Invest1 distinct publisher3 min readUpdated
The complaint against The Spaventa Group is not really about fake stock. It is about undisclosed spread, which makes it a problem for anyone sourcing AI-era private allocations.
The Investor · Invest desk
Compiled by The InvestorSomething wrong?How this is made
The SEC filed a complaint on Friday in the Southern District of New York against The Spaventa Group, a Long Island firm run by former broker Andrew Spaventa, which the agency characterises as a pre-IPO "boiler room" [1]. The allegation that should concern anyone buying private-company exposure is not that the shares were phantom: it is that, per the SEC, investors paid on average 46% more for their positions than Spaventa's own companies paid to acquire them, with the premium reaching 91% in some cases [5].
The scale, as reported by Fortune's Amanda Gerut: more than 100 agents making thousands of phone calls, selling the most requested names in the market, including Anduril, Anthropic, Perplexity and SpaceX [2][16]. More than $74 million came in across 11 private funds operated from offices on Long Island and in New Jersey, over four and a half years from December 2020 to June 2025 [4]. More than 800 people bought in; most were retail, more than 650 put in $100,000 or less, and over 100 were retirees, according to the SEC [3]. That is roughly $92,500 per investor [9] and about $1.4 million a month for 54 consecutive months [14]. One caveat on the source: the same Fortune newsletter later summarises the buyers as "most with $100,000 or more," which contradicts the SEC-attributed figure a few paragraphs above it [15]. Use the complaint's number.
The markup arithmetic is the case. A 46% average premium means roughly 68 cents of every dollar an investor handed over corresponded to what the shares actually cost the sponsor, and roughly 32 cents did not [10]. At the 91% extreme, about 52 cents [11]. If that average held across the full raise, the implied spread is about $23 million on roughly $51 million of cost [12]. Investors allegedly had no idea the markups were that high, and the pitch included a promise of no rip-offs from "unnecessary fees" [5]. Spaventa denied the SEC's claims when Fortune reached him by phone [6].
For comparison, the regulator's earlier pre-IPO action this year against Giovanni Pennetta alleged $10 million misappropriated while selling fraudulent shares in companies including Anduril; Pennetta pled guilty to one count of wire fraud [7]. The Spaventa figures are about 7.4 times larger [13]. Fortune's Allie Garfinkle, who has written about investors chasing phantom Anthropic shares, describes the secondary market as both massive and unregulated [8]. This is the first case at a scale that makes that description an enforcement fact rather than a warning.
The operating consequence is narrow and useful. The SEC's theory converts an undisclosed acquisition-to-sale spread into fraud [5], which means the two things worth demanding from any SPV or feeder offering a hot private name are the sponsor's own purchase price with a chain of title, and fee disclosure in writing that covers markup, not just management and carry. If a sponsor will not tell you what it paid, you are in the fact pattern the SEC just charged.
Watch whether the agency follows this with cases against other SPV sponsors and placement agents rather than only outright fabricators [7][1], and whether Spaventa contests the markup calculations or the disclosure claim [6]. Also watch the fee language in the next pre-IPO deck that lands in your inbox.
Follow any of these and your For You feed starts watching them — no settings page required.
Ranked by verification strength, evidence, and original report placement.
The SEC filed a complaint on Friday in the Southern District of New York against The Spaventa Group, a Long Island-based financial firm run by former broker Andrew Spaventa, which the SEC calls a pre-IPO "boiler room."
The SEC alleges Spaventa and his firm used a force of more than 100 agents making thousands of phone calls to sell shares in pre-IPO companies including Anduril, Anthropic, Perplexity and SpaceX.
More than 800 people bought in; most were retail investors, more than 650 put in $100,000 or less, and more than 100 were retirees, according to the SEC.
The alleged boiler room raised more than $74 million for 11 private funds run from offices on Long Island and New Jersey over four and a half years, from December 2020 to June 2025.
Despite a promise of no rip-offs from "unnecessary fees," investors paid on average 46% more for their positions than Spaventa's own companies paid to acquire them, with the premium as high as 91% in some cases, and investors allegedly had no idea the markups were so high, the SEC alleged.
Spaventa denied the SEC's claims when reached by phone by Fortune.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Specific figures, one publisher, allegations untested
Every number traces to one publisher's newsletter restating a colleague's story about a filed SEC complaint; the figures are unusually specific (dollar total, fund count, buyer bands, premium range, date window) and the filing itself is a checkable public document, which lifts evidence above anecdote. It is capped by three things: nothing here is adjudicated, the primary complaint is not supplied, and the piece contradicts itself on whether most buyers invested above or below $100,000.
No adoption signal in scope
The supplied material is an enforcement and market-structure story. It contains no release, deployment, benchmark, pricing, licensing or usage disclosure that could be read as adoption, and the deal items appended to the newsletter are unrelated funding announcements rather than evidence about this story. Inferring adoption from the number of buyers would conflate victim count with product uptake.
Framing runs slightly ahead of the allegation
The headline framing of an AI 'boiler room' scandal with 'massive hidden fees' implies fake or worthless stock, while the described allegation is narrower and more mundane: undisclosed spread over the sponsor's own acquisition cost. The dollar figures are real and attributed, which keeps the gap modest, but the piece presents unadjudicated allegations with scale-up language, generalizes from one Fortune-internal claim about an unregulated secondary market, and misstates its own buyer-size data in the recap line.
Self-referential newsletter, enforcement and defense both interested
The narrating source amplifies its own newsroom's story and explicitly cites the author's own prior coverage of phantom Anthropic shares as vindication ('a headline I'd been waiting to see'), which is a visible confirmation incentive inside a subscription deal newsletter. On the substance, the SEC is an interested party pressing charges and Spaventa is an interested party denying them, and no disinterested third party appears. Incentives are legible rather than hidden, which keeps this mid-range instead of high.
Moderate: verifiable filing, single lens, unresolved internal conflict
Confidence is moderate. The existence and contents of the complaint are the kind of fact that a single competent outlet reporting on a named docket usually gets right, and the derived arithmetic follows directly from the stated figures. But there is one publisher, one lens, no primary document in scope, no company response, and a self-contradiction on buyer ticket sizes that no other source can adjudicate.
invest
SEC's $74m pre-IPO case turns on the markup, not the access1 distinct publisher
invest
The two-year thesis: VCs are triaging portfolios, not reviewing them1 distinct publisher
product
OpenAI's CFO calls an IPO "another fundraise" while the model calendar slips2 distinct publishers
invest
Harvard's $2.2 billion SpaceX position is half its disclosed US equity book1 distinct publisher
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 18, 2026