On 26 August, Sebastian Siemiatkowski bought 692,506 Klarna ordinary shares at $14.3669, a total of $9,949,164. He filed it two days later, on the SEC’s deadline, and the filing carries no Rule 10b5-1 flag, so this was a decision rather than a scheduled trade.

Klarna is the Swedish company behind the “pay in four” option that sits at checkout on a large slice of the internet. Three Stockholm School of Economics students founded it in 2005, it won a full Swedish banking licence in 2017, and it now pushes $36.6 billion of purchase volume through 120 million consumers in a single quarter.

Shoppers split a purchase into interest-free instalments or defer it, and Klarna takes a fee from the merchant on the theory that people buy more when they can pay later. Siemiatkowski co-founded it and has run it ever since.

The company listed on the New York Stock Exchange on 10 September 2025 at $40 a share, raising $1.37 billion in the largest IPO of the year. That priced it at about $15.1 billion, already a heavy markdown from the $45.6 billion private investors had put on it in 2021. It now trades at $14.20.

Subscribe now

So the CEO has just spent $10 million on a stock that has lost roughly two thirds of its value since he rang the bell. To see why that is more interesting than it sounds, look at what else is in Klarna’s SEC filing record, which is not much.

Eighteen of Klarna’s nineteen Form 4s are the company handing shares to executives

Klarna’s insiders have filed 19 Form 4s since the company began Section 16 reporting in March 2026, days after its IPO lock-up expired on 9 March. I read all of them.

Eighteen carry transaction code A or M: equity grants and option conversions. The CFO Niclas Neglén, COO Camilla Giesecke, CTO Yaron Shaer, CMO David Sandström, CCO David Sykes, product chief David Fock and chief accounting officer Anthony Greenway appear repeatedly, always receiving. Siemiatkowski’s own two earlier filings are option grants in March and June.

The nineteenth is the August purchase. It is the only code P in the file. There is no code S anywhere: not one Klarna insider has reported selling a share.

Two officers came close, Sandström and Sykes each filed a Form 144 in March, notices of intent to sell 32,703 and 23,799 shares through J.P. Morgan under 10b5-1 plans adopted the previous December, worth $449,666 and $347,941. Together that is about $798,000 against a company then worth roughly $7 billion. No corresponding Form 4 sale ever appeared.

The stock fell 22.8% in one session after a quarter Klarna beat

Klarna’s Q2, reported 18 August, was good. GMV rose 18% to $36.6 billion, revenue rose 27% to $1.042 billion against a $997 million consensus, transaction margin dollars rose 42% to $446 million, and adjusted operating income tripled to $91 million. Credit provisions improved to 0.52% of GMV from 0.56%. Klarna Card went from 1.3 million active users to 6.5 million.

Then the company cut full-year revenue guidance to $4.08–4.16 billion from over $4.34 billion, and GMV guidance to $149–151 billion from more than $155 billion, blaming about $600 million of currency headwinds and a consumer slowdown in Germany, its largest market by volume. It also said its CFO of six years and CMO of nine would move on in early 2027.

Share

The stock closed at $15.06 that day, down 22.8% from $19.51 on 34.5 million shares, roughly nine times its recent average. It hit $14.00 two sessions later. Analysts cut hard: UBS to Neutral at $16, JPMorgan to Neutral at $18, Morgan Stanley to $17, Bernstein to $16.

Siemiatkowski paid $14.37, below every one of those targets, and 64% below the $40 the company listed at in September 2025.

His 17.5 million options need $91.80, and Klarna’s own report calls them materially out of the money

This is the part that reframes the purchase. In March 2025 the board granted Siemiatkowski 17,505,672 Class C share options vesting over four years. Expressed per ordinary share, the weighted average exercise price is $91.80, a 130% premium to the IPO price. The company’s own remuneration report states that the options “are materially out of the money and no value is realisable by Mr Siemiatkowski unless and until the ordinary share price exceeds the exercise price.”

At $14.20, Klarna would need to rise 546% for that package to be worth a cent.

His 2026 grants are barely better. March: 1,628,701 Class C options struck at $13.04 per ordinary equivalent. June: 1,166,563 struck at $17.90, already underwater on the day they were granted.

So the CEO’s entire incentive structure pays nothing at anything near today’s price. The 692,506 shares he bought at $14.37 are the only equity he holds that gains a dollar when the stock goes up a dollar. Read that either way you like: a man backing his own turnaround, or a man whose paper compensation has stopped functioning and who is buying the instrument that still works.

He owns no Klarna directly, and the vehicle that does also holds OpenAI and SpaceX

The Form 4 records the purchase as indirect, held through Flat Capital AB (publ), the Stockholm-listed investment company Siemiatkowski founded with Nina Siemiatkowski in 2013. He holds no Klarna shares in his own name. Flat Capital’s stake is now 25,344,322 shares, about 7% of Klarna and roughly $360 million, and the company reported the purchase to its own market as SEK 95 million.

Flat Capital has more than 25,000 shareholders of its own and holds stakes in OpenAI, SpaceX and Cerebras alongside Klarna. That is not the same as a founder writing a personal cheque, and anyone reading this as pure conviction should account for it.

Buys that break a long drought return +2.2% over 90 days; routine buys return nothing

Across ten years of Form 4 data, we find purchases at companies where insiders bought within the previous 90 days produce −0.4% alpha over the next 90 days against SPY. Purchases that break a drought of a year or more produce +2.2% at t=3.1. The pattern holds separately across 2020–22 and 2023–26, and it matches what Cohen, Malloy and Pomorski found in Decoding Inside Information: routine insider trading carries no information, and the returns sit with insiders who break their own pattern.

Klarna’s record contains no pattern to break. It contains grants, and now one purchase.

On August 11, 2026, Cardinal Infrastructure Group told investors it would earn thinner margins on much more revenue. The stock fell 36.2% that day and 42.2% across three sessions, from $60.00 to $34.66. On August 14, one day after Holzer and Holzer announced a securities investigation into the decline, seven insiders bought $8.54 million of Class A stock at a blended $38.54. Among them: founder and CEO Jeremy Spivey and CFO Michael Rowe Jr., neither of whom had ever reported a direct open-market purchase. The filings became public on August 17. The stock closed Friday at $43.50.

Every trade below is disclosed on SEC Form 4, filed within two business days of execution and aggregated at insidersignal.ai.

Subscribe now

Revenue guidance went up $210 million and the stock fell 36%

Second-quarter revenue was $226.9 million, up 114%, roughly 64% of it organic. Backlog reached $866 million, up 35%. Management raised full-year revenue guidance to $880 million to $900 million from $680 million, a $210 million lift to the midpoint against consensus near $718 million.

Then the margin line. Adjusted EBITDA was $28.1 million on a 12.4% margin, down from 18.6% a year earlier. The full-year target dropped to 16% to 18%, from the “20% plus” CFO Michael Rowe had reaffirmed on May 12 in those exact words. Revenue guidance was raised. Only the margin target came down.

Rowe attributed the compression to subcontracted labor and equipment rental costs “primarily in newer markets where we do not yet own full turnkey delivery capabilities.” Cardinal is winning more work than it can staff and equip, and renting the difference at someone else’s margin. Stifel and Oppenheimer both cut price targets on August 17, to $52 from $63 and to $70 from $80, keeping buy-equivalent ratings.

Three of the seven had never reported a direct open-market buy

Six executed on August 14, Wimmer on August 17. All seven are coded “P” for open-market purchase, all direct ownership. None of the filings indicates a Rule 10b5-1 plan, so none was scheduled in advance.

Spivey, Rowe and Anthony Wood Jr. held zero Class A directly beforehand, each ending with exactly what he bought. Spivey’s needs a caveat. Cardinal uses an Up-C structure, and his Schedule 13G reports 12,229,848 Class A-equivalent shares, about 45% of the class, held through LLC units paired with Class B stock. His economic exposure was never zero. What changed is that he committed $3.2 million of personal cash to the same security a public shareholder buys.

COO Benjamin Wood bought at about $51.30 in May and $39.48 in August, averaging down on his own operations. Director Ivy Zelman, a housing analyst at Credit Suisse First Boston from 1998 to 2007, has now bought on every leg down: the December IPO, $36.33 in March, $39.61 last week.

Holzer and Holzer announced its investigation on August 13, Block and Leviton on August 17, Hagens Berman on August 19. All three are plaintiff-firm solicitations. No complaint has been filed and no regulator has opened a proceeding. The offering they reference is Cardinal’s June 24 follow-on at $73.00. Every share was primary, and no insider sold into it.

The disclosure lag cost 1.8%. The signal was worth 10.9%.

The insiders paid $38.54 on average. A reader could not.

Six of the seven filings hit EDGAR on Monday, August 17. That day’s close was $39.23, sixteen cents above Friday’s close and 1.8% above the insiders’ blended price. That is the entry our backtester uses, the first price available to anyone who saw the filing when everyone else did. Friday, August 21 closed at $43.50, a 10.9% gain in four sessions against 12.9% for the insiders themselves.

CDNL still trades 27.5% below the $60.00 pre-earnings close and 54.9% below the $96.40 high set in late June. Across the same week, technology insiders sold $2.02 billion and bought $4.8 million. Cardinal was the largest buy cluster in any sector.

Q3 lands around December 2 and settles whether 12.4% was transitional

Rowe called the reset “transitional. It is not structural,” and named four second-half drivers, led by Allied Paving contributing three months of revenue at a 20%-plus margin once it closes in early October. Cardinal is paying about $120 million for Allied, roughly 5.5 times EBITDA, for a business running a 20.3% margin. Faced with a margin problem, it bought margin.

If Q3 recovers toward the 16% to 18% path, the insiders bought the bottom of a transition. If it does not, seven people bought the first leg down. The cluster printed at $34.50 and is quoted at $43.50, four sessions after anyone with an EDGAR bookmark could have seen it.

Subscribe now

All trades discussed are disclosed via SEC Form 4 filings, required within two business days of execution, available on EDGAR and aggregated at insidersignal.ai. Nothing here is financial advice.

The Gates Foundation sold its last Microsoft in the first quarter. Cascade has been buying garbage ever since.

On August 10 and 11, Bill Gates’ family office bought 501,664 shares of Republic Services for $108.1 million. Republic was up 1.7% on the year at the time. The S&P 500 was up 13.9% and closing at record highs on AI enthusiasm. Cascade Investment now holds 111,305,646 shares, or 36.4% of the company, worth roughly $23.8 billion.

Every trade here comes from SEC filings that insiders must submit within two business days. The purchase is Form 4 accession 0001104659-26-094887, filed August 12, and it is aggregated on insidersignal.ai.

Share

Cascade added 4,258 shares in three and a half years, then 1.5 million in three months

The position history explains why this trade matters more than its size:

Between February 2022 and August 2025, Cascade added 4,258 shares to a 110 million share position. That is a rounding error. Then it bought 1,488,814 shares for $310.5 million across three filings this year.

The August buy carried no 10b5-1 plan and came four days after earnings

Rule 10b5-1 plans let insiders schedule purchases in advance, which removes the timing signal. The aff10b5One flag on this Form 4 is set to 0, so this was discretionary.

Republic reported second quarter results on August 6 and raised full-year guidance. Cascade started buying on August 10, the next Monday. It executed seven tranches over two days at weighted-average prices of $214.43, $215.07, $216.15, $214.67, $215.68, $216.40 and $217.09, paying up as the stock rose rather than waiting for a better entry.

One distinction matters. Cascade files as a 10% owner, not as a director. Michael Larson, chief investment officer of both Cascade and the Gates Foundation Trust, has served on Republic’s board since October 2009 and chairs its Nominating and Corporate Governance Committee. Larson disclaims beneficial ownership of the shares.

Gates holds all three major waste companies across two separate vehicles

Cascade is Gates’ personal money. The Gates Foundation Trust is the endowment, and it runs a different book. Its most recent 13F shows $31.7 billion across 22 positions:

Republic Services does not appear in the Foundation Trust at all. Waste Management does not appear in Cascade. Between the two vehicles Gates has roughly $30.5 billion in trash collection, spread across all three major North American operators.

Microsoft is what got sold to get here. The Trust reported $4.76 billion of Microsoft in September 2025 and $3.72 billion in December. By March 31 the line was gone. Michael Larson runs both vehicles, so the desk that exited Microsoft was buying Republic weeks later.

Republic is shrinking its share count, so the stake grows without buying

Republic had 312,497,013 shares outstanding in April 2025 and 306,212,978 by July 2026. Its second quarter 10-Q shows $657 million of buybacks and $383 million of dividends in the first half, against a $3.0 billion authorization running through December 2026. The quarterly dividend rose to $0.625 from $0.580.

Cascade owns a larger percentage every quarter even when it does nothing, and a second holding makes the pattern explicit. Cascade filed a new Schedule 13D on AutoNation on February 10, stating that its ownership crossed 20% “solely due to a decrease in the number of shares of Common Stock outstanding as a result of the Issuer’s share repurchase program, and not as a result of an acquisition of shares.” By May the figure was 21.1%. Cascade’s 7,062,629 AutoNation shares never moved while the company retired 4.9% of its float in four months.

Subscribe now

Landfill methane is a subsidised profit centre

Republic’s 10-Q discloses a landfill gas-to-energy joint venture carried at about $305 million and Blue Polymers, a venture with Ravago that takes the olefins from Republic’s Polymer Centers. Second quarter tax expense reflected $41 million of net benefits from renewable energy investments, $78 million for the half, plus $10 million from renewable natural gas and electric vehicle infrastructure.

Landfills are among the largest methane sources in the United States. Capturing that methane earns Section 48 investment tax credits, so the founder of Breakthrough Energy owns 36% of a company that gets paid to do it.

Republic’s own executives have not joined him

Cascade is the only insider buying. The only other Republic insider activity in 2026 is director Sandra Volpe selling $380,304 in June and chief accounting officer Elyse Carlsen selling $77,877 in May, both small enough to be routine vesting sales. The full ledger is public.

What to watch

The Foundation Trust’s next 13F is due in mid-August and will show whether Waste Management got added alongside Republic. Republic’s buyback authorization expires December 31, 2026, and a renewal keeps the share count falling and Cascade’s percentage climbing.

Gates has committed to spending down roughly $200 billion and closing the foundation by 2045. The two portfolios funding that plan are concentrated in garbage collection, freight rail, and heavy equipment, which is a long way from the trade the rest of the market is making right now.

On August 5, 2026, at 4:02 p.m. Eastern, Pfizer director Mortimer Buckley reported buying 37,632 shares at $25.52. Sixty-two seconds later, fellow director Ronald Blaylock reported 39,231 shares at $25.46. The two filings carry sequential accession numbers and went through the same filing agent in the same sitting, for a combined $1,959,190.

They are the two largest open-market insider purchases in Pfizer’s recorded filing history. Pfizer’s executives and directors have sold $193 million of stock and bought $3.2 million since 2016.

Share

Pfizer insiders have sold $193 million and bought $3.2 million since 2016

Our records hold 89 Pfizer sale filings totaling $193.0 million against 8 purchase filings totaling $3.2 million. Insiders have sold sixty dollars for every dollar they bought.

