Last week I was reading a crypto news feed and hit something strange. A story about the U.S. Securities and Exchange Commission's push to open private markets to retail investors. Roughly two thousand words. The word "blockchain" appeared zero times.
That absence is the story.
I have spent eight years taking protocols apart at the assembly level โ Gnosis Safe in 2017, the Compound rate curves in 2020, the ERC-721 metadata layer in 2021. When a crypto-native outlet runs a policy piece with no crypto in it, that is not an editorial slip. It is a positioning move. Someone is building the runway before the plane lands.
The narrative being positioned is asset tokenization. The SEC, under chairman Paul Atkins, has proposed two changes: loosening the definition of "accredited investor" to include licensed professionals such as CPAs and CFAs, and permitting registered investment advisers to charge a 20% performance fee on private-market products. The proposal is real. It is also, as of this writing, only a proposal โ a fact the source article omitted entirely.
Two gates are opening. The gate on who can buy. And the gate on what advisers can charge. Neither is a blockchain story. Both will be sold to you as one.
To understand what is actually being shipped, you have to understand the machine it lands in.
Private markets โ private equity, private credit, venture, real estate โ hold assets that do not trade on an exchange. There is no continuous price. There is no order book. A share of a private credit fund is not a token with a bid. It is a contractual claim on a pool of loans whose value is estimated by a model, refreshed monthly or quarterly, and signed off by an administrator.
This is the world that "retailization" wants to open. The vehicles already exist: business development companies, interval funds, and evergreen funds. Evergreen funds are the important one. Unlike a traditional private equity fund that locks capital for ten years and returns it at the end, an evergreen fund offers periodic redemption โ usually quarterly, usually capped at a small percentage of net assets.
That cap is not a technicality. It is the entire architecture. The redemption cap is a gate. And the gate is the only thing standing between an illiquid asset and a bank run.
Now add two changes. First, more buyers โ CPAs, CFAs, and a broader class of "qualified" individuals who clear a credential bar rather than a wealth bar. Second, a fee structure that pays advisers 20% of the upside. In traditional private equity, this carried interest is standard because the limited partners are institutions that can negotiate, diligence, and wait. Here, the same fee is pointed at a customer base with shorter time horizons and weaker information.
The mechanics matter more than the marketing. A 20% performance fee is not a flat cost. It is a lever. It converts every basis point of reported return into adviser revenue. And the reported return is produced by a model. Hold that thought.
One structural note before the analysis. Private-market products are distributed, not discovered. You do not stumble into an interval fund; an adviser places you in one. That single fact determines where the value of this policy lands, and I will return to it.
The political frame matters too. Atkins was confirmed in 2025 and is a known deregulatory, crypto-friendly appointment. This proposal is not a neutral technical adjustment. It is one plank of a broader agenda to loosen the rules governing U.S. capital markets. That makes the policy powerful and reversible at the same time. A rulemaking that arrives by political will can leave by political will. Any model that treats the rule as permanent is mispricing a political variable as a technical constant.
Let me disassemble the engines this policy depends on. None of them is a front-end app. All of them are unglamorous, and all of them are where the risk lives.
Engine one: valuation. This is the oracle problem, and it is older than crypto. When I audited the interest rate models at Compound and Aave, I kept running into the same uncomfortable truth. The "market" rate those protocols published was not discovered by supply and demand. It was a curve someone chose โ a kink at a utilization threshold, a slope, a base rate. Elegant. Deterministic. And arbitrary in the sense that matters: the parameters were governance decisions dressed as mathematics.
Private-market NAV is the same move, scaled up. Mark-to-model valuation takes an asset with no observable price and assigns it one using comparable transactions, discounted cash flows, and โ critically โ management's own assumptions. There is no oracle to query. There is a spreadsheet and a committee.
Certainty is a bug in a stochastic world. A model that returns a single number for an asset that has no single price is not measuring value. It is manufacturing it.
Now stack the performance fee on top. The adviser's compensation is a function of the NAV. The adviser influences the inputs to the NAV. This is not a subtle conflict; it is a direct one. In my 2024 work auditing a financial institution's custody stack, I found the same pattern in a different costume: convenience had been prioritized over verification, and the verification layer was a signature, not a proof. The private-market equivalent is an administrator's signature on a valuation the adviser helped construct. Vested interest distorts the lens of analysis โ and here the analyst, the seller, and the fee collector are the same desk.
Engine two: the fee accrual. Carried interest sounds simple โ 20% of the gain. It is not. You must define the gain against what baseline, crystallize it at what interval, handle clawbacks if later marks fall, and decide whether the fee is computed on realized or unrealized appreciation. Every one of those choices is a parameter, and every parameter is a place where the number can be bent.
