The 100-Beneficiary Cap: A Forensic Reading of the Linqto SPV Case

$430 million is the number the SEC wants you to remember. I would argue the number that matters is 100.

Under §3(c)(1) of the Investment Company Act of 1940, a private fund escapes registration only if it holds fewer than 100 beneficial owners. That is the line. Cross it, and federal law reclassifies you as an investment company — with the disclosure regime, the independent governance, and the auditable books that status demands.

Linqto allegedly crossed it. Not once, but structurally. The mechanism, per the SEC's complaint filed October 9, 2026, against former executives Sarris and Endoso, was a lattice of parallel special purpose vehicles, each engineered to sit just below the threshold, each insulated from its siblings on paper, all of them feeding the same retail sales engine.

Three charges. Unregistered securities under §5 of the Securities Act. Fraud under §10(b) and Rule 10b-5. And the one that gets buried in the coverage: operating an unregistered investment company under §7 of the ICA. Three independent paths to liability, stacked so that the collapse of any single theory does not rescue the defendants.

I have spent nine years tracing capital through systems built to look clean. This one isn't clean. But the rot isn't where the headlines point. The headlines point at the fraud. The structure is the story.

The machine nobody audited

Linqto's pitch was simple, and it was a pitch I have watched mutate across three market cycles. Private markets — the pre-IPO equity of companies like SpaceX, Stripe, Databricks — had generated enormous returns for institutions and accredited insiders. Retail investors, locked out, wanted in. Linqto positioned itself as the bridge. Buy into a special purpose vehicle; the SPV holds a slice of the private company; you get exposure without a $250,000 minimum and without the paperwork.

This is not a novel structure. SPVs have been the plumbing of private markets for two decades. Pension funds use them. Family offices use them. The difference is that institutional SPVs are sold to sophisticated buyers who can interrogate the terms, and they are sold privately, under Reg D, to parties who understand what a capital call is and what a waterfall does.

Linqto allegedly took the same plumbing and pointed it at retail. That reorientation changes everything, because the regulatory exemptions that make SPVs viable are conditional on who is buying. Sell to an accredited investor under Rule 506(b) and you are operating inside a defined perimeter. Sell to a non-accredited investor, market publicly, and you have left the perimeter entirely. The exemption does not travel with the structure. It travels with the buyer.

The SEC's complaint alleges all three failures at once: sales to non-accredited investors, public marketing of what was characterized as a private placement, and an SPV architecture engineered to evade the 100-beneficiary cap that would have forced registration. Each of those is a separate violation. Together, they describe a platform that was operating outside the regulatory perimeter while presenting itself as a compliant bridge into private markets.

There is a specific detail in the allegations that I want to isolate, because it is the kind of detail that determines outcomes. The complaint asserts that Linqto continued to claim compliance even after internal awareness of non-compliance. That single fact does more work than any other in the filing. It is the difference between a company that misunderstood its obligations and a company that understood them and proceeded anyway. The first is a compliance failure. The second is scienter — the knowing state of mind that transforms a regulatory matter into a fraud case.

The lattice, and why it worked on paper

Here is where the forensics get interesting, and where I want to slow down.

A single SPV holding pre-IPO shares is, legally, a straightforward vehicle. It has a manager, it has investors, and under §3(c)(1) it can hold up to 100 beneficial owners before it must register as an investment company. The 100-owner limit exists precisely to keep these vehicles small and private — to ensure that a structure with real economic weight cannot hide behind a private-placement label while behaving like a public fund.

The alleged evasion was to run many SPVs in parallel. If each vehicle holds fewer than 100 beneficial owners, each one is facially compliant. But if the same operator runs dozens of these vehicles, all holding the same underlying asset, all sold through the same platform, to the same retail audience, then the aggregate is functionally a single investment company that has been sliced into regulatory invisibility.

The code doesn't lie, and neither do the cap tables — once you reconstruct them.

This is the pattern I recognize from my own work. In 2017, tracing the Parity Wallet hack, I spent four weekends manually stitching together wallet clusters that the attackers had deliberately fragmented. The attackers' entire strategy was structural: split the funds across 14 clusters, move them through mixers, and rely on the assumption that no one would bother to reassemble the whole. The defense was not encryption. It was fragmentation plus the hope that nobody would do the tedious work of reconstruction.

The SPV lattice is the same move in a different register. Fragment the beneficial ownership so that no single entity trips the alarm. Rely on the fact that the 100-owner test is applied vehicle by vehicle, not in aggregate. And trust that the sales operation looks like dozens of small private placements rather than one large public offering.