Former CEO Ian Read sold $22.9 million at $43.17 in March 2019 and $18.2 million at $37.36 in July 2018, part of $65.1 million across six filings. Albert Bourla sold $5.6 million at $41.94 on November 9, 2020, the day Pfizer announced its vaccine was more than 90% effective, cutting his stake 62%. That sale ran through a 10b5-1 plan adopted the previous August, so it was pre-scheduled and legal. It still drew a call for an SEC review. Mikael Dolsten sold $27.1 million across ten filings, Frank D’Amelio $28.8 million across six, Sally Susman $15.9 million, and Douglas Lankler $11.9 million.

Four people have bought Pfizer on the open market in ten years. Ronald Blaylock has done it three times. Former FDA commissioner Scott Gottlieb has done it three times, for $79,401, $77,005, and $36,997. James Smith bought $51,427 in October 2017. Mortimer Buckley bought for the first time on August 5.

The two August 5 filings account for 61% of every insider purchase dollar in the set.

Ronald Blaylock has bought Pfizer three times and doubled his check size on the third

Blaylock’s first purchase came on August 10, 2020, 13,000 shares at $38.55 for $501,085, establishing the position from zero. His second came on February 14, 2025, 19,457 shares at $25.65 for $499,072, lifting his stake 150%. The August 2026 buy was $998,821 and lifted his stake another 141.6%, to 66,938 shares. He has roughly doubled the position with each purchase.

His February 2025 filing reported 32,457 shares held, and the August 2026 filing implies 27,707 held immediately before the buy. The 4,750-share difference has no Form 4 behind it in our data, because we ingest purchase and sale codes and not the gift and trust-transfer codes that move shares without a trade. He has no recorded open-market sale of Pfizer, but something reduced the position by 4,750 shares in that interval. The figures above are reported holdings, not evidence that he never parted with a share.

His only other Form 4 activity in our data is on the sell side. He sold CarMax stock in July 2020 and again in July 2021, taking out $2.6 million.

His 2020 purchase at $38.55 is down 31% on price. The February 2025 buy at $25.65 is up 3.4%. Our follow-score model rates him 75 with a 100% win rate across two resolved trades, tagged for consistent alpha and high conviction. Two resolved trades is a small sample.

Mortimer Buckley waited 22 months on the board before buying $960,369 of stock

Buckley ran Vanguard as chairman and chief executive from 2018 until he retired in 2024. He joined Pfizer’s board in October 2024 and sits on its Audit Committee and its Governance and Sustainability Committee. He is also a director of Boeing.

He has filed three Form 4s. The first two follow the same pattern. He bought $497,420 of Boeing at $226.10 in August 2025, opening the position from zero, then another $499,966 at $224.20 in March 2026, doubling it. Both were open-market purchases rather than grants.

The Pfizer purchase was $960,369, nearly twice the size of his opening Boeing position. He waited 22 months after joining the board before buying any stock. He ran Vanguard for six years and sits on Pfizer’s audit committee, which gives him more access to the company’s numbers than most directors have.

Pfizer is up 5.1% this year, and both directors bought the day after earnings

Pfizer sits 7.3% below its 52-week high of $28.55. The directors bought the day after second-quarter earnings, which came in ahead of expectations. Revenue was $15.03 billion, up 2.6% year over year and above consensus. Adjusted earnings per share were $0.77 against the $0.69 analysts expected. Management raised the bottom of full-year revenue guidance to a $60.5 billion to $62.5 billion range.

August 4 was the earnings print, and volume ran 50% above the 60-day average of 41.5 million. Both directors filed on August 5. Blaylock is up 4.2% on his cost basis and Buckley 3.9%.

Directors are generally restricted to trading in the window that opens after an earnings release, so the day after second-quarter results is the first practical opportunity many of them get. That constraint accounts for the date but not for the size of either purchase.

The directors locked in a 6.76% dividend yield ahead of a 2027 patent expiry

At Blaylock’s $25.46 cost basis, Pfizer’s $1.72 annual dividend is a 6.76% yield. Buckley’s basis puts him at 6.74%.

A yield that high usually means the market expects a cut, and Pfizer’s comes from its patent schedule. Ibrance loses exclusivity in 2027. Eliquis and Vyndaqel follow in 2028. Analysts put the revenue hole at $15 billion to $17 billion through 2030, widening each year, against roughly $14.6 billion in dividends paid over the last six quarters.

Both directors committed about a million dollars each at that yield, and Buckley sits on the audit committee.

Starboard Value took a roughly $1 billion activist stake in late 2024 and pushed Bourla on cost discipline and R&D productivity. It never won a board seat and exited in November 2025. Buckley joined the board during that same period.

Pfizer’s controller sold 3,278 shares on the same day

Pfizer’s SVP and Controller Jennifer Damico sold 3,278 shares at $25.40 on August 5, about $83,261, cutting her position 10.3%. She has eleven sale filings in our records and zero purchases, including a $51,400 sale in June 2026 at $25.70.

Operating executives hold stock because they are paid in it, and they sell for taxes and diversification. Directors who buy on the open market are spending their own money on shares nobody granted them. The two sides of a Form 4 feed carry different information.

The two directors bought 23.5 times what the controller sold.

Big pharma has recorded five insider purchases in 90 days, and two are these

We checked open-market purchases across 19 major pharmaceutical tickers over the last 90 days.

Two are these Pfizer buys. The other three are Zoetis directors in May: Frank D’Amelio at $501,344, Michael McCallister at $233,273, and Paul Bisaro at $151,750. There were no purchases at Merck, Bristol Myers, Eli Lilly, AbbVie, Johnson & Johnson, Amgen, Gilead, Biogen, Regeneron, or Moderna.

D’Amelio was Pfizer’s chief financial officer for more than a decade and appears in our Pfizer records only as a seller, including $6.8 million at $44.09 in July 2019 and $6.0 million at $40.00 in August 2018. He now buys Zoetis, the animal health business Pfizer spun off in 2013.

Senator Tommy Tuberville disclosed a Pfizer sale 824 days after making it

On the same afternoon the two directors filed, Tuberville disclosed a Pfizer sale of $15,000 to $50,000 that he executed on May 3, 2024. The STOCK Act allows 45 days.

That disclosure was one of 133 transactions Tuberville filed on August 5, every one of them past the deadline, with the worst running 880 days late. He is currently running for governor of Alabama and opposes a congressional stock trading ban.

Pfizer is a popular name on the Hill. Ro Khanna has traded it eleven times in 18 months, seven buys and four sells. Susan Collins bought $15,000 to $50,000 in February 2026. French Hill sold $100,000 to $250,000 on the last day of 2025.

A Pfizer director bought roughly a million dollars of stock and the public knew within hours. A sitting senator sold the same stock and the public found out 27 months later, on the same afternoon.

Four developments that would confirm or undercut the signal

Both directors are up about 4% after three sessions, which is too short an interval to judge.

Watch whether Blaylock buys again. He has bought at $38.55, $25.65, and $25.46, doubling the position each time rather than averaging in on a schedule. A fourth purchase at a similar size would extend the pattern.

Watch whether Buckley doubles his position. His Boeing sequence opened at roughly $500,000 and added roughly $500,000 seven months later. If the same rhythm holds at Pfizer, the follow-on lands around March 2027 and would indicate a considered position rather than a director meeting ownership guidelines.

Watch the dividend through the 2027 Ibrance expiry. The case for buying Pfizer at a 6.76% yield depends on that payout surviving the patent expiries, and the next several declarations will settle it.

Watch the sell side. Sale filings outnumber purchases 89 to 8 over the last decade, and a sustained shift in that ratio would make August 5 part of a trend rather than an outlier.

All trades referenced are public disclosures from SEC Form 4 filings and STOCK Act periodic transaction reports. Prices, share counts, and ownership changes come from the filings themselves. Track every Pfizer insider transaction as it files on the Pfizer company page.

Scribe Therapeutics priced its IPO at $15.00 a share on July 23, 2026, upsized to the top of its range (GlobeNewswire). The stock opened near $25 the next morning and hit $25.49 intraday, a 67% pop. A week later it closed at $18.36, still up 22% from the offer price but down 28% from the high (Investing.com). Six Form 4s and a Schedule 13D later, the real story isn’t that “insiders bought.” It’s which insiders, and which ones didn’t.

A Clinical-Stage CRISPR Biotech Backed by Sanofi and Eli Lilly

Scribe is a clinical-stage CRISPR biotech in Alameda, California, co-founded by Nobel laureate Jennifer Doudna. Its ELXR platform edits genes epigenetically, silencing them without cutting DNA. Lead candidate STX-1150 targets PCSK9 to lower LDL cholesterol, backed by preclinical data showing over 50% LDL-C reduction in non-human primates that held for roughly 18 months after a single dose under 1.0 mg/kg. First-in-human data is expected in the first half of 2027. The company holds a 2023 collaboration with Sanofi worth up to roughly $1 billion in milestones, a partnership with Eli Lilly’s Prevail subsidiary, and more than $25 million in CIRM grants across two preclinical programs. Q1 2026 showed a $17.4 million net loss against $49.7 million in cash, and management guides the combined cash and IPO proceeds into the first half of 2029.

Subscribe now

$155.5 Million Raised at $15 a Share, Option Fully Exercised

Scribe sold 8,580,000 shares at $15.00, and underwriters exercised their full 1,287,000-share option, bringing the IPO to 9,867,000 shares and $128.7 million gross (GlobeNewswire). Alongside it, Sanofi bought 500,000 shares in a private placement at the same $15.00 price, about $7.5 million, unregistered and separate from the public offering. Combined gross proceeds: $155.51 million. Total shares outstanding after both closings: 18,864,386.

Six Form 4s, and Real Buying Closer to $64 Million Once You Remove the Double Count

Six Form 4s hit around the IPO date, all at $15.00 a share unless noted:

Two of those rows are the same money. Carl Gordon is OrbiMed’s Managing Partner, and his personal Form 4 covers the identical 1,000,000-share block reported a second time through the OrbiMed entity filing under standard Section 16 attribution rules. Count it once: OrbiMed’s actual commitment is $15 million, not $30 million. That resets the honest total for fund and strategic buying to roughly $64 million against a $155.5 million raise: Avoro at $36.3 million, OrbiMed at $15 million, a16z at $5 million, Sanofi’s $7.5 million placement, and about $85,000 from the two named officers.

Share Insider Signal

The CEO Took Options, Not Risk, and Eli Lilly Got Diluted Instead of Doubling Down

CEO Benjamin L. Oakes filed a Form 4 too. It shows two option grants, 264,784 shares and 429,599 performance-based shares, both struck at $15.00, plus an earlier 12,969-share grant at $19.13. Transaction code “A,” cost basis zero. He bought no stock. Neither did the other named directors or co-founder Brett Staahl, who received option grants only. Co-founder Jennifer Doudna filed nothing in the window.

Eli Lilly is the more interesting gap. The prospectus disclosed a non-binding indication that Lilly, then holding about 12.4% of the company, might buy enough IPO stock to hold up to 10.9% post-offering, with the explicit caveat that it could buy “more, less or no shares” (MedCity News). Lilly filed no Form 4, no 13D, no 13G. Its post-IPO stake landed at roughly 6.5%, meaning the deal diluted Lilly rather than the pharma partner topping up. That kills the tidier version of this story where two drug giants doubled down side by side. Sanofi wrote a real check. Lilly did not.

Avoro’s 16.37% Stake, Disclosed Two Days After Everyone Else’s Filings

The number that anchors the whole story surfaced two days after everyone else’s filings. On July 30, Avoro’s Behzad Aghazadeh filed a Schedule 13D disclosing beneficial ownership of 3,088,888 shares, 16.37% of the company (TradingView). The stake breaks into three pieces: 697,650 shares from automatic conversion of Series B preferred at closing, 2,333,333 shares bought directly from underwriters in the IPO for roughly $35.0 million ($31 million through Avoro Life Sciences, $4 million through Avoro Ventures), and 57,905 shares bought in the open market for about $1.26 million.

The lock-up runs 180 days from the prospectus date, extended by Scribe to September 28, 2026. But it carves out an exception: shares bought in the IPO or bought in the open market after closing are free to sell before that date. Avoro’s converted preferred is locked. Its cash purchases are not. That distinction matters for anyone reading this as a hold-to-catalyst commitment rather than a large, liquid position a specialist fund can trim whenever it wants.

Why Avoro and OrbiMed’s Exit History Is the Real Case for “Smart Money”

Avoro, formerly venBio Select, reported roughly $8.39 billion in discretionary assets as of December 31, 2025, with $13.55 billion in total regulatory AUM, and a portfolio that has run through 32 IPOs including argenx, Syndax, and Karyopharm (Crunchbase; Tracxn). Aghazadeh chaired Immunomedics until Gilead bought it for roughly $21 billion in 2020, and Avoro has also exited Endocyte to Novartis and MyoKardia to Bristol Myers Squibb.

Carl Gordon has made the Forbes Midas List multiple times. His OrbiMed exits include Peloton Therapeutics, sold to Merck in May 2019 for $1.05 billion cash upfront plus up to $1.15 billion in milestones, announced the day before Peloton’s own scheduled IPO (MIT), and SpringWorks, sold to Merck KGaA. These are the exact profiles behind a “smart money anchored this deal” thesis. The caveat: anchoring biotech IPOs is a routine part of how crossover funds deploy capital, not a special signal reserved for winners. A good track record raises the odds. It doesn’t remove the risk.

Volume Has Halved Since the IPO, and No Analyst Has Set a Price Target Yet

Volume tells its own story. SCTX traded 540,000 shares a day on average right after the IPO. By July 30 that had dropped to somewhere between 150,000 and 325,000, well under half the early pace. The 67% pop is a week old. No sell-side analyst has published a price target yet, because the underwriters (Leerink Partners, Goldman Sachs, Guggenheim Securities, Wells Fargo Securities) are still in the standard post-IPO quiet period. That typically lifts around 25 days after pricing, putting the first coverage in mid-to-late August.

What Actually Changes the Read: August Coverage, the September 28 Lock-Up, and 2027 Trial Data

Strip out the double-counted OrbiMed filing and the story stops being “insiders bought Scribe” and becomes something narrower and more useful: two specialist biotech funds with real exit histories, plus a16z and Sanofi, put about $64 million behind a CRISPR IPO on day one. The CEO, the founders, and the other named officers took options, not risk. Lilly walked away with less ownership than it started with. What’s left to watch: whether analyst coverage in August backs the pop or the fade, whether Avoro and OrbiMed use their unlocked IPO shares to trim once September 28 passes, and whether STX-1150’s first human data in 2027 gives any of this a fundamental reason to hold rather than trade.

Data sourced from SEC EDGAR Form 4 and Schedule 13D filings, Scribe Therapeutics press releases, and public market data. Nothing here is investment advice.

On Wednesday, July 22, 2026, the House passed the Stop Insider Trading Act, 232 to 198. Every vote against it came from a Democrat. Thirteen Democrats voted yes. The bill bars members of Congress, their spouses, and their dependent children from buying individual stocks. It does not require anyone to sell what they already own, which makes the practical effect narrower than the headlines suggest.

What H.R. 7008 actually restricts

The bill prohibits purchases of “covered investments,” defined as securities issued by publicly traded companies plus synthetic equivalents like options and derivatives. Existing positions are grandfathered. Members can hold them indefinitely, or sell them after filing public notice 7 to 14 calendar days before the sale with the Clerk of the House or the Secretary of the Senate.

Diversified mutual funds and ETFs are exempt. So are small business interests and assets in trusts where no covered individual controls the trustee.

