp as much as possible to appear legit to unsuspecting buyers."
The data supports the skepticism selectively. Autonolas (OLAS), one of the earliest AI agent governance tokens, has declined 99.6% from its all-time high, per CollectiveShift. Its marketplace has turned over just $89,000 in cumulative volume. In Q3 2026, the DAO voted to have staking contracts on seven chains report unspent OLAS balances back to Ethereum — a governance action required to prevent pointless dilution.
ElizaOS shut down its foundation in August 2026 after settling a class-action lawsuit, according to OwnYourMind. The ELIZAOS token has no staking, no fee accrual, no burns, and no governance voting. Founder Shaw Walters stated publicly that development would continue without another token launch.
Venice AI (VVV) presents the counter-case. VVV is explicitly not a governance token — holders have no voting rights and there is no DAO. Governance is centralized by design. Instead, staking VVV grants proportional access to daily inference capacity: a stake-for-compute model that has driven 3,000%+ appreciation in 2026, per Datawallet. The team argues that governance tokens create misaligned incentives where holders vote for financial benefit over platform health.
Meanwhile, a new entrant — AGENT, from a Zürich-based project — announced a December 1 community presale for a 1-billion-supply ERC-20 token designed to settle payments across a multi-agent AI system, per GlobeNewsWire. The token has no burn mechanism and no governance function — it is positioned purely as a payment instrument, which raises the question of why it cannot simply use USDC over x402.
The x402 protocol — revived from the dormant HTTP 402 status code by Coinbase in May 2025 — has become the default payment rail for machine-to-machine transactions. Per Stablecoin Insider, the protocol processed 138 million cumulative payments on Base through Q3 2026, up from 1,500 at mid-2025.
The x402 Foundation, backed by Visa, Mastercard, Google, and Cloudflare, launched in July 2026, recording 75 million transactions in its first 30 days, per CoinDesk. Stripe shipped x402 support as "Machine Payments" on February 10, 2026.
The protocol charges zero fees beyond gas costs and settles in 2 seconds. Weekly agent-to-agent transfers peaked at 8.7 million in the week of August 17, moving approximately $367,950 — roughly $0.04 per payment, per Bitget.
This is the structural threat to AI agent tokens. If machine-to-machine payments settle in stablecoins over a free, open protocol backed by Visa and Google, the utility argument for project-specific payment tokens collapses. AGENT's December presale, Venice's DIEM companion token ($1/day API credit), and Kite's native KITE token all compete with a payment rail that requires no new token.
Three token unlocks totaling $879.9 million hit the market in the first week of November 2026, per BeInCrypto:
| Token | Unlock Size | Value | % of Supply | Date | |-------|------------|-------|-------------|------| | Bitway (BTW) | 101.63M BTW | $109.66M | 3.37% | Nov 2 | | Kite (KITE) | 640M KITE | $131.88M | 17.41% | Nov 3 | | Hyperliquid (HYPE) | 180K HYPE | $15.72M | 0.08% | Nov 6 |
Kite's unlock is the most structurally significant for the AI governance thesis. The L1 blockchain, designed for agentic payments with verifiable identity, unlocked 640 million tokens to team, advisors, and early investors — a 17.41% supply expansion one year after listing. Per The Industry Spread, different unlock trackers disagree on the cliff structure, adding uncertainty to sell pressure projections. Tokenomist lists both the Bitway and Kite unlocks as awaiting verification.
The broader November calendar is heavier. Per CoinAlert News, Hyperliquid leads a total $1.72 billion November unlock schedule, with HYPE unlocks concentrated later in the month.
Pendle executed one of 2026's most consequential tokenomics overhauls on January 20, replacing the vePENDLE vote-escrow model (two-year locks) with sPENDLE, a liquid staking token with a 14-day unstaking period, per Coin Bureau. Under the legacy model, vePENDLE holders received 80% of swap fees, per Tokenomics.com. The transition to sPENDLE maintained fee distribution but eliminated the long-term alignment mechanism, signaling that even successful vote-escrow models face liquidity pressure from holders.
Morpho reached $11.478 billion in TVL with $5.61 billion in active loans by October 2026, per Eco.com. Its governance-minimized design — where deployed Blue markets are immutable and governance cannot change their parameters — represents the anti-governance position. MORPHO token utility is limited to DAO voting on new market parameters, not existing ones. Institutional adoption by Coinbase, Apollo, and Bitwise validates the approach.
EigenLayer activated production slashing on April 17 with zero slashing events recorded through May, per BlockEden. ELIP-012, approved in March 2026, directs incentives to fee-generating AVSs and routes fees into buybacks. But the EIGEN token is down ~97% from peak, and protocol TVL has halved. The $292 million Kelp DAO bridge exploit in April — attributed to the Lazarus Group — exposed the restaking stack's counterparty risk.
Compound's Proposal 612 illustrated governance dysfunction in October. Community delegate ugurmersin alleged the Compound Foundation converted 8.42M DAI from reserves into 344,780 COMP to swing votes, per The Defiant. The proposal to extend treasury withdrawal delays from 2 to 10 days passed with 1.89M votes for vs. 1.04M against — then was cancelled on October 7 for unstated reasons.
Value flows in AI x crypto divide into three tiers:
Tier 1 — Real yield to token holders: Pendle (sPENDLE holders receive 80% of swap fees, 8–15% APR in 2026). Uniswap's fee switch generated $48.94M in cumulative protocol revenue through September 2026 at a $325K daily run-rate across seven networks. The top 12 DeFi protocols collectively distributed ~$800M/month in buybacks and fee distribution by May 2026, per prior webthreepedia analysis.
Tier 2 — Utility without governance: Venice (VVV) stake-for-compute grants proportional inference capacity. Kite (KITE) is a payment settlement token for agentic transactions. Neither distributes fees to holders; value accrual is structural (network usage demand).
Tier 3 — No value accrual: OLAS ($89K total marketplace volume, 99.6% price decline). ELIZAOS (no staking, no fees, no burns, foundation dissolved). Most AI agent tokens catalogued in the awesome-crypto-ai-agents list.
The x402 protocol creates a Tier 0 baseline: machine-to-machine payments that require no token, charge no fees, and settle in stablecoins. Every AI agent token must now justify why it exists alongside this free alternative.
The AI agent token sector is bifurcating. On one side: protocols that generate fees and distribute them to holders (Pendle, Uniswap), or offer genuine utility-for-stake (Venice). On the other: a growing cohort of AI tokens with no governance, no fee accrual, and no justification for existing when x402 settles machine payments in stablecoins for free.
The SEC's November 2 FAQ update gives this bifurcation regulatory teeth. Tokens that confer voting rights without genuine decentralization now face a clearer path to securities classification. Tokens that function as payment instruments face the simpler question of why stablecoins won't do.
For token holders, the message is structural: governance rights are worth something only when the protocol generates revenue, distributes it, and operates without a central party controlling the outcome. In AI x crypto, that combination remains rare.