The SEC's decision to charge both the unregistered-securities path and the unregistered-investment-company path tells me they did the reconstruction. They added up the SPVs. They counted the owners. And they concluded that the aggregate breached the threshold that the fragmentation was designed to stay under. When a regulator pursues two structural theories that only cohere if you have done the tedious aggregate math, it means the tedious aggregate math is already in the file.

The pricing problem is a data integrity problem

The fraud charges are the ones that will dominate the coverage, and they deserve a forensic reading of their own, because the specific allegations are unusually falsifiable.

The SEC alleges that Linqto told investors its pricing was set by an algorithm, and that in reality pricing was manual. It alleges that the platform showed shares as "sold out" when they were not, to manufacture scarcity. And it alleges that investors were told they were buying below market when the pricing was above fair value.

Read those three together and you see a single, coherent failure: the platform's front end was a presentation layer that did not correspond to the back end's actual logic. The interface said "algorithm." The reality was a person with a spreadsheet. The interface said "sold out." The reality was inventory still on the shelf. The interface said "below market." The reality was a number that no independent valuation could support.

This is, at its core, an oracle problem. And it is the exact problem that every tokenization project in the current cycle claims to have solved.

When I audited Aave's governance in 2020, I wrote a script to scrape 5,000-plus on-chain voting records and correlate voter wallets with proposal outcomes. The finding that mattered was not that voting happened. It was that 15% of voting power sat with 12 entities. The system was transparent — every vote was on-chain — but transparency of the mechanism did not mean transparency of the control. The display said "community governance." The data said "twelve wallets."

Linqto's display said "algorithm." The data, if the SEC is right, said "a human deciding what to show you." Different era, different technology, identical structure: a presentation layer optimized for trust, layered over a back end optimized for extraction.

The reason this is falsifiable — and the reason I think the defendants are in serious trouble — is that manual pricing leaves fingerprints. An algorithm produces a distribution. It has a variance profile. It responds to inputs in a way that can be modeled. Manual pricing, dressed as algorithmic pricing, produces a distribution that looks like a human made choices: clustered round numbers, price changes that coincide with business events rather than market events, and timestamps that betray deliberation.

I have watched teams try to fake algorithmic behavior by scripting outputs. It never holds up. The autocorrelation is wrong. The response to volatility is wrong. A real pricing engine breathes. A fake one sits still and then lurches. Internal system logs, message queues, and deployment records will show exactly which one this was. The complaint's willingness to allege it publicly suggests the SEC already has the logs.

The evidence chain is internal, and that is the point

Here is what the coverage is missing. This is not a case built on investor testimony or reconstructed intent. It is a case built on the platform's own records — and that makes it, in the language I use for my own work, a high-confidence evidence chain.

Internal emails establish the gap between what was said publicly and what was known privately. System logs establish whether pricing was algorithmic or manual. Cap-table records establish how many beneficial owners each SPV actually held. Marketing materials establish whether the offering was private or public. And the platform's own repeated compliance claims, made even after internal awareness of non-compliance, establish scienter — the knowing state of mind the fraud counts require.

Between the hash and the human, there is a silence. This case is about what fills that silence.

In on-chain forensics, the chain records what happened but not why. A wallet moves funds; the ledger notes the transfer but not the motive. The human intent lives in the gap, and the investigator's job is to close that gap with external evidence — communications, timing, counterparty relationships. The Linqto case is the inverse. The human intent is documented in emails, and the job is to reconcile it with what the systems actually did. Either way, the verdict comes from the same place: the space between what the machine recorded and what the person meant.

The §5 trap

The fraud charges will generate headlines, but the §5 charge is the one that should terrify the defense, because §5 is a strict liability provision.

The SEC does not need to prove intent, knowledge, or fraud to establish a §5 violation. It needs to prove that a security was sold without registration and without a valid exemption. That is the entire test. This decouples the case from the hardest thing in securities litigation: scienter. In a fraud case, the SEC must show the defendants knew the pricing was fake, knew the "sold out" claims were false, knew the market-price comparisons were misleading. That is a subjective inquiry, and subjective inquiries are where defendants win.

In a §5 case, none of that matters. If the SPV interests were securities — and under Howey they almost certainly were, as investments of money in a common enterprise with profits expected from the efforts of others — and if they were sold to non-accredited investors without registration, the violation is complete. Intent is irrelevant.

The SEC structured the complaint so that the fraud counts and the §5 count are independent. Even if a jury rejects the fraud theory entirely, the §5 count survives. That is why the agency pursued both. It is not indecision. It is a redundancy engineered to guarantee a finding of liability. The defendants can win the argument about what they intended and still lose the case about what they sold.

The investment company path

The third charge — unregistered investment company — is the one most coverage will underweight, and it is the most structurally revealing.