Penalties rise sharply from current law. Violations cost the greater of $2,000 or 10% of transaction value, plus any net gain realized, and cannot be paid from campaign funds or office allowances. The STOCK Act’s existing penalty is a $200 late-filing fee that ethics committees have routinely waived since 2012. No member has ever been criminally prosecuted under the STOCK Act. DOJ closed its last COVID-era trading probe, into Senator Richard Burr, on January 19, 2021, without charges against anyone.

H.R. 7008 takes effect 180 days after enactment.

Share

Why 198 Democrats voted against a stock trading ban

Section 3 of the bill is a national voter photo ID mandate, drawn from the SAVE America Act and written into the same legislation through the Rules Committee. It takes effect at 90 days.

Ranking Democrat Joe Morelle called it a sham with a poison pill. The Campaign Legal Center said the trading provisions fail to address the underlying problem because ownership survives untouched.

With the voter ID title attached, the bill cannot reach 60 votes in the Senate. Multiple outlets describe it as dead on arrival.

Subscribe now

Hawley’s Senate bill is the one that would end congressional trade data

Josh Hawley’s HONEST Act (S. 1498), formerly the PELOSI Act, mandates divestment rather than a purchase freeze and extends to the president and vice president. It advanced out of committee 8-7 on July 30, 2025, and has never received a floor vote.

Hawley’s own substitute amendment pushed divestment to each official’s next term. Rand Paul objected that this shielded Trump specifically. The amendment passed 8-6.

The purchase ban starves the buy side and accelerates the sell side

No new purchases means no new buy-side Periodic Transaction Reports, the highest-value input in the dataset. The decay is gradual rather than immediate, because grandfathered positions keep producing sales for years.

Sell-side data improves. Under current law a sale surfaces up to 45 days after execution. Under H.R. 7008, sales are pre-announced 7 to 14 days in advance, moving disclosure from roughly six weeks late to two weeks early.

Substitution is the longer-term risk. Activity migrates into exempt vehicles: diversified funds, home-state concentrated funds, independent trusts, spouse occupational accounts. Those disclose less, or disclose nothing usable.

Subversive’s NANC held roughly $282 million as of July 22 at a 0.74% expense ratio. Its Republican counterpart, ticker GOP since March 2025, runs $58 to $75 million depending on the data provider. NANC has returned about 93% since its February 7, 2023 inception against roughly 79% for the S&P 500, a gap driven by mega-cap tech concentration rather than stock selection. ETF.com analysts have noted that a trading ban leaves no disclosures for the strategy to track.

Bryan Steil returned 62.5% last year

Steil, who chairs House Administration, sponsored this bill. Per Unusual Whales, he returned 62.5% in 2025, placing him among the top performers in Congress. Only 32.2% of members beat the market that year.

Public support for a congressional trading ban sits at 86%, per a University of Maryland survey of 2,625 registered voters. The bill the House passed preserves every existing position, exempts the vehicles where most member wealth already sits, and carries a voter ID mandate that stalls it in the Senate.

All trades referenced here are publicly disclosed through STOCK Act Periodic Transaction Reports and aggregated on platforms including insidersignal.ai.

The decision point is whether the Senate strips Section 3 and moves the trading provisions separately. That is the only version with a path to enactment, and the point at which buy-side disclosures begin to disappear.

Nancy Pelosi is the most-copied trader in America. Her portfolio has beaten the market for years. By one tracker’s math it returned 45.6% in 2024 and 54.1% in 2025, and on a single copy-trading app, 19,219 people mirror her trades. “What is Nancy Pelosi buying” is one of the most-searched questions in retail investing.

Here’s the downside: you cannot actually copy her best trades (believe me, I’ve tried). Her biggest winner is an Nvidia position her husband opened in 2024 at $113, now near $200, and you cannot go back and buy it at her price.

Congressional trades are files up to 45 days after it happens, so by the time you see the disclosure, the move has usually already run. The ETFs built to mirror Congress make the point. They have roughly matched the S&P, because their returns come from the same mega-cap tech sitting in every index fund.

Subscribe now

BUT → there’s a faster insider signal and almost nobody is tweeting about it. In 2026, copying biotech insiders who bought their own stock on the open market returned about 57%. Their purchases hit public record in two business days.

I wrote a full technical deep dive on biotech insider copying here.

It’s exactly what I’ve been telling my followers for over a year. (I even wrote a public article about it on Substack)

Every trade below comes from an SEC Form 4, the filing insiders and 10% owners must submit within two business days of a purchase. It all sits on EDGAR and on insidersignal.ai, and trading on it is legal.

Share

Biotech insiders see the data first, and 2026’s M&A boom made the sector a target

A drug company is worth what its data says it is worth, and generally the data goes to the insiders first. Interim trial numbers. The tone of an FDA meeting. Whether a manufacturing batch held.

Decades of research show insider buying predicts returns most reliably in research-heavy firms, and biotech is the purest example there is.

The moment is 2026. After years of dead money, big pharma is staring down a wall of patent expirations and paying up to fill it. Biopharma M&A hit roughly $106 billion across 201 deals by late June, on pace for the best year since before the pandemic, with the average deal jumping to $527 million from $365 million in 2025.

Roche paid $7.1 billion for RVT-3101. Merck paid $10.8 billion for tulisokibart. The FDA turned friendlier, and the IPO window reopened. When a sector is this catalyst-rich, the people who can read the catalysts have the most to act on.

A $3M founder buy popped 115%, and a top fund poured in $100M

The names make the pattern concrete.

In June, CervoMed director and 10% owner Joshua Boger, the man who founded Vertex Pharmaceuticals, bought 955,414 shares of the small neurology company for about $3 million. When the Form 4 posted, CRVO jumped as much as 115% intraday. A founder who has read more clinical data than almost anyone alive put cash in at the market price, and traders read it exactly as intended.

Artiva Biotherapeutics is the slow-burn version. RA Capital, one of the sharpest specialist funds in the business, anchored a $300 million raise at $11.52 in May with close to $100 million, then kept buying the pullback in the open market down in the $7s and $8s. It now holds around 16.8 million shares. GC Cell, the Korean company that co-invented Artiva’s cell therapy, bought right alongside it. The stock started 2026 in the low single digits and trades near $10 now, after its off-the-shelf cell therapy posted strong data in autoimmune disease.

The list runs way deeper than this table. At Nuvation Bio, a veteran pharma venture investor on the board kept adding as the cancer program gained traction. And in the 13F filings, the biggest specialists, Baker Brothers, OrbiMed, Perceptive, and RA Capital, rotated toward risk. They crowded into clinical winners like ABVX, where institutional ownership jumped about 68% in a single quarter.

View all these trades at insidersignal.ai

The same funds are now selling into strength, a sign the easy money is made

The clearest proof the early buying paid off is what those same funds are doing now.

After the run, specialists are selling into strength. OrbiMed sold about $39 million of Enliven after a 130% rally. A Palvella insider trimmed following a 620% move. Veracyte and Monte Rosa insiders took profits after gains of 120% and 400%. BVF cut Kymera into strong data.

This all looks like a harvest after a year of good positioning. That cuts two ways, and the second way matters for you. When the best-informed holders shift from buying to trimming, the easy money is usually already made.

Watch small-cap code-P buys, especially clusters in the same week

Do your own research, I’m not a financial advisor.

Here’s my secret to catching the next biotech breakout:

Strip it down. In small and mid-cap biotech, the buys worth your attention are code-P open-market purchases, and the ones worth the most are clusters of insiders buying inside the same week.

The catch is speed, the same thing that sinks the Pelosi trade. By the time these purchases surface in a weekend roundup, the first move is often gone.

That’s why I made Insider Signal to track the filings the second they go public, so you see the cluster forming instead of reading about it a week later.

Methodology and disclosure: Every transaction referenced comes from SEC Form 4 filings, which corporate insiders, directors, and 10% owners must file within two business days of a trade. Congressional trades are disclosed under the STOCK Act, with up to a 45-day lag. All filings are public through EDGAR and aggregated on insidersignal.ai, and trading on them is legal. Return figures are 2026 year-to-date: the biotech insider-copying result is from InsiderSignal’s first-half 2026 backtest using realistic fills; the biotech sector (XBI) and S&P 500 figures are year-to-date total returns through mid-July 2026. None of this is investment advice. Clinical-stage biotech is high risk, and every investment can lose money.

On March 19, 2026, Rich McCormick sold his stake in L3Harris Technologies, a Pentagon contractor that builds tactical radios, missiles, and avionics. Eighty-five days later, on June 12, he bought L3Harris again. Both trades ran through the same retirement account. Both were small, valued between $1,001 and $15,000. The entire time, McCormick sat on the House Armed Services Committee and its Tactical Air and Land Forces subcommittee, the panel that oversees Army and Air Force procurement.

That round trip is a clean way into a portfolio most people have never examined. McCormick has filed 52 stock trades since taking office, 35 purchases and 17 sales, according to data aggregated on insidersignal.ai. The tickers read like a map of his committee assignments. And two dozen of them did not reach the public until nearly two and a half years after the law required him to disclose them.

Every trade here is public. Members of Congress must file periodic transaction reports under the STOCK Act, the 2012 law that requires disclosure of individual securities trades within 45 days. The filings land in the House clerk’s database and get aggregated on platforms like insidersignal.ai. Tracking them is not exotic. The information is public. The question is what it shows.

Subscribe now

Three committees that reach into defense, health, and tech

McCormick is a physician and a retired Marine and Navy officer. He flew the CH-53E Super Stallion helicopter, later served as a Navy Medical Corps doctor, and worked as an emergency room physician at Gwinnett Medical Center and Northside Hospital before winning his House seat in 2022. He represents Georgia’s 7th District, the suburbs north of Atlanta. He is running for reelection, with a Republican primary set for May 19, 2026.

In the current Congress he holds three committee seats. Each one touches a slice of the market he trades.

Armed Services oversees the Department of Defense and the roughly $900 billion National Defense Authorization Act. His subcommittees are Tactical Air and Land Forces, which covers Army and Air Force weapons procurement, and Cyber, Information Technologies, and Innovation. That jurisdiction sits directly on top of contractors like L3Harris. McCormick worked the FY2027 defense bill in committee this spring, one of nearly 1,000 amendments the panel reviewed over 14 hours.

Oversight and Government Reform is the House’s main investigative committee. Its reach runs across federal agencies and technology policy.

Science, Space, and Technology covers federal research, NASA, and much of the tech and telecom landscape. He sits on its Investigations and Oversight and Space and Aeronautics subcommittees.

Now hold that list next to his trades. Defense: L3Harris. Healthcare: Johnson & Johnson, UnitedHealth, Abbott Laboratories. Technology: Microsoft, Cisco, Analog Devices. Financial and telecom infrastructure: Intercontinental Exchange, the owner of the New York Stock Exchange, and American Tower, an $81 billion cell-tower REIT. The portfolio and the jurisdiction line up almost stock for stock. OpenSecrets noted the overlap directly: some of his purchases were in industries that fall under one or more of his committees.

Two dozen trades that surfaced nearly two and a half years late

Share

The clearest documented problem is timing. In March 2023, McCormick bought about two dozen individual stocks. He listed them on his annual financial disclosure, which every member must file. But he did not file the separate periodic transaction reports until 2025, nearly two and a half years past the 45-day deadline, according to an OpenSecrets analysis published that September.

The 2023 haul was broad and blue chip. Apple. Cisco. Costco. Emerson Electric. Home Depot. Johnson & Johnson. L3Harris. Microsoft. NextEra Energy. PepsiCo. UnitedHealth. Walmart. Taken together, the purchases were worth between $24,024 and $360,000. Disclosure only requires broad ranges, so the exact figure stays hidden.

The stocks were not the only late filing. McCormick also disclosed buying between $100,000 and $250,000 in U.S. Treasury bills roughly two months after that deadline. Neither delay carries a criminal penalty. The standard consequence for a STOCK Act violation is a $200 fee. That is part of why late filings keep happening across both parties.

In 2026, the buying and selling ran both directions

Then McCormick went quiet. Through all of 2025 he did not buy or sell a single individual stock, based on his disclosures. In 2026 the activity restarted, and it looked different.

The trades came in clusters, and they moved both ways at once. In the filing covering March 19, he bought Abbott, Automatic Data Processing, BlackRock, Home Depot, Microsoft, UnitedHealth, and the Vanguard Mega Cap Growth ETF. In the same report, he sold Analog Devices, Johnson & Johnson, L3Harris, and NextEra. On April 17 he sold American Tower. On June 12 he bought Abbott twice, plus Intercontinental Exchange and L3Harris, and sold Analog Devices, BlackRock, Cisco, UnitedHealth, two broad-market ETFs, and the same Vanguard growth fund he had bought in March.

Track the individual names across those filings and a pattern appears. UnitedHealth: bought in March, trimmed in June. BlackRock: bought in March, trimmed in June. The Vanguard growth ETF: bought in March, sold in June. L3Harris: sold in March, bought in June. Every trade sat in one of two accounts, the Growth Partners IRA and the Growth Partners Roth IRA. Every trade fell in the same $1,001 to $15,000 band.

Why this looks like a managed account, not a hot tip

Here is the part that cuts against the easy narrative. This does not look like insider trading.

The fingerprints point somewhere duller. Buying and selling the same stock within weeks. Trading a name in the traditional IRA while trading it in the Roth. Moving broad index ETFs alongside single stocks. Capping every ticket at $15,000. That is the signature of a professionally managed, rules-based portfolio being rebalanced on a schedule, not a lawmaker acting on something he heard in a closed briefing. There are no options. No concentrated bets. No purchases parked right before a market-moving announcement. Whatever drives these trades, it is not conviction, and nothing in the record points to foreknowledge.

So the McCormick story is not the Pelosi story. It is quieter, and in a way more useful. A member can hold and trade the exact companies his committees regulate, churn them through IRAs for years, and draw almost no attention, precisely because the trades are small and mechanical. The signal that flies under the radar here is not a suspicious win. It is the ordinariness of the conflict.

McCormick is one of more than two dozen members flagged this year

He is not alone, and that is the point. More than two dozen federal lawmakers have been cited for STOCK Act disclosure violations in recent months. His healthcare positions are a crowd trade. McCormick was one of several members who bought Johnson & Johnson and UnitedHealth as those stocks climbed in 2026, with UnitedHealth up more than 29% on the year.

All of this is unfolding while Congress argues, again, about banning member stock trading. Bills to end the practice have drawn support in both parties. A version has moved through the House. Critics say it is stuffed with exemptions. Nothing has passed. A handful of members have stopped trading on their own, including Reps. George Whitesides, Greg Landsman, and Julie Johnson and Sen. Peter Welch. Until a ban becomes law, the disclosures are the only window the public gets, and the window keeps opening late.

The number that isn’t public yet

McCormick’s filings answer some questions and leave others open. The disclosed ranges hide the real dollar figures. The reports do not say whether an outside advisor directs the accounts, which would settle the intent question. And the raw periodic transaction reports behind the 35-buy, 17-sell tally have not been reconciled trade by trade, so the exact per-company counts are still soft.

What is not soft: a sitting congressman on three committees holds and trades the defense, healthcare, and technology stocks those committees oversee, and he kept two dozen of them out of public view for nearly two and a half years. The trades themselves may be the most boring thing about them. That is exactly why they are worth watching.

