$75 million. That is the number quoted in every headline about the SEC's proposed crypto exemption framework, and it is the exact number already available under Regulation A+ Tier 2 — same twelve-month window, same underlying disclosure philosophy. Run the comparison the other direction and the anomaly gets sharper. The startup carve-out in the same proposal caps issuers at $5 million across four years. Regulation CF gives an issuer $5 million every twelve months. Annualized, the new "relaxed" regime hands early-stage issuers $1.25 million — one quarter of what an existing, unremarkable, non-crypto-specific rule already permits.
That is the anomaly worth auditing. A framework marketed as an accommodation to an asset class that has spent a decade insisting it cannot fit existing exemptions turns out, on parameter inspection, to fit existing exemptions almost exactly — and in its earliest tier, more tightly than the rules it supposedly replaces.
Follow the gas. Always. Here the gas is the fee table, and the fee table says this is not deregulation.
Context: what was actually published
Release No. 33-11434 entered the Federal Register on August 21. Comments close October 20. Structurally the proposal does four things: it creates a startup exemption capped at $5 million over four years; it creates a tiered path permitting $75 million in any twelve-month period; it introduces a conditional safe harbor for the "investment contract" determination; and it attaches a Rule 104 disqualification provision that explicitly does not apply retroactively.
Two omissions sit inside that structure. A&O Shearman's read flags that the proposal creates no exemption for intermediaries. Rule 500 keeps dealers outside the preferred application. The SEC acknowledges the secondary-market gap — in a footnote, not in the operative text.
That drafting choice is not accidental. Admitting a regulatory void in a footnote preserves legal honesty while avoiding a formal record of the vacuum in the rule's body. It is the standard technique for minimizing political exposure. Anyone who has read enforcement-adjacent rulemaking recognizes the pattern: the agency knows, the agency says so quietly, the agency does not commit.
Meanwhile the parallel legislative track failed. The Clarity Act did not clear cloture, reportedly 49–50 against a 60-vote threshold. Market-structure legislation is dead for this Congress. At the state level, Colorado's algorithmic decision-making statute proceeds on its own schedule, imposing documentation duties, adverse-outcome disclosure within 30 days, and a 45-day human review with an independent reviewer empowered to overturn automated outputs.
And five securities-law academics — John Coates, former SEC General Counsel; John C. Coffee Jr. of Columbia; James D. Cox of Duke; Merritt B. Fox of Columbia; Joel Seligman, former Rochester president — published a joint objection arguing that crypto investment contracts require no separate treatment at all. When that particular configuration of names signs one document, it is not industry comment. It is a doctrinal challenge with a citation trail.
The parameters do not lie
The first thing I do with any disclosure package — a fund factsheet, a protocol treasury report, a regulatory release — is restate every headline number on a common per-period basis. In 2024, auditing six months of spot ETF flow data against price action, the entire finding emerged from that one normalization step; the raw daily prints were noise, the period-adjusted series was signal. The same move applies here.
| Reg CA parameter | Existing analogue | Direction | |---|---|---| | Startup: $5M over 4 years | Reg CF: $5M per 12 months | Tightened — annualized capacity is 25% of existing | | Tiered: $75M per 12 months | Reg A+ Tier 2: $75M per 12 months | Cloned — zero institutional increment | | Investment contract safe harbor | No precedent | The only substantive innovation | | Intermediary exemption | None | Structural gap | | Rule 104 disqualification | Non-retroactive | Amnesty for the existing stock |
The genuine benefit is not capacity. It is legal certainty — the safe harbor converts an unresolved Howey question into a defined status with conditions. That is worth something real to issuers. But it is worth nothing to the venue that has to list the asset afterward, because the venue got no exemption, no safe harbor, and no clarity.
This is where the five scholars and the rule text converge. Their argument was that crypto investment contracts need no bespoke regime. The parameter table confirms the SEC quietly agreed: it reused existing caps, existing disclosure logic, existing exemptions. If the capacity is copied, then the case for "this asset class requires special institutional machinery" collapses under its own drafting.
The fungibility paradox
Here is the technical core of the gap, and it is a mathematics problem before it is a legal one.
Consider token T. Sold by the issuer with promotional effort and a roadmap, T is very likely an investment contract. Sold anonymously on a secondary venue by a holder with no relationship to the issuer, T may not be. That bifurcation — institutional sales in, programmatic sales out — was the operative holding in SEC v. Ripple, and it produced a genuinely strange object: one asset, two legal statuses, determined by transaction history rather than by the asset.
Now put that object on a venue. Token T is fungible. There is no bit on the ledger that distinguishes "T that once moved through an investment contract" from "T that has only ever circulated freely." An intermediary asked to apply status-based treatment to a fungible instrument is being asked to do something the data structure cannot express. Its rational responses are three: refuse to list, delist broadly, or list a very short whitelist and absorb concentration risk.