If the SPV lattice is treated as a single investment company that never registered under §7 of the ICA, then Linqto was not merely selling unregistered securities. It was operating an unregistered fund. The distinction matters because the ICA imposes obligations that go beyond disclosure: independent governance, custody rules, valuation requirements, and periodic reporting. An unregistered investment company is, in effect, a fund that has opted out of the entire investor-protection apparatus that funds are supposed to carry.

This is why the parallel-SPV allegation is the structural heart of the case. The fragmentation was not a bookkeeping choice. It was the mechanism that allowed the operator to avoid the fund-level obligations that would have applied to the aggregate. Strip the fragmentation away and the structure is a large, retail-facing private fund with no registration, no independent oversight, and no valuation discipline. That is not a technical violation. It is the entire investor-protection regime, sidestepped.

The tokenization mirror nobody wants to look into

Now the part that should worry anyone building in this cycle.

The Linqto structure — private assets, wrapped in vehicles, sold to retail, priced by an opaque mechanism — is precisely what a generation of tokenization and real-world-asset platforms are rebuilding on-chain. The pitch is that blockchain fixes the trust problem. The SPV becomes a smart contract. The cap table becomes a token registry. The pricing becomes an oracle. The beneficial-owner count becomes a queryable variable.

I want to be precise here, because I am not a tokenization skeptic in general. I am a skeptic of the claim that moving a broken structure onto a ledger repairs the structure.

Start with the cap. A tokenized fund that represents beneficial ownership of pre-IPO equity faces the same §3(c)(1) question as a paper SPV. Does a token holder count as a beneficial owner? Almost certainly yes, if the token conveys economic rights. So a tokenized vehicle with 10,000 holders is not magically exempt because the holders are pseudonymous. If anything, the on-chain record makes the breach easier to prove, not harder — because the holder count is right there, queryable, immutable, and timestamped. The transparency that tokenization advertises is the transparency that prosecutes.

Then the pricing. Every tokenized asset needs a price. Pre-IPO equity has no public market, so the price must come from somewhere: an internal model, a secondary transaction reference, or a designated oracle. In every one of those cases, the price is a human judgment wearing a technical costume. The oracle does not discover the price of a private company. It reports what someone decided the price should be. Calling that decision "on-chain" does not make it less of a decision. It makes it a decision with a permanent audit trail.

And that is the twist. In the Linqto case, the SEC had to reconstruct the manual pricing from logs and emails. In a tokenized version, the pricing inputs, the oracle updates, and the resulting token prices would all be on-chain — meaning the gap between the marketing claim ("algorithmic") and the reality ("human") would be visible to anyone with a block explorer. The structure that hid behind fragmentation on paper becomes a structure that documents its own failures in real time.

We don't need better technology to commit this fraud. We need the same humans with a nicer interface.

The governance vacuum, restated

There is a second thread here that connects directly to the governance work I have been doing for six years, and it is the thread most coverage will skip.

The SEC charged individuals, not just the company. Sarris and Endoso are named personally. That is a deliberate choice, and it signals that the agency views them as control persons — the parties who directed the scheme rather than merely profited from it. When regulators reach past the corporate shell to the humans, it is because the shell is either insolvent or insufficient. The corporate veil is not a shield when the humans are the ones who built the weapon.

This is the same lesson I keep arriving at from the other direction. In DAO governance, the fiction is that token holders decide. In practice, a handful of whales and their delegates decide, and the token holders ratify. Turnout sits below 5% on most proposals. The "community" is a presentation layer over a concentration of control. Linqto's "platform" was a presentation layer over two people making the calls.

The difference is that a DAO's control concentration is visible on-chain to anyone who queries it, while Linqto's was buried in a lattice of SPVs. But the underlying question is identical: when a structure claims to distribute power, who actually holds it? And the answer, in both cases, is almost never the people the marketing describes.

The officer-and-director bar, and what it tells us

The remedy package the SEC is seeking deserves attention, because it reveals how the agency is sizing the case.

Injunction. Disgorgement plus prejudgment interest. Civil penalties. And an officer-and-director bar.

That last item is the tell. The officer-and-director bar is the most severe non-monetary remedy in the SEC's toolkit. It does not just punish; it removes the defendants from the industry. It is typically reserved for cases where the agency believes there is a risk of recurrence and a pattern of harm serious enough to justify permanent exclusion. Seeking it here means the SEC is not treating this as a technical registration failure. It is treating Sarris and Endoso as operators whose continued participation in financial markets is itself a hazard.

The disgorgement claim carries its own weight. The 2021 National Defense Authorization Act wrote the SEC's disgorgement authority into the Securities Exchange Act and set a ten-year limitations period for it. That matters enormously here. Before the NDAA, the SEC's disgorgement power was constrained by the five-year window the Supreme Court endorsed in Kokesh. With a ten-year reach, the SEC can potentially claw back sales going back to 2021 — which, on the alleged scale, puts the disgorgement base in the hundreds of millions.