We track filings like these the moment they hit the House clerk’s database. Subscribe to catch the next one before it scrolls past.

Leave a comment

All trades discussed are disclosed through periodic transaction reports required under the STOCK Act and are available in the House clerk’s database and aggregated on insidersignal.ai. Dollar amounts are reported in ranges, not exact figures. Nothing here is investment advice.

On June 29, 2026, Coastlands Capital LP bought 25,000 shares of Adial Pharmaceuticals at $2.60. The firm now holds 373,322 shares, a 7.18% increase to a position that already ranked among its largest. The trade posted to SEC EDGAR as a Form 4 on July 1 at 10:30 AM ET. Small ticket, but the signal gets louder once you know who Coastlands is.

All trades discussed here are public. SEC Form 4 filings are required within two business days of execution and are available through EDGAR and aggregated on platforms like insidersignal.ai. This is not inside information. It is public information, read closely.

Matthew Perry ran one of biotech’s sharpest value funds before starting Coastlands

Coastlands Capital LP is a San Francisco investment adviser formed in 2025. It manages roughly $785 million for pooled funds, runs lean with five employees, and keeps a low public profile outside regulatory filings, according to FinTRX and SEC adviser records. The principal owner and managing partner is Matthew D. Perry (no relation to Matthew L. Perry).

Perry is not a random micro-cap buyer. He was president of BVF Partners, the Biotechnology Value Fund, one of the more respected value-oriented biotech investors in the market. BVF made its name taking concentrated positions in small, mispriced drug developers and pressing for value. Perry left, launched Coastlands, and carried that approach with him.

Adial is not his only concentrated healthcare bet. Coastlands is also a 10% owner of Pasithea Therapeutics. The pattern repeats: take a large stake in a small, beaten-down biotech, then wait. When a person with that resume adds shares, it is worth reading the filing rather than scrolling past it.

Subscribe now

Adial did a 1-for-25 reverse split in February just to stay on Nasdaq

To understand why Adial is a $6 million company, rewind to winter. On February 5, 2026, Adial executed a 1-for-25 reverse stock split. The move cut the share count from about 27.8 million to about 1.1 million. The reason was blunt. Adial needed to regain compliance with Nasdaq’s $1.00 minimum bid price or risk delisting.

Reverse splits are a distress flare. Companies run them when a stock has fallen so far that exchange rules threaten to remove it. Adial fit the profile. Its shares had dropped roughly 74% over the prior year heading into the split, and the stock later touched an all-time low of $1.31 on May 15, 2026. This is the wreckage Perry started buying: a company the market had largely written off.

In June, Adial stopped being an alcohol-drug company and became a colitis company

Then Adial reinvented itself. On June 11, 2026, it acquired Azora Therapeutics in an all-stock deal and changed its entire reason for existing.

For years, Adial’s lead asset was AD04, a genetically targeted serotonin-3 receptor antagonist, essentially low-dose ondansetron, for Alcohol Use Disorder. The program was not vaporware. Adial completed a successful End-of-Phase-2 FDA meeting in August 2025 and was planning a Phase 3 trial. After the merger, that program is the sideshow.

The new lead is AT177, an oral, colon-targeted aryl hydrocarbon receptor (AhR) agonist for ulcerative colitis, spun out of Stanford’s SPARK program. The pitch: a drug that activates locally in the colon with minimal systemic exposure, sidestepping the safety issues of AhR drugs that circulate through the body. Ulcerative colitis is a large, chronic market where many patients never reach durable remission on current biologics, which is the gap AT177 is aimed at.

The deal handed control to the newcomers. Former Azora holders ended up with about 86.9% of Adial on a fully diluted basis. Legacy Adial holders kept about 13.1%. In practice, Adial did not acquire Azora so much as Azora reverse-merged into Adial’s Nasdaq listing.

The $32 million that came with the deal explains the $2.60 price

Alongside the acquisition, Adial lined up up to $64 million in private placement financing from specialist biotech investors. The first tranche was about $32 million, structured as pre-funded warrants to buy 11,780,948 common shares at $2.7489 each. A second $32 million tranche is tied to hitting Phase 1 initiation triggers in ulcerative colitis. The PIPE closed June 16, 2026.

That $2.7489 warrant price is the anchor. Coastlands’ June 29 open-market buy at $2.60 came in just below it. Perry did not pay a premium. He added at close to the same price the financing syndicate received, on the open market, weeks after the raise closed. On June 30, ADIL traded between $2.56 and $2.70, so the $2.60 buy was an at-market add. Not a bargain grab, not a chase.

Share

Why a 10% owner adding on the open market is a specific kind of signal

Insider buying is not one thing. A director buying 500 shares to look supportive is noise. A sophisticated 10% owner increasing an already-large position is a different message.

Three things make the Coastlands buy notable. First, Perry was already committed through the PIPE. Adding more on the open market is not establishing a position, it is reinforcing one. Second, he bought after the news, not before it. The reverse split, the Azora merger, and the financing were all public by June 29, so this reads as a bet on execution from here, not positioning ahead of a known catalyst. Third, the buyer has watched thousands of small biotech stories play out. When that person adds at roughly his own entry price, it says the thesis still looks intact to one of the most informed holders on the tape.

None of that guarantees the stock works. It is a data point about conviction, not a forecast of the outcome.

The risks are exactly what you’d expect from a $6 million biotech

The bear case writes itself. Adial has a market cap near $6 million, no revenue, and a lead drug that will not start Phase 1 until mid-2027. The stock carries a beta around 2.77, meaning it swings far harder than the market in both directions.

Dilution is the structural overhang. The company just issued pre-funded warrants covering nearly 12 million common shares, and the Azora preferred stock converts into large blocks of common after stockholder approval. Legacy holders already sit at roughly 13% of the company. More shares can hit the market as warrants exercise and preferred converts. A 10% owner buying is encouraging, but it does not repeal the math of a micro-cap that runs on repeated capital raises.

The share-count math doesn’t cleanly reconcile

One honest caveat. Coastlands’ figures across filings do not line up neatly. In a Schedule 13G reflecting holdings as of June 17, 2026, the firm reported 262,167 shares, a 9.98% stake based on 2,625,943 shares outstanding. That number excluded pre-funded warrants above a 4.99% ownership cap. The June 29 Form 4 shows 373,322 shares.

The gap between those two numbers is not explained on the face of the filings. It most likely traces to how pre-funded warrants and the 4.99% blocker get counted across different form types. The takeaway is directional, not precise. Coastlands sits around or above 10% of a tiny company, and it is still buying. Anyone modeling an exact percentage should read the underlying documents rather than trust a single headline figure.

What to watch: Phase 1 in 2027 and whether Perry keeps buying

The forward calendar is thin but specific. Adial guided to Phase 1 initiation for AT177 in mid-2027. That milestone matters twice, because it also unlocks the second $32 million financing tranche. Next earnings are scheduled for August 19, 2026. Two analysts rate the stock a Buy with price targets clustered between $6 and $8, well above the current $2.60 area, though targets on a $6 million micro-cap deserve heavy skepticism.

The clearer tell will be Coastlands itself. If Perry keeps buying, watch for an amended Schedule 13D or additional Form 4s. A 10% owner who buys once might be filling an allocation. A 10% owner who keeps adding through the summer is making a statement. The filings will show which, within two business days of each trade.

Adial spent the first half of 2026 doing everything a cornered micro-cap does. Reverse split. Reverse merger. Dilutive raise. New story. The most informed buyer on the tape looked at all of it and bought more. Whether that is smart money or a value investor catching a falling knife is the question the next twelve months will answer.

Nothing here is investment advice. All figures are drawn from public SEC filings and market data as of July 3, 2026. Do your own research.

On May 29, 2026, Paul Pelosi bought 200 Uber call options. A $50 strike, expiring March 2027. The filing reached the House Clerk on June 23, valued between $500,001 and $1 million, and it was the Pelosis’ first Uber position ever. The stock was down about 30% from its record high when he bought. InsiderSignal

Deep in the money, plus a bigger Intel bet

The contracts control 20,000 Uber shares at $50 each. Uber closed at $70.40 the day he bought, so the calls were already about $20 in the money. This is not a lottery ticket. Deep-in-the-money calls behave like leveraged stock: they move nearly dollar-for-dollar with the shares, decay slowly, and cap the loss at what you paid. Paul Pelosi has run this same play for years.

The same filing carried a bigger Intel bet: another 200 calls, same $50 strike, same March 2027 expiry, but valued between $1 million and $5 million. Both stocks have been punished this year. Pelosi is buying the bottom of both.

Why Pelosi became the most-copied trader in Congress

The mechanics matter here. The STOCK Act gives members of Congress and their spouses up to 45 days to disclose any trade over $1,000. Nancy Pelosi does not place these trades herself. Her husband Paul, a venture capitalist, runs the money, and her office has said for years that she owns no individual stocks and has no hand in the transactions.

But the returns are what made her a brand. Unusual Whales pegged her 2025 portfolio at 20.1%, beating the S&P 500’s 16.6%. The two years before that were not close: about 65% in 2023 and 70.9% in 2024. Beating the market once is luck. Doing it by that much, that often, from a seat with a view of every bill and antitrust case in the country, is why a whole industry now exists to copy her. Capitol Trades, Quiver Quantitative, Autopilot’s Pelosi Tracker, the NANC ETF.

Share

The Nvidia playbook that returned $5.8 million

For the template, look at Nvidia. On November 22, 2023, Paul Pelosi bought 50 Nvidia calls at a $120 strike, expiring December 2024, for an estimated $1.8 million. Then Nvidia split 10-for-1, and the position became 500 contracts at a $12 strike. The Pelosis exercised in December 2024 and walked away with 50,000 shares. All in, they spent roughly $2.4 million. By Benzinga’s July 2025 math, the trade was up about $5.8 million.

It is the same shape as the Uber trade: deep in the money, long dated, a megacap tech name everyone already knew. And here is the part that matters for anyone watching the filings. The Nvidia calls also showed up weeks after the purchase, yet copying them the day they went public still caught most of the run.

It does not always work, though. In July 2022 the Pelosis sold 25,000 Nvidia shares near $165. Held through the split, that stake would be worth many times more today. Even the best-tracked portfolio in Washington leaves money on the table.

Uber is 30% below its high. Wall Street says buy

By its own numbers, Uber is cheap. First-quarter 2026 revenue hit $13.2 billion. Gross bookings rose 21% in constant currency, the third quarter running above 20%. Free cash flow over the trailing year set a record at $9.8 billion, the company is in the middle of a $20 billion buyback it authorized last August, and Uber One just passed 50 million members who now account for more than half of all bookings.

The stock, though, sits near $70. That is about 30% below its October 2025 high of $100.10, close to its 52-week low. Wall Street thinks that is a mistake. The consensus rating is Strong Buy, the average target sits between $104 and $108, and the high on the Street is $150. Pelosi bought the dip the analysts say should not be there.

The bear case is real, and it has a name: Waymo. It is scaling its own robotaxi network and turning from Uber partner into Uber rival in Dallas, London, and San Francisco. To avoid getting cut out of its own market, Uber is now spending billions to build a self-driving fleet, more than $300 million each into Lucid and Nuro. That fight, not the earnings, is what is sitting on the stock.

Share Insider Signal

The catch: this may be her last big trade

One more thing makes this filing strange. Pelosi is retiring. She is not running again in 2026, and her trade disclosures stop after January 2027. She has also come out for a ban on congressional stock trading. Senator Josh Hawley’s bill, renamed the HONEST Act, cleared a Senate committee on an 8-to-7 vote on July 30, 2025, and Pelosi backed the amended version. The Uber and Intel calls may be among the last big trades that portfolio ever files.

For the record: no evidence has ever shown Pelosi trades on inside information. She says she has no involvement, and the trades are her husband’s.

The one number that isn’t written yet

None of this is advice. It is one public filing, read closely. Strip out the noise: the most-watched trader in Washington just opened her first Uber position, near a 52-week low, using the same deep-in-the-money setup that worked on Nvidia. The rest of us found out 25 days later. Uber reports earnings again around August 5. What the calls are worth by March 2027 is the one number nobody has yet.

All trades discussed are publicly disclosed through STOCK Act periodic transaction reports filed with the House Clerk and aggregated on platforms including Capitol Trades, Quiver Quantitative, and insidersignal.ai.

On August 12, 2025, Thompson Lin bought 18,600 shares of his own company at $22.76. Over the next two days, three directors followed him in. The stock, Applied Optoelectronics (NASDAQ: AAOI), had just missed earnings and was trading near $22.

It closed at $191.55 on June 15, 2026, up 756% from where the insiders bought. By the next morning it had eased to $176.81, a gain of about 690%.

That is a 756% gain. The five insiders put in about $1.27 million combined. Those same shares are worth roughly $10.9 million today.

Share

This is the kind of signal that hides in plain sight. Every trade was filed with the SEC on Form 4 within two business days. Every one posted publicly the moment it filed. Anyone watching could have bought alongside them.

Here is what they saw.

The cluster

Five buys over three days. No sells. All from the top of the company.

• Thompson Lin, President and CEO, bought 18,600 shares at $22.76 on August 12. About $423,000.

• William Yeh, director, bought 18,000 shares at $22.54 on August 13. About $406,000, an 8% increase to his stake.

• Mike Chen, director, bought 7,500 shares at $22.91 on August 13. About $172,000.

• Thompson Lin bought another 8,075 shares at $21.56 on August 14.

• Che-Wei Lin, director, bought 4,609 shares at $21.64 on August 14. About $100,000.

Clusters matter more than single buys. One insider buying can mean anything. A bonus to reinvest, a routine top-up, optics for the next proxy. Five insiders buying in three days, with the CEO going twice, means the people closest to the order book agreed on one thing. The stock was too cheap.

The math on copying them is simple. Ten thousand dollars spread across those trades near $22 would be worth nearly $80,000 today. Thompson Lin’s two buys cost roughly $597,000. That position is worth about $4.7 million now.

Subscribe now

Why AAOI was trading at $22

The stock was cheap for a reason that looked bad on the surface.

AAOI reported second-quarter results on August 7, 2025. Revenue hit $103 million, up 138% from a year earlier. But the company posted a GAAP loss of $0.16 per share, about double what analysts expected. The miss came from heavy spending on research and qualification work for its next generation of transceivers.

The market read the loss and sold. Shares had already slid about 20% over the prior month into the low $20s. Back in April 2025, during the tariff selloff that hit the entire optics sector, AAOI had touched $9.71.

So by mid-August the picture was a company growing revenue triple digits, spending hard to win new business, and trading near multi-year lows. The street saw a money-losing optics vendor. The insiders saw something else.

What the AAOI insiders could see that the market couldn’t

This is the part that made the trade.

Five months earlier, on March 13, 2025, AAOI had signed a deal with Amazon. Amazon received a warrant to buy up to 7.9 million AAOI shares at $23.70. The warrant vests as Amazon buys product, up to $4 billion worth over ten years, tied to a manufacturing and development agreement with Amazon Data Services.

Read that number again. Four billion dollars. AAOI did $455 million in revenue for all of 2025. The Amazon agreement alone was worth close to nine years of the entire company’s sales.

The insiders signed that deal. They knew the qualification timelines. They knew their 800-gigabit transceivers were in final testing with a major hyperscaler. They knew the cable business was inflecting at the same time. In August they backed that knowledge with their own cash, near the low. Every buy was legal, filed, and public the next day.

