Welcome to the 3x/weekly CEO Note for PRO readers of the 51 Insights digital asset newsletter.
Hey, it's Marc.
Crypto wanted a law. It got nine answers.
What happened: On Thursday the SEC's staff answered nine questions about crypto tokens. Two drew the headlines. A token buyback on a working network is not a promise of profits. And a liquid staking token, the tradable receipt you get when you stake crypto, can count as a commodity.
The timing matters. The Senate killed the Clarity Act, the bill meant to settle all of this in law, ten days earlier. And Hester Peirce, who ran the SEC's crypto work, leaves on October 2.
In plain English:
A token isn't a security by itself. The promise sold with it can be, as in "buy now, we'll build the network."
Once the team keeps that promise, the token is just a token.
On Thursday SEC staff said what a team can do after that point: keep building, market what works today, and buy back the token.
Liquid staking tokens now have a legal category, and some count as commodities.
None of this is law. It's staff guidance, and a future SEC can take it back.
The details 👇
Exhibit 1. Nine answers reduce to eight practical changes. Almost all of them switch on only after the network works, and handing promises to a foundation, a common exit plan, no longer ends the security. Source: SEC Division of Corporation Finance FAQ (Sept 25, 2026); SEC staff statement on liquid staking (Aug 5, 2025); 51 Insights analysis.
Once a network works, teams can build, market and buy back the token
Start with how a token becomes a security in the first place. In March the commissioners voted that the token isn't the security. The promise sold with it can be. "Buy now, we'll build the network" is a bet on the team. So for a token to stop being a security, some combination of three things has to happen. My understanding is that Thursday's answers speak to all three:
The team keeps its promises, and grades itself. Each issuer, the staff wrote, "determines the thresholds that must be met to achieve functionality."
The team promises less to begin with. Marketing what the product does today is fine, and so are "indefinite aspirational statements," as long as nothing promotes profit.
Someone else takes the promises over. That door is now closed. If a foundation assumes them, the security goes with them.
Two more answers drew the headlines. Once a network works, a team can buy back its token without that counting as a promise of profits. And a liquid staking token, the tradable receipt you get when you stake crypto, is just a receipt if the issuer can't lend out your deposit. If a protocol issues it, it can count as a commodity.
Builders loved it. Uniswap founder Hayden Adams called the answers "bangers."
So here's what this really is: a map of life after launch, drawn by staff, on top of a rule the commissioners voted for in March.
Why it matters: each issuer now writes its own exit test
Who decides when a network "works"? Each issuer, the staff wrote, "determines the thresholds that must be met to achieve functionality."
If you get to write the test, and the test is whether you kept your promises, the cheapest way to pass is to promise less. A team that promised "mainnet, an open-source client, 100 validators" can point at three facts and walk out. A team that promised to become "the leading DeFi platform" never can.
That has an odd side effect: the fewer promises you make, the sooner you're free. Lawyer Jacob Robinson warned that this "runs the risk of disincentivizing disclosure." Gabriel Shapiro put it more sharply: securities law is being "disrupted" by "incentivizing making fewer commitments to investors."
Of course you might expect the SEC to stop this before it gets silly. A measure the measured party gets to define is a textbook setup for Goodhart's law. But I actually don't know how much stopping is coming. Buybacks are the test case:
Two projects did 89% of this year's $638M in buybacks. That's the total through August 25, per Allium Labs data cited by the FT: Hyperliquid spent $370M and Pump.fun $200M ($570M ÷ $638M).
They're opposite cases. Hyperliquid's buybacks run in code, "in a fully automated manner," its docs say. Pump.fun is a company that chose its own policy: all revenue for nine months, then half from April.
Peirce reads it more narrowly than the text does. She replied: "if you have a central party, you can't rely on this FAQ." The buyback answer itself never says that. Plan around Peirce anyway.
Pitching buybacks as yield before launch now counts against you. Before the network works, a buyback pitched "as creating yield or return" can be the very promise that makes the token a security. Decks that lead with buyback math just created evidence against themselves.
Now follow that one step further. If a token with a buyback gives holders most of what a share gives them, why sell tokenized equity at all? Shapiro asked exactly that: "if you can get people to buy a coin in the style of BNB, HYPE, PUMP, etc., with minimal regulation, why voluntarily take on the burdens of selling them equity?"
That's a real threat to the tokenized-stock story everyone is excited about. The bigger trend, in his words, is "get all the benefits of equity with none of the burdens."
We think that second-order effect will matter more than the buyback headline.
Exhibit 2. Two projects did 89% of this year's $638M in buybacks ($570M ÷ $638M). Hyperliquid's run on code; Pump.fun's policy is set by a company. The FAQ treats the two the same way. Peirce's reading would separate them. Source: Allium Labs data via Financial Times; Hyperliquid documentation; CoinDesk; 51 Insights analysis.
PS: Your next client could be reading this.
We help companies such as BCG, Avalanche, Coindesk, MoonPay and others create AI-native growth engines that build category mindshare, reach a 100k+ B2B digital asset audience, and generate qualified enterprise demand.
Between the lines
Two people can undo it. The FAQ has "no legal force or effect," in the staff's own words. The March framework can be withdrawn by another vote. From October 2 the SEC has two commissioners. Shapiro warned that if Democrats later take control and try to undo this "after an entire market has organized around it, the resulting chaos will be something to behold."
Courts aren't bound by staff answers. Investors can still sue, and judges apply the Supreme Court's old Howey test. Shapiro noted that a federal judge has already held that "certain forms of incentive alignment" can be enough, promise or no promise. Jennings said: "No way this holds up in court or any future administration."
We've seen this movie once. In June 2018 the head of the same SEC division said in a speech that Ether's network had become decentralized enough that Ether wasn't a security. The industry built around that line for years. It came with the standard caveat that the views were his own, and it didn't stop the lawsuits that followed. I'm guessing 2018 isn't quite the right reference class, because this time a Commission vote sits underneath. But the FAQ itself is the same kind of paper as that speech.
Only a final rule makes this durable, and comments close October 20. The SEC proposed that rule, Regulation Crypto Assets, in August, with a safe harbor for teams that finish what they promised. Comments close October 20. Undoing a final rule takes a whole new rulemaking.
Exhibit 3. The staff FAQ rests on a Commission interpretation that a future majority can withdraw by vote. Only a final Regulation Crypto Assets, with comments due October 20, would make a reversal slow. Source: SEC; Federal Register; Yahoo Finance; CoinDesk; 51 Insights analysis.
Subscribe to Pro to read the rest.
Become a paying subscriber of Pro to get access to this post and other subscriber-only content & join 35k+ others.
Upgrade NowA subscription gets you:
- Stop wasting 15+ hours/week reading noise; get the 3 signals that matter 5min/week, zero research
- Build your digital asset playbook with the same intelligence as BlackRock, DTCC, and Goldman
- Access to full archive of 700+ publications & conversations
- Early access to 51 Terminal