But the defendants have one meaningful counter. Under Liu v. SEC, disgorgement cannot exceed the wrongdoer's net profits and must be directed to investors. That gives them room to argue for deductions — operating costs, amounts already returned to investors. It is a real defense. It is also a defense that collapses if the SPV flows are characterized as a circular transfer, because there are no legitimate net profits to deduct when the money was moving in a circle.

The forensic reconstruction is the whole case

Strip away the legal taxonomy and what remains is a reconstruction problem, and that is why this case reads differently to me than it does to a securities lawyer.

The SEC's job is to reassemble a structure that was designed to resist reassembly. Count the owners across the SPV lattice. Trace the pricing decisions to their human source. Establish that the marketing was public when it claimed to be private. And show that the people at the top knew, because the records prove they knew.

I have done versions of this work at a smaller scale — 14 wallet clusters after Parity, 50,000-plus BAYC transactions during the NFT peak, the divergence between UST's redemption rate and its market price before the Terra collapse. The pattern is always the same. The structure hides in fragmentation. The truth hides in the aggregate. And the analyst's entire value-add is the willingness to do the tedious reconstruction that the structure was betting nobody would do.

Volume spikes don't announce fraud. The absence of volume, matched against a claim of "sold out," does.

That is the detail I keep returning to. A platform that claims scarcity — "sold out," "below market," "algorithmic" — is making a claim about the relationship between supply, demand, and price. Each of those claims is testable against the platform's own data. If the order book was empty while the interface said "sold out," the discrepancy is the crime. If the price was manual while the interface said "algorithmic," the discrepancy is the crime. The fraud, in this reading, is not a lie told in isolation. It is a gap between two systems that the operator controlled — the one investors saw and the one the operator used.

The contrarian read

Here is the angle I expect almost no one to take, and it is the one I believe.

The consensus lesson from Linqto will be: "This is why private markets need more regulation," or, from the crypto side, "This is why private markets need to be tokenized on-chain." Both are wrong.

The first is wrong because the regulations already existed. §5 existed. §3(c)(1) existed. The prohibition on public marketing of private placements existed. Linqto did not exploit a gap in the rules. It allegedly operated in direct violation of rules that were on the books and well understood. More regulation would not have stopped this any more than the existing regulation did. The failure was enforcement latency, not legislative absence.

The second is wrong because it mistakes a transparency problem for a technology problem. The claim is that tokenization would have prevented Linqto because everything would be on-chain and auditable. But the failures here were not failures of record-keeping. They were failures of honesty. The cap-table evasion, the fake pricing, the phantom scarcity — these are decisions made by people, and people make the same decisions whether the ledger is a database or a blockchain. Tokenization does not make a liar honest. It makes the lie permanent.

What actually distinguishes a trustworthy structure from an untrustworthy one is not the medium of the ledger. It is the alignment of incentives and the presence of adversarial verification — someone with the standing, the resources, and the motive to reconstruct the aggregate and check it against the claim. In Linqto's case, that someone turned out to be the SEC, and the reconstruction took years. In a tokenized world, the reconstruction could be instant. But only if someone bothers to do it. The tool is available. The will is not guaranteed. And a tool that no one uses is indistinguishable from no tool at all.

There is a deeper implication for the AI-agent economy I have been tracking this year. If algorithmic pricing becomes the default marketing claim for every platform — human-operated or machine-operated — then the verification burden shifts entirely to the auditor. We are already seeing a version of this in DeFi lending, where roughly 40% of activity is driven by arbitrage agents rather than humans. When the operators are agents and the claims are algorithmic, the gap between the claim and the reality becomes the single most important thing to measure. Linqto is the preview. The pricing was manual, the claim was algorithmic, and the gap was the fraud. In a machine-driven market, that gap will be the standard failure mode, not the exception.

Takeaway

Watch the sweep, not the verdict.

The Linqto case will take years to resolve, and the headline outcome will be a settlement or a trial that most people forget by the time it concludes. The signal that matters is what the SEC does next. If the agency treats this as a template — a demonstration that the SPV-and-retail model is prosecutable on structural grounds — expect a series of follow-on actions against comparable platforms in the next two quarters. The trigger will not be a new scandal. It will be this case's establishment of the reconstruction method.

For anyone building in tokenized private markets, the forward-looking question is not whether your ledger is transparent. It is whether you have run the aggregate count, checked your pricing against an independent reference, and confirmed that your public claims survive contact with your own data. The code doesn't lie. But the code is only as honest as the person who decides what it records — and the silence between the hash and the human is where every one of these cases is actually decided.