The computing angle: optical transceivers are the AI chokepoint

To understand why AAOI ran 756%, you have to understand what it makes and where it sits in the AI buildout.

An optical transceiver converts electrical signals into light and back. Inside an AI data center, these devices move data between GPUs, between servers, and between switches. Copper works for short runs. Past a few meters, at the speeds AI clusters need, you need light. Light means transceivers. Lots of them.

The ratio is the whole story. A modern AI training cluster uses roughly six high-speed optical transceivers for every GPU, according to McKinsey work cited by Tech Times. So every time Nvidia ships another rack of accelerators, transceiver demand climbs with it. Nvidia sells the shovels. Companies like AAOI sell the connectors that let the shovels work as one machine.

The speeds are climbing fast. The industry is moving from 400-gigabit parts to 800-gigabit, and now to 1.6-terabit. Each jump roughly doubles the data a single port can carry, and each jump resets the race. TrendForce projects the market for AI optical transceivers will grow from $16.5 billion in 2025 to $26 billion in 2026, up 58% in one year. Shipments of 800-gigabit-and-faster parts are expected to rise from 24 million units to nearly 63 million over the same stretch. The demand sits downstream of hyperscaler spending, and that spending is going vertical. Alphabet, Microsoft, Amazon, and Meta are on track to spend around $650 billion in capital expenditure in 2026, up from roughly $380 billion in 2025, most of it aimed at AI compute and the networks that feed it.

What makes AAOI different from a box assembler is that it builds its own lasers. The laser chip is the hard part of a transceiver and the part in shortest supply. AAOI fabricates its own in Texas and Taiwan. That matters for two reasons. Margin, and supply security at a moment when laser chips are the bottleneck for the whole industry.

How tight is that bottleneck? On March 2, 2026, Nvidia invested $2 billion each into Lumentum and Coherent, two laser suppliers, to lock up priority access to EML laser chips, the high-speed lasers used in advanced transceivers. That move pushed lead times for everyone else past 2027. When Nvidia is paying $4 billion to secure laser supply, a company that makes its own lasers and runs domestic capacity is worth a second look.

That is the reshoring angle, and it is real. AAOI is spending up to $300 million on a new 210,000-square-foot plant in Sugar Land, Texas, with a groundbreaking on February 10, 2026. In late April 2026, Governor Greg Abbott announced a $20.9 million Texas CHIPS grant to support the expansion, part of more than $279 million in total investment and 500 new jobs. The company plans to triple its laser capacity in Texas by mid-2027 and build more than half of its 1.6-terabit modules in the United States. For hyperscalers worried about tariffs and Chinese supply chains, a domestic transceiver maker is a hedge worth paying for.

The competition is fierce and mostly Chinese. Zhongji Innolight did about $3.3 billion in 2024 revenue, up 123%, and holds more than half of Nvidia’s 800-gigabit module orders. Eoptolink grew 179% to about $1.2 billion. Together they control roughly 60% of the 800-gigabit market and price 20% to 25% below Western suppliers. AAOI is the smaller player. Its pitch is vertical integration and a manufacturing base outside China. The Amazon deal says at least one hyperscaler is buying that pitch.

The run-up: from $22 to $191

The insider buys came in August. The payoff came in waves.

• December 10, 2025: AAOI announced its first volume order of 800-gigabit transceivers from a major hyperscale customer.

• February 26, 2026: the company reported record fourth-quarter revenue of $134 million, up 34% from a year earlier, and guided to more than $1 billion in 2026 revenue. The stock jumped 63% in a day.

• March 9, 2026: AAOI landed its first 1.6-terabit order, worth more than $200 million. Shares rose another 16%.

• March and April 2026: a stack of new 800-gigabit orders, $53 million, then $71 million more, pushing fresh orders past $124 million in a matter of weeks.

By June 2026, AAOI had locked in more than $324 million in combined 800-gigabit and 1.6-terabit orders. Management said demand was running ahead of supply through at least mid-2027. Growth was no longer the question. Capacity was.

The chart shows the rest. A slow grind through the back half of 2025, then a near-vertical climb starting in February 2026 as the orders hit the tape. The bulk of the 756% move happened in about four months.

The full ownership picture

AAOI is a founder-run company. Thompson Lin started it in 1997 and still serves as chairman, president, and CEO. He has historically held around 2% of the stock. The board leans on the founder’s network and the Taiwanese-American optics and chip industry.

The directors who bought in August are part of that network. Che-Wei Lin runs ASMedia and is a corporate vice president at ASUSTek, the Taiwanese computer-hardware giant. He is a chip executive, not a relative of the founder, despite the shared surname. William Yeh and Mike Chen are longtime directors. Their August buys, an 8% and a 4% increase to their stakes, were meaningful against what they already held.

Insiders own roughly 5% of AAOI in total. The rest sits with institutions and one strategic player.

The strategic player is Amazon. The warrant gives it the right to about 7.9 million shares, which would make it one of the largest holders if fully vested. That stake grows only as Amazon keeps buying product, which ties the biggest customer’s interest directly to the stock. It is a clean alignment, and a rare one for a company this size.

Institutions hold the majority, historically led by BlackRock, Vanguard, and State Street. The float is volatile and heavily shorted. Citron Research disclosed a short position in 2026. The stock carries a beta above 3, so it moves about three times as hard as the market in both directions.

The insiders are now selling

Here is the part most write-ups skip.

The same insiders who bought at $22 have been selling at much higher prices. According to Form 4 filings, the CEO sold around 306,945 shares near $95 in early 2026. The CFO, Stefan Murry, sold close to 300,000 shares in February. Through the spring, directors and officers sold tens of millions of dollars more as the stock pushed toward $200.

This does not mean the story is over. Insiders sell for many reasons, taxes and diversification chief among them, and most of these people still hold large stakes. But it does mean the same group with the best view of the business were buyers at $22 and sellers at $95 to $200. That is a signal too.

The analysts are bullish and still lag the price. Rosenblatt’s Mike Genovese carries the high target at $220. Raymond James moved to $160. B. Riley stayed neutral at $129. The consensus target sits at or below where the stock already trades, which means Wall Street sees limited upside from here on its own numbers.

The valuation is stretched. At $191, AAOI trades around 22 times sales with no positive GAAP earnings. It funds the Texas buildout by issuing stock, including a $600 million share program launched in May 2026. Every new share dilutes the gain. The other risks are concentration, with the top three customers at about 90% of revenue, Chinese competition, and the long history of optics being a brutally cyclical business. AAOI itself peaked on Microsoft demand in 2017, then fell about 90% when that single customer pulled back.

The reversal may have already started. The stock slipped almost 8% in a single session this morning. On a name with a beta above 3, a move like that is ordinary. It is also a preview of how fast a 700% stock can travel in the other direction.

The pattern, and where to find the next one

The AAOI trade was not a lucky guess. It was a readable signal. Five insiders, the people who signed a $4 billion deal with Amazon and watched the order book fill, bought their own stock near a multi-year low. They filed the paperwork. The filings posted within days. The stock then did exactly what the insiders’ own cash predicted.

The lesson is not that AAOI goes higher from here. The selling says the easy money is already gone. The lesson is that the signal was sitting there in August 2025, free, for anyone who knew where to look.

That is what we do at insidersignal.ai. We track every Form 4 the moment it files, flag the clusters, and surface the trades that matter before the rest of the market catches on. The AAOI buys were on the platform in August. The next cluster like it is filing somewhere right now.

Go find it.

All trades discussed here are public, filed with the SEC on Form 4 within two business days of execution, and available on insidersignal.ai. This article is for information only and is not investment advice.

Datadog insiders sold $261 million of stock in the last 90 days. They bought zero.

The latest installment hit EDGAR yesterday: on June 8, CEO Olivier Pomel sold 127,141 shares for $29.5 million, his seventh sale of 2026 and his largest yet. The stock is up 67% year to date. Analysts are raising targets. The people who run the company are selling into every one of them.

Every trade below is a public SEC Form 4 filing, linked to its full record on insidersignal.ai.

The exit, top to bottom

Pomel’s 2026 sales: March 2, March 16, April 9, May 11 ($25.4M), May 26 ($18.9M), June 2, June 8. Roughly $98.5 million this year. The June 2 sale covered tax withholding, and the repeating 127,141-share tranches look like a scheduled 10b5-1 plan, not a panic exit.

But the pattern is bigger than him. In 90 days, 14 distinct Datadog insiders filed 39 sales worth $261.4 million, including $34.3 million in one block from board member Dev Ittycheria, MongoDB’s CEO. Open-market buys: zero.

The same shape runs up and down the AI stack. Over the 12 months through late April, insiders at Nvidia, Palantir, and Broadcom sold a combined $4.6 billion, per Form 4 data cited by Motley Fool. Nvidia insiders bought nothing in the open market all year. CrowdStrike insiders sold $163.9 million in 90 days with zero buys, $67.6 million of it from CEO George Kurtz, who has filed 16 sales since mid-May, most recently June 8. At C3.ai, down 20% this year, insiders sold $13.5 million while exactly one bought: $77,300. Even Goldman Sachs fits: $35.6 million sold in three months by the C-suite itself, CEO David Solomon and General Counsel Kathryn Ruemmler ($13.4M) included, with no purchases.

This is market-wide. In our database, insider sell filings have outnumbered buys four to eight to one in every month of 2026.

The scoreboard

Trailing 90 days of insider open-market activity, from our filing database:

Where the buying actually went

Insiders didn’t stop buying. They stopped buying what’s popular.

COE (51Talk). CEO Jack Jiajia Huang files a Form 4 almost every week, and almost every one is a buy. His latest, accepted June 10, bundles six purchases from May 29 to June 5: 165,240 shares, $4.1 million. Since May 8 he has bought roughly $18.2 million across five weekly filings; over six months, Quiver Quantitative counts 19 purchases totaling about $48.5 million, zero sales. One CEO, $50 million of his own money, into a falling price, week after week. Zero headlines.

Robinhood (HOOD). Director Meyer Malka is $55.3 million deep in two weeks: $20.0M on May 28, $15.1M on June 3, $20.2M on June 5. We covered the first tranche when it hit [LINK TO PRIOR HOOD ARTICLE]; the campaign has nearly tripled since.

Borr Drilling and MP Materials. Borr director Tor Olav Troim wrote a single $5.0 million ticket on June 9. MP’s COO bought $543,000 the same day, his second buy in three weeks.

Our cluster screen caught more: $14.4 million of TXO from oil veteran Bob Simpson, six Horizon Technology Finance insiders buying together on June 9 and 10, and one trap: ten Quantinuum insiders bought $24.6 million on June 5, but every leg printed at exactly $60.00, an offering, not open-market conviction. That distinction is why raw feeds mislead.

Add it up: over $75 million of conviction buying in two weeks, none of it in an AI darling, almost none of it covered anywhere.

How to read it

Insiders sell for many reasons but buy for only one. Research on insider trades (Lakonishok and Lee, 2001) finds the buy signal is strongest in small, uncovered companies, where this month’s buying showed up, and weakest in large caps, where the selling is concentrated. So: ignore single sales, weight repeated selling with zero offsetting buys, and treat multi-insider or multi-week buying campaigns as the highest-value signal on the tape. June gives you a live example of all three.

What to watch

Whether Huang’s COE campaign extends into a seventh month. Whether Malka files a fourth Robinhood tranche. Whether Datadog’s 90-day sell total crosses $300 million at the next vesting cycle. Each answer arrives as a Form 4, two business days after the trade.

The insiders closest to the AI rally are treating it as liquidity. The insiders writing buy tickets are doing it in oil rigs, rare earths, brokerage apps, and a Chinese education stock nobody screens. One side is wrong.

The Insider Signal tracks SEC Form 4 filings weekly and surfaces the patterns headlines miss. Every trade above links to its full filing record, free to view, at insidersignal.ai. Nothing here is investment advice.

On May 28, 2026, an investment fund tied to Robinhood director Meyer “Micky” Malka bought 249,000 shares of HOOD on the open market.

• Average price: $80.39.

• Total: $20 million.

It was the first open-market purchase by Malka’s funds after more than 18 months of selling. It landed one trading day before the stock closed up 11%.

Twenty million dollars is small against Malka’s total exposure and Robinhood’s roughly $80 billion market cap. The reversal is the story. The same funds that returned $210 million by selling HOOD in 2024 and 2025 just bought shares back. Discretionary. Open market. No prearranged plan.

All trades here are public, disclosed in SEC Form 4 filings, required within two business days and aggregated on insidersignal.ai. Nothing below alleges illegal conduct. It documents what insiders filed and what the timing shows.

The trade that broke the pattern

The Form 4 was filed June 1 and signed by Malka. The transaction code is “P.” That letter matters. Code P means an open-market purchase made at the insider’s discretion. Not an option exercise, not a vesting event, not a sale running on a 10b5-1 plan. Someone decided to buy, at that price, in that window.

The buyer is Bullfrog Capital, a Ribbit Capital vehicle. The 249,000 shares filled between $80.05 and $80.68. The fund now holds 3,493,427 shares. Malka discloses the position as fund capital and disclaims beneficial ownership except for his financial interest. This is Ribbit money, not a personal checkbook. Keep that straight before calling it a founder betting his savings.

The weight comes from the context. For 18 months the funds sold. On October 8, 2024, they sold 3,225,806 shares at $25.07, about $80 million, under a 10b5-1 plan. A separate plan sold 3,219,314 shares between $40.11 and $42.57, about $130 million. That alone is $210 million in selling. For a venture fund, distributing shares back to investors after an IPO is mechanical, not directional. A buy is different. A buy says the stock is worth more than the cash.

Who Micky Malka is

Malka grew up in Caracas during hyperinflation. He co-founded a broker-dealer as a teenager, then built Patagon.com, sold to Banco Santander in March 2000 for more than $700 million. In 2012 he founded Ribbit Capital, which invests only in fintech. Robinhood is its signature bet. Ribbit has put in more than $500 million since 2014 and led the $3.4 billion emergency round in February 2021 that kept Robinhood solvent through the GameStop squeeze. Malka joined the board on March 23, 2022.

This is not a generalist director nibbling at the stock. It is the head of Robinhood’s most committed backer, who carried the company through its near-death moment, adding at a price he picked.

The hedge six days earlier

Here is the wrinkle. On May 22, the same fund filed a separate Form 4 reporting an options hedge: a three-legged FLEX put position expiring December 31, 2027, covering 30,000 shares per leg. It sold puts at the $100 strike, bought puts at $75, and sold puts at $45.

The key number is size. Thirty thousand shares per leg against 3.49 million held. Under 1% of the stake. This does not contradict the buy. It reads as routine risk management on a concentrated position. The fund capped a sliver of its downside, then added a much larger position on top. “Unhedged conviction trade” is the wrong phrase. “Managed the risk, then bought more” is the right one.

What Malka bought into

He bought into a controversial quarter. Robinhood reported Q1 2026 on April 28. Revenue hit $1.067 billion, up 15%. Net income was $346 million, up 3%. Then the miss: adjusted EPS of $0.38 against a $0.39 consensus, the first miss after four straight beats. The stock fell about 6% after hours.

The mix had shifted hard. Crypto revenue fell 47% to $134 million. That was the scare. But total transaction revenue still rose to $623 million because event-contract revenue jumped 320% to $147 million, on a record 8.8 billion contracts traded. Gold subscribers reached 4.3 million, up 36%. Funded customers hit 27.4 million, up 6%.