The proposal does not cite Ripple. It does not engage the proposition that the token itself is not the security — a formulation SEC officials have used in public remarks. This is the largest analytical blind spot in the entire debate, and it is not a footnote-sized problem. Without a resolution of fungibility, an intermediary exemption would be a license to violate an obligation that cannot be discharged. Volatility exposes leverage; here it exposes the fact that the rule's silence is doing load-bearing work.
Where the liquidity actually goes
Map the incentives. Issuers gain a compliant path and their issuance appetite rises. Intermediaries gain nothing and their registration cost is unchanged, so listing and market-making appetite falls. Primary supply increases while secondary absorption capacity is constrained.
That imbalance resolves into one of three equilibria. Issuance onshore with trading offshore — classic regulatory arbitrage. A persistent liquidity discount on US-compliant assets, because venue scarcity is a cost. Or intermediary concentration, in which only entities large enough to absorb registration cost — the already-registered broker-dealers — carry the flow, which centralizes the market the framework claims to open.
There is a reflexive fourth effect the discussion consistently misses. Onchain-native DEX venues sit outside the registered-broker framework entirely. When a compliant CEX bears registration cost and a DEX bears none, the regulatory burden functions as a subsidy to the decentralized venue. Pressure pushes liquidity toward the exact venues the SEC is least able to supervise. A framework designed to domesticate the market may instead relocate it — not into a void, but into a jurisdiction-free substrate.
Colorado's human-in-the-loop veto
The state-level design deserves separate treatment because it is, technically, the most aggressive accountability architecture currently in force anywhere. Documentation duties are procedural and enforceable. The 30-day adverse-outcome disclosure is close to infeasible for automated systems making thousands of decisions daily — per-decision notice does not scale. But the 45-day human review with an independent reviewer holding override authority is a hard structural constraint: it forces a human veto point into an automated pipeline.
For a lending agent or a risk-scoring model, that is implementable. For a high-frequency execution agent, it is not. The predictable outcome is bifurcation: either compliance cost is passed to users, or agent services exit the state. Neither is a consumer-protection win, and both are measurable if anyone bothers to instrument them.
The missing comparator
Now the frame error. This is consistently described as a domestic vacuum. It is not empty. MiCA is fully applicable across the EU with CASP licensing in force; Singapore, Hong Kong, and the UAE have operating licensing regimes. Against that backdrop, the consequence of a US primary/secondary split is not market paralysis — it is export. Capital and builders do not wait in a void; they route around it. The comparative matrix, which the debate never draws, has the US behind on both issuance and secondary infrastructure simultaneously.
The contrarian read
October 20 is not a catalyst. The proposal's content has been public since August 21; the comment deadline is a procedural node with no new information embedded in it. If the market prices regulatory clarity, then the repricing already occurred in August and the deadline itself is noise. The genuinely negative event was the cloture failure, because it removes the second rail. The industry has now lost both federal paths at once — a fractured primary framework and no secondary framework — and only one of those was priced.
And the deeper contrarian point: the gap is not negligence. Closing intermediary coverage requires touching Section 18 preemption and the statutory definition of a dealer. That exceeds the authority of this rulemaking. The footnote is the most honest sentence in the document precisely because the agency cannot fix what it has correctly described. A regulator that admits a boundary in small print is behaving better than one that pretends the boundary does not exist — and worse than one that asks Congress to move it.
Data Integrity Check
Sources: SEC Release No. 33-11434 and Federal Register publication (primary, high confidence); A&O Shearman analysis (expert opinion, high); the five-scholar objection (academic, high); Colorado statute text (primary, high); DOJ non-intervention reporting (secondary, medium). Three items require verification before citation. The docket identifier referencing 2026 conflicts with an October 20 date that falls on a Monday in 2025 and a Tuesday in 2026 — one of the two is a transcription error. Colorado's algorithmic statute was enacted with a February 2026 effective date and deferred by special session to June 30, 2026; a January 1, 2027 date does not match my records and should be checked against the statute, not against commentary. The cloture vote count and date are directionally credible but unconfirmed. Sample sizes on the cited insurance and agent case studies are two — descriptive, not inferential.
Takeaway
Watch three signals next quarter, not one. Whether the first private suit names a US-facing intermediary over a secondary listing of a token originally sold as an investment contract — that is the litigation path filling the gap the rule declines to fill. Whether Rule 104's non-retroactivity survives comment, since that clause is the framework's moral price. And the spot volume ratio between DEX venues and compliant CEX venues for US-accessible pairs, because that ratio is the only honest measure of where the burden is actually landing. Code is law; math is evidence. The parameters already told us what this rule is.