In the DeFi summer of 2020 I wrote about the "ethical debt" of yield farming, and the backlash taught me something I still use: when an incentive mechanism is complex enough that its participants cannot compute their own outcome, the mechanism is doing work the participants cannot see. A carried-interest engine running on an illiquid asset is exactly that. The fee is real and immediate. The valuation it is computed against is soft and revisable.
Engine three: accreditation verification. The policy expands who qualifies by credential โ CPA, CFA, and similar. That turns a wealth test into a verification problem. Someone has to check the credential, continuously, at onboarding and on an ongoing basis. This is where I expect the most under-priced infrastructure to appear: an accreditation-verification API market, sitting between professional bodies like the AICPA and the CFA Institute and the platforms that need to gate access.
The regulatory logic here is quietly radical. It says, in effect, that wealth is not a proxy for investment maturity โ a credential is. That is a defensible idea. It is also a bet that a CPA understands private credit. Most do not. They understand accounting, not the correlation structure of a leveraged loan book. The credential proves diligence, not domain expertise.
Engine four: distribution. This is the one the technologists always miss. Private-market products are placed, not found. The moat is not the product โ products in this space are highly substitutable โ it is the channel. Whoever already holds the client relationship captures the policy's value.
The ultimate channel is the retirement account. If private assets ever enter the 401(k) menu at scale, the retail penetration curve goes vertical. That is the real unlock, and it is gated by infrastructure that does not yet exist: daily-valued NAV, liquidity buffers, and suitability engines that can survive a regulator's scrutiny. Until that plumbing is built, "retail access" means a wealthy professional clicking through an interval fund. Not a mass market. A credentialed niche.
And be skeptical of the scaling story. Private markets are low-frequency, non-standard, and relationship-driven. The consumer-internet logic โ more users, more data, more value, exponential โ does not transfer. There is no viral loop in a ten-year lockup.
The liquidity architecture deserves its own engine, because it is where the source article went silent. Every one of these vehicles rests on a duration mismatch. The investor can ask for money back on a schedule. The underlying asset cannot be sold on that schedule. Evergreen and interval funds paper over the gap with redemption caps. This works โ until it does not. In 2008, non-traded REITs froze. In recent cycles, interval funds have hit their caps and told investors, politely, to wait. Silence before the block confirms the truth. The gate is the silence. When a fund suspends redemptions, it is not announcing a failure. It is announcing that the asset was never liquid and the schedule was always a fiction.
Here is where the crypto industry is about to fool itself, and I want to be precise about it.
The narrative forming in the tokenization crowd is that RWA โ real-world assets โ will fix private-market liquidity. Put the fund share on-chain, the argument goes, and it becomes transferable. Secondary markets emerge. The illiquidity problem dissolves.
It does not. Tokenizing an illiquid asset does not make it liquid. It makes the illiquidity programmable โ and faster to propagate. If the underlying loans cannot be sold, a token representing them has no honest bid either. You have not created a market. You have created a more efficient way to discover that there is no market.
Worse, you have added two failure points. The first is the valuation oracle. Moving NAV on-chain means an oracle must publish the number, and that oracle is a signer โ often a single signer, often the same administrator whose model produced the number off-chain. You have not removed the conflict. You have automated it. The protocol does not lie; the interface does. In this architecture, the oracle is the interface.
The second is settlement. The "decentralized" layer that RWA tokenization would run on is, in practice, a sequencer. One operator, ordering transactions, defining finality. I have said for two years that decentralized sequencing is a PowerPoint, and this is the case where it stops being academic. If the settlement layer for tokenized private credit is a single sequencer, then the gate is no longer a redemption cap. It is a liveness property controlled by one party. The same rebranding playbook that turned Ethereum projects into "Bitcoin Layer 2s" will turn a permissioned database into a "tokenized private market." The label changes. The trust model does not.
The honest version of this technology would publish NAV as a range with an attestation, not a point estimate with a signature. I have not seen a single proposal do that.
The policy is real, and the opportunity inside it is real. But the value will not accrue where the marketing points. It will accrue to whoever controls distribution โ the incumbent asset managers with existing clients and brand trust, not the startups that educate the market and get harvested.
My forecast is narrow and falsifiable. Watch the redemption gates. The first interval or evergreen fund that suspends redemptions at scale โ not a small cap breach, a real one โ will be the moment retail investors learn what "illiquid" means, and the moment the SEC's own fraud-prevention language collides with its deregulatory agenda.
The gate opens first. The oracle is audited last. Ask which one you are actually buying.