The split in reading is the whole debate. If Robinhood is becoming a platform that no longer lives on crypto, this quarter proved it: crypto cratered and revenue still grew 15%. If the trading mix still drives the stock, you saw the crypto number and sold. The market did the second thing on the day. Malka did the first thing four weeks later.

The valuation problem

The stock is expensive. At roughly $90, the market cap sits near $80 billion and the trailing P/E in the mid-40s. J.P. Morgan’s Ken Worthington said it plainly when he set a $90 target: about 45 times his 2026 estimate. Schwab trades near 19 times forward earnings. Interactive Brokers near 32. Consensus growth is about 18%. A 45-times multiple on 18% growth needs the platform story to keep delivering. Any stumble compresses it toward Schwab, and that is a long way down.

Wall Street stays bullish, with a Strong Buy rating across 35 analysts and a median target near $97.50. The timing of Malka’s entry is the tell. He bought at $80.39 with the stock well off its $153.86 high and sentiment soured by the crypto print. He bought weakness. He did not chase. The next session, May 29, HOOD closed at $94.30, up 11%, on volume 122% above average. The buy already looks well-timed. The easy part of the move is gone.

What the data says about buys like this

Insider purchases get attention and insider sales mostly do not, for a reason. Lakonishok and Lee, in the Review of Financial Studies (2001), found firms with heavy insider buying beat firms with heavy selling by 4.8% over the next year, size-adjusted. The finding: purchases carry information, sales do not. Insiders sell for a hundred reasons. They buy for one.

Two caveats. The effect was strongest in small companies, and Robinhood is an $80 billion large-cap, where the signal runs weakest. And the study is an aggregate, not a single-stock predictor. So treat the Malka buy as what it is: a large, discretionary purchase by one of the best-informed directors on the board, reversing his own funds’ selling. The most informative category of insider trade. One data point. It raises the odds. It does not settle them.

Context cuts the other way too. Co-founder Baiju Bhatt is still selling, most recently 67,422 shares on May 20 for about $5.08 million, on a 10b5-1 plan. Scheduled, not directional. His trust still holds about 47 million shares. So the picture splits: one insider trims mechanically, another buys at his discretion near the lows. Malka’s is the only directional bet in the window, and it points up.

What to watch

The headline is not “director buys, stock goes up.” A director whose funds returned $210 million of Robinhood stock over 18 months stopped selling and bought $20 million back, near the lows, into a quarter everyone hated, six days after capping a sliver of his downside. That combination is rare, and it is the cleanest directional signal HOOD insiders have sent since the IPO.

Three things would move the read. A second insider buying would turn one signal into a cluster, which carries more weight. A follow-on purchase by Malka’s funds would confirm conviction. A return to plan-based selling, or a bigger hedge, would suggest the buy was a gesture.

For now the filing says one thing clearly. The person who knows Robinhood as well as anyone outside the C-suite looked at $80 and decided it was cheap. The market, one day later, started to agree.

All transactions described are disclosed in public SEC Form 4 filings available through EDGAR and aggregated on insidersignal.ai. This article documents publicly reported insider activity and is not investment advice. It does not allege that any party traded on material non-public information.

Leave a comment

On May 20, 2026, SpaceX filed paperwork to go public. The news led every financial outlet in the world. Estimated valuation: $1.5 to $2 trillion. Largest IPO ever.

It was not actually news. The signal had been sitting in public filings for five months. Here is what happened.

December 13, 2025: A Memo From the CFO

SpaceX is private, but it has thousands of shareholders. Employees, early investors, venture funds, late-stage buyers. Every six months or so, the company runs a tender offer where some of them can sell shares back to the company or to outside buyers at a fixed price.

To do that, SpaceX has to set a price. The CFO, Bret Johnsen, sends a memo to shareholders explaining the new mark. He has done this every cycle since 2011.

On December 13, 2025, Johnsen’s memo set the share price at $421. That valued SpaceX at roughly $800 billion. The memo also previewed an IPO in 2026.

Fortune reported the memo the same day. SpaceX did not deny it.

Subscribe now

The Price Doubled in Five Months

Here is the valuation ladder leading up to the IPO:

• June 2024: $210 billion

• December 2024: $350 billion

• July 2025: $400 billion

• December 2025: $800 billion

• February 2026: $1.25 trillion (after merging with Musk’s AI company)

• May 2026: $1.5 to $2 trillion target

Look at the jump from July to December. The valuation doubled in five months after barely moving for a year. That kind of move does not happen by accident in a private company. SpaceX was marking shares up to a level that made sense for a near-term public listing.

If you saw the July to December jump and understood what it meant, you had until May 20 to position.

Share Insider Signal

Google Owns $100 Billion of SpaceX and Almost Never Talks About It

In January 2015, Google invested $900 million in SpaceX at a $10 to $12 billion valuation. That bought about 7.5% of the company.

Google has never voluntarily disclosed the size of that stake in its quarterly or annual reports. Not in eleven years.

In April 2026, six weeks before the IPO filing, Bloomberg reported that an Alaska regulatory filing showed Google’s stake at 6.11% as of December 31, 2025. At a $2 trillion IPO valuation, that stake is worth roughly $100 billion. That is about 4% of Alphabet’s entire market cap, sitting on the balance sheet, with almost no analyst coverage modeling it.

Anyone who read the Bloomberg report could have bought Google stock as cheap, liquid, legal SpaceX exposure six weeks before the filing repriced the asset.

A Mutual Fund Jumped 19% in One Day

The cleanest public signal of the December repricing came from a mutual fund.

Baron Partners Fund is a $10 billion fund run by Ron Baron. He bought SpaceX in 2017 at a $20 billion valuation. He has never sold. As of March 2026, SpaceX is roughly 33% of the fund, with Tesla at another 20%.

On December 4, 2025, Baron Partners Fund returned 19% in a single day. A diversified mutual fund does not move 19% in a day unless one of its biggest holdings just got marked up. SpaceX had just been internally repriced.

Baron Partners Fund costs $2,000 to access at any brokerage. The fund’s daily share price is the cheapest private-market price discovery available to normal investors. The December 4 spike told you what the December 13 memo would say, nine days before Fortune reported it.

Show Image

Where to Actually Buy SpaceX Right Now

If you wanted SpaceX exposure between December 2025 and May 2026, the realistic options were:

Google stock (GOOGL). Roughly 5% of Alphabet’s value is a SpaceX position. The cheapest, most liquid way to get exposure. You also own Search, YouTube, Cloud, and Waymo.

Baron Partners Fund (BPTRX). 33% SpaceX, 20% Tesla, $2,000 minimum, available at any brokerage. The fund tracks SpaceX’s latest internal valuation, which is usually well below the IPO target.

XOVR ETF. A regular ETF that holds roughly 23% SpaceX through a clean structure. Trades like any stock. Bought at scale before the filing.

ARK Venture Fund (ARKVX). 17% SpaceX, $500 minimum, but with quarterly limits on selling.

Avoid Destiny Tech100 (DXYZ). Currently trades at roughly three times the value of its actual holdings. The premium will collapse after the IPO when SpaceX becomes buyable directly.

Avoid private SPV platforms unless you understand exactly what legal structure is holding the shares (see the next section).

Linqto Was the Warning Sign

In March 2025, a pre-IPO investment platform called Linqto shut down. In July 2025, it filed for bankruptcy. About 13,000 customers thought they owned SpaceX shares. They did not.

Linqto held the shares through legal entities that were never properly set up. Creditor counsel called it “a fraud case” with markups of 30% to 150% over what Linqto actually paid. The bankruptcy plan was confirmed in February 2026.

The lesson: not every platform selling pre-IPO SpaceX shares actually owns them in a way that survives a bankruptcy. The structure matters more than the marketing. Customers paid extra to platforms that turned out not to legally hold the underlying asset.

What Sophisticated Holders Did Not Do

Musk has not sold significant shares in any tender. Founders Fund, the first institutional investor, has not trimmed. Gigafund, the vehicle Luke Nosek spun out in 2017 specifically to keep funding SpaceX, has not exited. Valor Equity, with 7.3% of the company, has not exited.

The sellers in tenders have been mid-tenured employees and a few early angels. The largest, best-informed holders have held.

That is a positioning signal. The people closest to the company are not taking liquidity even at $800 billion.

What to Watch Between Now and December

The IPO is targeted for the week of June 8, 2026, with first trading around June 11 or 12. A few things to track:

• The first five trading days will likely see forced buying from index funds. FTSE Russell announced fast-track inclusion rules that put SpaceX in major indexes after just five sessions.

• Q3 2026 earnings will be the first quarter where Starlink’s financials are officially broken out. Current estimate: $11.4 billion in revenue, $7.2 billion in cash profit, 10 million subscribers.

• December 8, 2026 is when insiders can legally start selling. This is the 180-day lockup expiry. Supply hitting the market typically pressures prices.

• Any reactivation of the Musk-Trump feud. In June 2025, the Pentagon was reportedly directed to review every SpaceX contract within days of a Twitter spat. Government revenue is roughly 80% of SpaceX’s non-Starlink launch business.

The Bigger Point

The information needed to position for the SpaceX IPO was sitting in public documents the entire time. A CFO memo reported by Fortune. An Alaska regulatory filing reported by Bloomberg. A mutual fund’s daily share price. An ETF’s quarterly disclosures.

The investors who saw it bought Google, Baron Partners, XOVR, or ARK Venture between December 2025 and April 2026 at prices that already look cheap. The investors who waited for the May 20 filing are paying 30% to 100% more for the same exposure.

The SpaceX IPO will be written up as a Musk story, a Mars story, a Starlink story. The real story is that none of the signals were hidden. They were just spread across enough different documents that no single news source put them together.

That is what insider signal investing actually is. Not Form 4 filings alone. Reading the disclosures that exist outside the usual places, and connecting them before anyone else does.

All filings discussed are publicly available through the SEC, the Alaska Division of Banking and Securities, mutual fund disclosures, and reported financial press. Aggregated insider trading data at insidersignal.ai.

On May 11, 2026, Robert Duggan bought 660,233 shares of Pulse Biosciences at $19.69 for $12,999,987. By May 14, the stock hit a 52-week intraday high of $27.98. The position currently sits on roughly $4.2 million in unrealized gains. (Insider Signal trade page)

Across his full 49.4 million share stake, the same eight-day move lifted his paper position by more than $300 million.

Duggan has been chairman of Pulse Biosciences since November 2017. He now holds 49,450,345 shares, roughly 71.6% of the company. He has been adding for nine years.

What makes this trade worth a deep look is the track record behind the buyer.

Share

The Pharmacyclics Playbook

Duggan bought into Pharmacyclics in 2008 when shares traded between $1 and $3. He took over as chairman and CEO. He drew no salary. He drove the development of ibrutinib, sold under the brand name Imbruvica.

In May 2015, AbbVie acquired Pharmacyclics for approximately $21 billion in cash and stock at $261.25 per share. Duggan personally walked away with more than $3.5 billion.

He ran a similar playbook at Summit Therapeutics. In September 2024, Summit’s HARMONi-2 trial showed its bispecific antibody ivonescimab cut the risk of disease progression or death by 49% compared to Merck’s Keytruda in first-line non-small cell lung cancer. Median progression-free survival came in at 11.14 months versus 5.82 months. Summit stock rose 583.7% in calendar year 2024. (GuruFocus)

When this man writes a $13 million check at 81 years old, the signal carries weight.

The Cluster

Duggan did not buy alone. CEO Paul LaViolette purchased 15,000 shares the same day at the same $19.69 price for $295,350. The buy raised his direct holdings from 846 shares to 15,846, an increase of 1,773%. (Investing.com)

LaViolette is a former Boston Scientific Chief Operating Officer. He spent 15 years at the company in cardiovascular leadership roles, helping build the franchise that today sells Farapulse, the dominant pulsed field ablation product in U.S. atrial fibrillation. He took the Pulse Biosciences CEO seat on January 10, 2025. (MassDevice)

So the biotech investor with one of the best track records of his generation and the operator who helped build Boston Scientific’s competing PFA franchise both wrote checks on the same day at the same price.

Why the Stock Was Cheap

Pulse Biosciences reported Q1 2026 results on May 4. Revenue came in at $401,000. GAAP net loss landed at $18.6 million. Cash and equivalents stood at $68.3 million against quarterly burn of $14.6 million. The stock dropped 10.47% pre-market and 6.35% in aftermarket trading. (Investing.com Q1 transcript)

That drawdown set the entry. The board authorized insider purchases on May 6, two days after the print. Duggan and LaViolette executed on May 11 during an open trading window.

One detail matters. The shares were acquired through the company’s at-the-market equity offering program with TD Securities, established under the February 19, 2026 Equity Distribution Agreement. Mechanically, the $13.3 million went to the company’s balance sheet. Functionally, it is fresh personal capital reported on Form 4 as an acquisition at the prevailing bid. (Stocktitan Form 4/A)

The Catalyst Stack

The buy followed months of clean clinical data on the nPulse Cardiac Catheter System.

February 5, 2026: Late-breaking data on 150 European patients at the AF Symposium showed 100% acute pulmonary vein isolation, 100% procedural success at six months, and 96% success at 12 months.

February 20, 2026: Oppenheimer raised its PLSE price target from $22 to $30. Mizuho maintained Buy.

April 7, 2026: The NANOPULSE-AF pivotal IDE trial opened enrollment at St. Bernards Medical Center in Jonesboro, Arkansas. Approximately 215 patients across up to 30 U.S. and European sites. Enrollment completion expected early Q4 2026. (Stocktitan 8-K)

April 25, 2026: Updated data at Heart Rhythm 2026 showed 90% Kaplan-Meier freedom from atrial fibrillation, flutter, or tachycardia at one year in the five-second ablation cohort. (Pulse Biosciences press release)

May 13, 2026: Management presented at the BofA Securities 2026 Healthcare Conference, two days after the cluster buy.

The Competitive Setup

Pulse Biosciences is not yet competitive with Boston Scientific’s Farapulse, which crossed $1 billion in revenue during its first full year of U.S. commercial launch in 2024. Medtronic’s PulseSelect captured additional share. The U.S. PFA market exceeded $500 million in 2024 and continues to grow. (Fierce Biotech)

The bear case has assumed Pulse will struggle to differentiate. The 90% one-year durability data from Heart Rhythm 2026 challenges that view. So does the fact that the former operator of Boston Scientific’s electrophysiology business now runs the company and just bought stock in size.

NANOPULSE-AF topline data is expected in 2028. Between now and then, the story depends on enrollment cadence, interim safety, and capital. With Duggan holding 71.6% and a demonstrated willingness to backstop financings, the 2024 rights offering raised $60 million, the 2026 ATM has funded the runway. Financing risk is structurally lower than the burn rate alone suggests.

What This Means for the Signal

Our backtests put the C-Suite plus Healthcare combination at a Sharpe ratio of 6.22. The healthcare sector punishes generalists and rewards insiders with visibility into binary clinical events. Duggan and LaViolette have that visibility.

Two earlier Insider Signal calls fit the same pattern. OrbiMed’s $10.9 million Shattuck Labs purchase delivered 131% returns in 15 days. Eli Lilly’s insider cluster bought the orforglipron crash bottom and saw 50% to 63% returns in three months.

The Duggan trade is up 32.5% in eight trading days. The short-term pattern fits.

The longer pattern, the one that produced Pharmacyclics and Summit Therapeutics, takes years to play out. Eight days does not tell the story. The real test arrives in 2028 when NANOPULSE-AF reads out.

View the full trade on Insider Signal

All trades discussed are publicly disclosed via SEC Form 4 filings, required within two business days of execution. This information is available through the SEC’s EDGAR database and aggregated on platforms like insidersignal.ai.

April 28, 2026. Pershing Square Inc. prices its IPO. Bill Ackman buys 800,000 shares at $23.77. Cost: $19,015,801. The Form 4 hits EDGAR May 4. The stock grinds higher. May 11: PS rips 24.46% to close at $52.09. May 12: $50.83. Today, May 14, the stock has pulled back. The 800,000-share lot peaked up roughly $22.7 million. The round trip is the point.

Founders sell into IPOs. They almost never buy. When one does, and the disclosure lands on the same day the stock posts its biggest single-day rally since listing, every part of the playbook is on display.

The Trade: 800,000 Shares at $23.77 on IPO Pricing Day

Per SEC Form 4 filed May 4, 2026: William A. Ackman (Director, CEO, Chairman, 10% owner) purchased 800,000 shares of Pershing Square Inc. on April 28, 2026 at $23.77. Total: $19,015,801.

After the trade, Ackman’s economic control sits at roughly 94.6 million shares. Direct: 1,500,000 common. Indirect: 92,878,204 fully vested M Units redeemable into 86,493,537 PS common shares one-for-one. The same filing recorded a 300,000-share gift and a 6,874,237-share contribution back to the issuer.

All trades cited are public via SEC Form 4, aggregated on insidersignal.ai. Note: some aggregators display PS as NASDAQ-listed. Per CNBC, Bloomberg, and the NYSE, PS trades on the NYSE.

Share

Five Criteria That Separate Real Insider Buys From Noise

Most insider buys are noise. Five criteria separate signal from noise. This trade hit all five.

Founder, not director. Director purchases of a few thousand shares are routine. Founder purchases at meaningful scale are rare. Ackman is the operator who built the company.

At the deal price, not the dip. Ackman did not buy after PSUS broke 18% below issue. He bought on pricing day, at the same price institutions paid. That is conviction underwriting.

Scale. $19 million is meaningful even at an $8 billion net worth. 5,000-share board buys can be ignored. $19 million by a CEO cannot.

Concentration. Ackman runs a 10-name portfolio. He is paid to know what conviction looks like. Writing a $19 million check to his own ticker is a public statement that the market is mispricing the asset.

Disclosure timing as catalyst. The April 28 trade hit EDGAR May 4, four business days later. From May 4 forward, the buy was public. The stock ground higher across the week and ripped 24.46% on May 11. The Form 4 was the catalyst.

A trade that hits all five is rare. When you see one, the question stops being “is this a signal” and becomes “what is the exit.”

PS Is the Manager. PSUS Is the Fund. PSH Trades in Europe.

PS is not a fund. It is the asset manager.

Pershing Square Inc. (NYSE: PS) collects management fees on Pershing Square’s funds. Pershing Square USA (NYSE: PSUS) is the closed-end fund holding the concentrated portfolio. Pershing Square Holdings (LSE/AEX: PSH) is the long-running European-listed fund.

The combined PS plus PSUS IPO raised $5 billion. PSUS dropped 18% on day one. PS held up.

2025 financials: $30.7B AUM, $20.7B fee-paying, 96% permanent capital. Revenue $762.5M. GAAP net income $249.8M. Implied valuation at $28: $11.2B. Cumulative net returns since 2004: 2,600%+ vs 836% for the S&P 500.

Show Image

Pershing Square Holdings logo. Image: Wikimedia Commons.

From IPO Pricing to a $52.09 Peak in 13 Days

April 29, 2026: PS opens around $24.

May 1, 2026: Ackman and CIO Ryan Israel host an X Spaces walking through the structure.

May 4, 2026: Form 4 disclosing the $19M IPO-day buy hits EDGAR. The market now knows.

May 7, 2026: PSH annual general meeting confirms fee cut and Q2 dividend.

May 11, 2026: PS rips 24.46% to $52.09. Biggest single-day move since listing.

May 12, 2026: $50.83 close. The lot is up 114%, roughly $22.7M unrealized.

May 13 to 14, 2026: Pullback. The catalyst was identified. Buyers had bought. Mean reversion.

The textbook arc held. Insider files. Stock grinds higher. Momentum crowd piles in. Stock peaks within a week or two. Pullback follows.

Three Ways to Exit a Trade Like This

Most retail investors who chase insider-buy headlines treat the entry as the hard part. It is not. The exit is.

Framework 1: Hold Until the Insider Sells.

Ackman’s horizon for PS is permanent. He is building a Berkshire-style vehicle. He will not sell at $50, $80, or $200. Copy his trade as a long-term hold and you stay through every drawdown. The risk: your hand is weaker than his. Most copy-the-insider trades fail because the follower folds at the first shake.

Framework 2: Scale Out During the One-Week Disclosure Window.

If the Form 4 itself is the catalyst, the trade is finite. Form 4 hit May 4. Stock peaked May 11. Window: one week. Enter on the disclosure. Scale out one-third on any single-session pop above 10%. Scale out another third within 48 hours. Trail the rest at the 20-day moving average. Mechanical exits beat predictions.

Framework 3: Trim at Price-to-Revenue Anchors.

PS at $24 traded near 7x trailing 2025 revenue. PS at $52 traded near 14x. The same business does not double in value in two weeks because of an X Spaces and a Form 4. Set scaled-out exits at price-to-revenue thresholds. Trim 25% at 10x. Another 25% at 12x. Another 25% at 14x. Anchor the exit to the business, not to price action.

The common thread: decide before you buy. Catalysts produce mean reversion. Plan the exit at entry, capture the catalyst. Plan the exit during the rally, chase tops.

Three Lessons for Reading the Next Form 4

Three takeaways.

First, this is what a clean signal looks like. Founder. Deal price. Meaningful scale. Concentrated portfolio. Public catalyst around the disclosure. Save it as a reference example.

Second, the catalyst structure repeats. Trade. Disclosure four to five business days later. One-to-two-week momentum window. Mean reversion. The pattern shows up across hundreds of insider-buy trades on insidersignal.ai. Recognizing it beats predicting any single trade.

Third, exit planning is the entire game. Ackman’s lot peaked at $22.7 million unrealized on May 11. Today the gain is smaller. He has a permanent horizon and does not care. A copy-trader who chased the May 11 rally without an exit plan does.

Insider signals are not predictions. They are documented decisions by people who know more than any outside observer can. The reader’s job is to recognize the signal, understand the catalyst, and have the exit planned before it plays out.

The next Form 4 from any PS insider is the read that matters.

All trades referenced are publicly disclosed via SEC Form 4 filings. Data aggregated at insidersignal.ai. Informational only. Not investment advice.

Sources

• SEC Form 4, Ackman, Pershing Square Inc., April 28, 2026

• StockTitan Form 4 detail

• MarketScreener insider buy summary

• CNBC: Pershing Square IPO debut

• Bloomberg: Ackman’s $5B Pershing IPO pricing

• Timothy Sykes: PS soars after IPO and insider buying

• Pershing Square 2025 Annual Report

• Insider Signal trade page

Between March 23 and April 21, 2026, four separate short positions hit the oil futures market. Combined notional value: $2.72 billion. Each trade preceded a Trump or Iran announcement that pushed crude prices down. A fifth trade, a $1.5 billion buy of S&P 500 E-mini futures, cleared in the same one-minute window as the first oil short.

The CFTC opened a probe in mid-April (Bloomberg, April 15). Four congressional letters have pressed the SEC and CFTC for action. Bloomberg, the Financial Times, Reuters, and LSEG have confirmed every trade size through exchange records.

No traders have been publicly identified.

Four trades, four announcements

March 23, 6:49 a.m. ET. Roughly 6,200 Brent and WTI contracts cleared in one minute. Notional: $580 million. Average volume for that time slot over the prior five days was 700 contracts. Nine times normal. Fifteen minutes later, Trump posted a five-day strike postponement on Iranian power plants and “very good and productive conversations” (FT via Yahoo Finance). The S&P 500 surged. Stocks rallied roughly $1.7 trillion in minutes.

April 7, 19:45 GMT. A short of 8,600 lots split between Brent and WTI. Notional: $950 million. LSEG called the volume “entirely atypical for that time period” (Reuters via Investing.com). Roughly three hours later, Trump announced a two-week ceasefire on Truth Social. WTI fell 16.41%. Brent fell 13.29%. Both posted their biggest single-day declines since April 2020 (CNBC).

April 17, 12:24 GMT. A 7,990-lot Brent short. Notional: $760 million. Twenty minutes later, Iranian Foreign Minister Abbas Araghchi posted on X that the Strait of Hormuz was “completely open” (Al Jazeera). Brent fell 11.5% on the day.

April 21, 19:54 GMT. A 4,260-lot Brent sell-side trade. Notional: $430 million. Placed during post-settlement hours when volume is typically minimal (Reuters via Investing.com). Sixteen minutes later, Trump extended the ceasefire indefinitely. The S&P 500 and Nasdaq closed at records.

Subscribe now

Why these trades stand out

Four flags repeat across the four positions:

• Volume multiples of 9x to 30x normal for the time slot

• One-minute concentration windows

• All four positioned for prices to fall

• All four preceded announcements that did, in fact, send prices falling

The CFTC’s Large Trader Reporting System captures every position above its thresholds. The agency almost certainly already knows who placed each trade. Public identification requires an enforcement subpoena, not voluntary disclosure.

Who is asking questions

Rep. Ritchie Torres (D-NY) has sent three separate letters to the SEC and CFTC, one for each new trade (Torres). Rep. Sam Liccardo (D-CA) cited violations of the Securities Exchange Act of 1934, the Commodity Exchange Act of 1936, and the STOCK Act of 2012. His letter said “the timing indicates bets were placed by those with advance knowledge of the President’s action” (Liccardo).

Sen. Elizabeth Warren and Sen. Sheldon Whitehouse filed a joint CFTC letter on the $950 million trade (Senate Banking). Sen. Chris Murphy called the $1.5 billion S&P E-mini position “mind blowing corruption” (The Hill).

Paul Krugman wrote that the pattern raises the question of “whether decisions about war and peace in part serve the cause of market manipulation rather than the national interest” (Krugman Substack).

The White House warned staff against placing Iran war bets on prediction markets on April 10 (CNBC).

Share

The Hegseth sub-plot

A Morgan Stanley broker acting for Defense Secretary Pete Hegseth contacted BlackRock in February 2026 about purchasing iShares Defense Industrials Active ETF (IDEF) shares. The contact happened in the weeks before the February 28 strikes on Iran. The order did not execute because the ETF was not yet available through Morgan Stanley (CNBC, March 31).

Hegseth told the Senate Armed Services Committee that “no one owns me.” The Pentagon demanded retraction. The FT stood by its reporting.

The Hegseth inquiry is named. The oil trades are anonymous. Together they sketch the picture of an environment where war timing was tradable information.

What we don’t know

We don’t know whether one actor placed all four oil shorts or whether the positions were coordinated. We don’t know if the trader was a US person. We don’t know whether anyone in the executive branch held positions. The STOCK Act covers Congress, not the White House.

The STOCK Act gap is also why public Form 4 and Periodic Transaction Report data on members of Congress matters more than ever. Rep. Dave Taylor (R-OH) reportedly sold two oil company stocks during the war window. Every member’s filings are public. Track them at insidersignal.ai/politicians.

When one person can move oil 13% with a Truth Social post, the people closest to that decision hold a position no other market participant has. The May 5 ceasefire is fragile. Iran fired on US Navy ships. The UAE was attacked. Trump paused Project Freedom tanker escorts (CNBC).

More announcements are coming. More tests of the pattern.

All trade data is sourced from public reporting by the Financial Times, Bloomberg, Reuters, and LSEG, drawn from exchange records at CME Group and Intercontinental Exchange. The CFTC investigation is publicly disclosed. No individual trader has been publicly identified or charged. This article does not allege illegal insider trading. It documents publicly reported trade timing alongside publicly reported announcement timing.

In February, I wrote about ricosuave666, the Polymarket account that won four straight bets on Israel-Iran escalation, vanished, came back to bet on Iran attack timing. He went silent again the same week the IDF opened a classified leaks investigation.

My argument then was simple, Prediction markets had become an intelligence channel, and nobody was prosecuting it.

Until now.

Share

On December 26, 2025, U.S. Army Master Sergeant Gannon Ken Van Dyke opened a Polymarket account. Between December 27 and January 2, he placed 13 bets on Venezuela and Maduro contracts. He staked approximately $33,034. He walked away with approximately $409,881 in profit.

The trigger was Operation Absolute Resolve, the classified U.S. mission to capture Maduro and his wife. Van Dyke helped plan and execute it and had signed nondisclosure agreements.

He used a VPN, routed funds through a foreign cryptocurrency vault, and asked Polymarket to delete his account three days after the operation became public.

In the predawn hours of January 3, 2026, U.S. forces apprehended Maduro and his wife in Caracas. Hours later, Trump announced the operation.

Van Dyke’s “Yes” contracts resolved at $1.00. He withdrew most of the proceeds the same day.

On April 23, 2026, the U.S. Attorney for the Southern District of New York unsealed Van Dyke’s indictment. The CFTC filed a parallel civil complaint.

Both agencies called it the first insider trading case ever brought involving event contracts. The CFTC also invoked the so-called “Eddie Murphy Rule” for the first time, the Commodity Exchange Act provision named after Trading Places that bars federal employees from trading on nonpublic government information.

If convicted, Van Dyke faces up to 50 years across five counts.

This is the case that ends the debate. The CFTC, DOJ, and at least seven states are now competing to prove insider trading laws apply to prediction markets.

Subscribe now

The trades

Van Dyke’s positions were not subtle.

He started with sub-$100 “Yes” share purchases on contracts including “U.S. Forces in Venezuela by January 31” and “Maduro out by January 31.” On January 2, the eve of the raid, he placed more than $26,000 in concentrated bets.

The contracts paid out. He had operational knowledge of the timing, but the market did not.

The legal theory matters more than the dollar figure here. Prosecutors charged Van Dyke under commodities fraud, wire fraud, and unlawful monetary transaction statutes, anchored by misappropriation of material nonpublic information.

The same theory applies to corporate insiders trading event contracts on confidential business information. Law firm analyses now warn that Polymarket itself could be liable under similar theories.

So now, the “prediction markets are different” argument is dead.

The CFTC sets the table

The Van Dyke case did not arrive without warning. The CFTC spent the first quarter of 2026 building the legal scaffolding.

February 25, 2026: The CFTC Division of Enforcement issued an advisory tied to two enforcement actions on KalshiEX. It reaffirmed that trading event contracts on misappropriated MNPI can violate Section 6(c)(1) of the Commodity Exchange Act and CFTC Rule 180.1.

March 12, 2026: The CFTC issued an advance notice of proposed rulemaking on prediction market oversight, specifically requesting comment on how inside information should be regulated in event contracts.

March 31, 2026: CFTC Director of Enforcement David Miller delivered a speech at NYU Law School naming five enforcement priorities. Insider trading led the list. Miller attacked what he called the “myth” that insider trading laws do not apply to prediction markets, calling that view “flatly wrong.”

The Van Dyke indictment dropped 23 days after Miller’s speech.

Kalshi suspends three congressional candidates

On April 22, 2026, Kalshi disclosed disciplinary actions against three U.S. political candidates who bet on the outcomes of their own races.

• Matt Klein, Minnesota state senator running for the 2nd Congressional District. Bet $50 on his own primary. Fined $539.85, suspended five years. Cooperated.

• Ezekiel Enriquez, Republican primary candidate for Texas’s 21st Congressional District. Bet under $100. Fined $784.20, suspended five years. Cooperated.

• Mark Moran, former Virginia Senate Democratic primary candidate now running as an independent. Bet $100 across two markets. Refused to settle. Fined $6,229.30. Later told the AP he placed the bets deliberately because he “wanted to get caught.”

The fines are small, but they’re meant to act as a warning. Kalshi has opened roughly 200 insider trading investigations in the prior year and positions itself as the “first line of defense” the CFTC has demanded.

The platform is making compliance it’s new moat. Kalshi is valued at around $22 billion with weekly volume exceeding $1 billion. Its outdoor ads in D.C. now read “We ban insider trading” and “We operate within U.S. law.” in contrast with Polymarket.

On April 15, 2026, CEO Tarek Mansour spoke at the Semafor World Economy summit and called on the DOJ to prosecute prediction market insiders.

“If you commit insider trading on Kalshi, that can and will at some point be a federal crime,” Mansour said. “I actually do expect the DOJ to prosecute some of these cases.”

Eight days later, SDNY unsealed Van Dyke.

The state lawsuits

Federal enforcement is one front, but State attorneys general have opened a second.

On April 23-24, 2026, Wisconsin AG Josh Kaul filed three lawsuits in Dane County against Kalshi, Polymarket, Robinhood, Coinbase, Crypto.com, and affiliated entities.

The complaints accuse the platforms of facilitating illegal sports betting under Wisconsin Statute 945.03(1m). Sports contracts reportedly account for around 90% of Kalshi’s business, generating over $1 billion in annualized revenue.

Wisconsin is one front in a multi-state war:

• Arizona filed criminal charges against Kalshi for operating an unlicensed gambling operation.

• Nevada courts extended an order on April 3, 2026 blocking Kalshi from offering sports markets.

• New York AG Letitia James sued Coinbase and Gemini on April 21 over allegedly illegal prediction-style products.

• The CFTC has sued Connecticut, Arizona, and Illinois directly to defend its claim of exclusive federal jurisdiction.

Federal courts are split. The Third Circuit recently sided with Kalshi on sports contracts. Courts in Ohio, Maryland, and Nevada have ruled against the platform. The conflict is heading to the Supreme Court.

Why this matters for traders

The legal infrastructure that took 90 years to build around securities markets is being compressed into 90 days for prediction markets. Here’s what matters:

The MNPI perimeter just expanded. Compliance teams at hedge funds, public companies, and federal agencies now need to treat event contracts as in-scope for insider trading policies. Point72 and Balyasny have reportedly already banned employees from trading on prediction markets. Expect that policy to become standard. Anyone with access to material nonpublic information, whether on M&A, FDA decisions, congressional votes, or military operations, is now exposed to commodities fraud and wire fraud liability if they trade event contracts on it.

Surveillance is the moat. Kalshi’s 200 investigations and three candidate suspensions are not a marketing exercise. They are the price of staying inside the regulatory perimeter. The Van Dyke case shows Polymarket’s offshore status no longer protects U.S. citizens with a U.S. data trail.

Liquidity will reprice information edge. When suspicious clusters of accounts can profit five and six figures on binary military catalysts with no enforcement, the implied probabilities are noise. When the CFTC and DOJ start prosecuting, the noise drops out. That makes prediction markets more useful as a forecasting tool and less useful as an insider’s ATM.

Leave a comment

The pattern moving forward with insider trading

Insider trading enforcement in any new market follows a consistent arc. The market grows. Suspicious patterns accumulate. A regulator issues a warning. A speech follows. Then a single case lands that proves the legal theory works. After that, enforcement scales.

Equity markets ran that arc over decades. Crypto ran it over years. Prediction markets ran it in 2 months days, from the CFTC’s February 25 advisory to the April 23 Van Dyke indictment.

The first corporate insider case has not landed yet. The legal theory works just as well against a CFO trading event contracts on earnings, an FDA reviewer trading on approval timelines, or a private equity associate trading on deal closes. It is coming.

The headline is that 13 trades were enough to anchor a federal prosecution, and the legal theory now applies to every event contract on every regulated platform.

When we wrote about ricosuave666 in February, the closing line was that the markets remain open, the insiders keep betting, and anyone paying attention can watch information asymmetries play out in real time. The markets are still open. The insiders are still betting.

But the prosecutions have caught up, and they’re likely just getting started.

All trades, indictments, and regulatory actions discussed are publicly disclosed via SDNY filings, CFTC complaints, state court records, and platform disclosures. For ongoing coverage of insider trading patterns across equity and event markets, visit insidersignal.ai.

On March 23, 2026, a trust controlled by the wife of Rep. Ro Khanna (D-CA) purchased shares of Intel (INTC) at $1,000 to $15,000 and Oracle (ORCL) at $15,000 to $50,000. The trades were filed under the STOCK Act on April 7. Intel closed that week around $42. Oracle closed at $153.65. One month later, Intel is near $66 and Oracle is at $184. That is a 57% gain on Intel and a 20% gain on Oracle in 31 days.

View all their trades on insidersignal.ai/politicians/K000389.

Subscribe now

Why Intel went up 60%

Intel had been stuck in a congestion zone between $42 and $48 for months. Then catalysts fired in sequence.

April 1: Ireland Fab 34 buyback. Intel agreed to repurchase Apollo’s 49% stake in its Leixlip fabrication facility for $14.2 billion. The deal consolidated ownership of one of Intel’s most advanced manufacturing sites.

April 7: Terafab. Intel joined Elon Musk’s Terafab consortium alongside Tesla, SpaceX, and xAI. The target is 1 terawatt per year of compute for AI.

April 9: Google Cloud. Intel signed a multiyear data center chip partnership with Google covering future Xeon generations and custom IPUs. INTC closed up 4.7%.

April 16: INTC closed at $68.50, an all-time high.

April 21: HSBC Street-high $95. HSBC upgraded from Hold to Buy and raised its target from $50 to $95. BNP Paribas upgraded the same day from Underperform to Neutral.

April 22: Elon Musk confirmed on the Tesla earnings call that Tesla will use Intel’s 14A process for Terafab Austin. INTC popped 3% after hours.

Intel has added roughly $137 billion in market cap since the March 21 Terafab reveal. The stock is up 270% from its April 2025 low near $18.

Khanna’s trust bought before any of this was public. The Ireland buyback came nine days later. Terafab came fifteen days later. Google Cloud came seventeen days later. HSBC came twenty-nine days later. The timing speaks for itself.

Share

Why Oracle went up 20%

Oracle was the bigger position. The trust sized it at $15,000 to $50,000, up to 50 times larger than most of the other tech holdings.

Two weeks before the purchase, Oracle reported fiscal Q3 2026 results. Revenue up 22% year-over-year. Cloud up 44%. IaaS up 84% with four straight quarters of acceleration. Remaining performance obligations hit $553 billion, up 325%, including the $300 billion OpenAI Stargate deal. FY2027 revenue guide: $90 billion.

Then the catalysts fired.

March 24: Oracle launched 22 AI-driven Fusion Agentic Applications. Bank of America set a $200 price target.

March 31: Oracle expanded AI infrastructure for U.S. government customers, adding Nvidia B300 GPUs, xAI Grok, and Nvidia Nemotron to OCI government regions. The announcement lands squarely in Khanna’s subcommittee jurisdiction.

April 14: Oracle expanded its Bloom Energy partnership to 2.8 GW of fuel cells for AI data center power. ORCL up 4.9%.

April 16: AWS multicloud. Oracle and AWS announced direct private link integration. ORCL rallied 14% to 20% over the next five sessions.

April 22: Google Cloud expansion. Oracle integrated its AI Database with Google Gemini Enterprise across 15 regions. ORCL closed up 3.68% at $184.12.

Consensus Oracle price target across 35 analysts is $246 to $261. Mizuho and Guggenheim sit at $400.

Intel is a Hold. Oracle is a Buy. The trust owns both.

The other 34 trades

Intel and Oracle sat inside a 36-stock cluster bought across March 19 and March 23, all filed April 7.

The larger positions at $15,000 to $50,000: Honeywell, BWX Technologies, Oracle, Stryker, and Mid-America Apartment. BWX Technologies is a nuclear components supplier with direct defense revenue. Both Oracle and BWXT sit in sectors where Khanna’s Armed Services subcommittee on Cyber, Innovative Technologies and Information Systems has oversight.

The tech winners in the smaller basket are significant. Western Digital up 56%. KLA Corp up 20%. CrowdStrike up 16%. PayPal up 12%. Verisign up 10%. Software laggards like Workday, Autodesk, and Fair Isaac are down single digits, but position sizing means the dollar-weighted result skews heavily toward the winners.

The S&P 500 is up 8.5% since March 23. The SOXX semiconductor ETF is up 28%, its best April ever. The trust front-loaded the exact sleeve of the market that ripped.

The track record

From January 2024 through April 2026, trades associated with Khanna have produced an estimated 112.1% excess return over the S&P 500. That is nearly three times Pelosi’s 38.5%.

Anthony Pompliano posted on April 8: “Nancy Pelosi take a seat. There is a new king.” Elon Musk quote-tweeted: “Wow.”

In 2025, Khanna logged 4,081 trades totaling roughly $50 million, the most of any member of Congress. His wife’s trust netted a 143% return on NVIDIA bought in February 2024. Net worth: $232.7 million.

The paradox

Khanna’s statement to Newsweek on April 15: “I have been a leader to ban Congressional stock trading, do not trade, and have no input in the trades filed by my wife’s trust.”

Primary challenger Ethan Agarwal, backed by Garry Tan and Chamath Palihapitiya, is not buying it. On April 20 he posted his first attack ad: “Today is the beginning of the end for @RoKhanna. We are highlighting Ro’s hypocrisy around $600,000,000 of trades since being elected to Congress.” Primary is June 2.

Meanwhile the HONEST Act advanced out of Senate Homeland Security Committee on a bipartisan 8-7 vote, with Sen. Josh Hawley joining Democrats. It bans members and spouses from trading individual stocks with a 180-day divestment window.

Khanna sits on the subcommittee that oversees defense AI procurement. His wife’s trust trades AI and semiconductor stocks. The trust beats the S&P 500 by 112%. He advocates banning the practice that produces the returns.

Intel reports Q1 2026 earnings today after the close. Options markets are pricing an 11% to 12% swing. If Intel lands the quarter, the position may have doubled in under 45 days.

All trades are publicly disclosed under the STOCK Act, available through House financial disclosures and insidersignal.ai/politicians/K000389.

Intel up 60% in one month. Oracle up 20% in one month. Thirty-six stocks in one filing. 112% excess return over the S&P. The most prolific trader in Congress says he wants to ban the game. He just keeps winning it instead.

On March 25, 2026, Rep. Josh Gottheimer (D-NJ) purchased Microsoft call options valued at up to $1.1 million. The trades were disclosed in a STOCK Act filing on April 8, 2026. Two tickets, both expiring June 18, 2026. Microsoft reports fiscal Q3 2026 earnings on April 29, 2026.

This is the most leveraged political bet on a known earnings catalyst in the tape this week. It is also the kind of trade that happens when the trader is close to the story.

The trade

Per the Form PTR Gottheimer disclosed April 8:

March 25, 2026: MSFT call options, strike $320, expiry 6/18/2026, value $500,001 to $1,000,000.

March 25, 2026: MSFT call options, strike $325, expiry 6/18/2026, value $50,001 to $100,000.

Both positions sit out of the money at the time of purchase. Both expire 51 days after Microsoft reports earnings. This is not a hedge or a stock surrogate. It is a directional bet on a specific catalyst.

Subscribe now

Why this specific catalyst

Microsoft announced on April 8 that Q3 FY2026 earnings release will be April 29, 2026 after market close. The last print mattered: Q2 FY2026 adjusted EPS of $4.14 on revenue of $81.27 billion landed on January 28, and MSFT dropped nearly 10% the next day as Azure growth slowed to 39%.

The stock has spent the subsequent two months digesting that move. Short-dated calls struck just above the spot price are the classic expression of a single view: that Q3 Azure comes in better than the buyside fears. Gottheimer paid a premium to get that exposure.

Who is doing the buying

Josh Gottheimer has three data points that make this trade unusual.

He was a Microsoft executive. Before Congress, Gottheimer served as General Manager of Corporate Strategy and Advertising at Microsoft from 2012 to 2015. He knows the product lines, the sales motion, and the internal culture of the company he is now leveraged long.

He sits on the House Permanent Select Committee on Intelligence. He is Ranking Member on the NSA and Cyber Subcommittee and serves on the Subcommittee for Defense Intelligence and Overhead Architecture. In that seat he receives classified briefings on cybersecurity threats and federal AI deployment. Microsoft is one of the largest commercial providers of both.

He co-chairs the House Commission on Artificial Intelligence. The commission was established in December 2025 to scrutinize AI’s impact on the economy and recommend policy. Gottheimer is one of two co-chairs. Microsoft’s enterprise AI business is the single largest driver of the Q3 2026 print analysts are trying to model.

The track record

Gottheimer is not new to Microsoft options. A January 2025 Capitol Trades analysis flagged nearly $40 million in MSFT options across his account. The March 25 trade fits a pattern, not a first-time event.

Since 2017, Gottheimer has made more than 3,300 disclosed transactions and compounded at roughly 22.93% annually against the S&P 500’s ~10%. He has not been charged with any STOCK Act violation. He has not been the subject of any SEC, DOJ, or House Ethics investigation for insider trading. Everything here is legal and disclosed.

It is also conspicuous. A sitting Intel Committee member, former Microsoft executive, and AI commission co-chair, leveraged long the company that sits at the center of the AI policy conversation, through short-dated call options timed to the earnings print.

Leave a comment

Not the only MSFT buyer in Congress this week

Rep. Richard McCormick (R-GA) disclosed an April 7 filing that included a Microsoft common stock purchase alongside buys of Abbott Labs, BlackRock, Home Depot, and ADP, and sales of Johnson & Johnson, L3Harris, and Analog Devices. The dollar amount is small ($1,001 to $15,000) and does not compare to the Gottheimer options book.

What it does do is confirm that at least two members of Congress positioned long into the April 29 Microsoft print during the same disclosure window.

Share

What to watch

April 29 earnings. Azure growth is the swing factor. A print above 40% year over year would re-rate the stock higher, put both Gottheimer call strikes in the money, and validate the trade. A miss would put Azure below 39% and leave the $320 and $325 strikes worthless at expiration.

Additional disclosures. Any PTR filed by Gottheimer between now and June 18 that trims or closes the position will be public within 45 days. The exit trade tells you more than the entry.

STOCK Act reform. H.R. 1908 and the Gillibrand-Hawley Senate bill remain pending. Trades like this are the case study both bills use.

All trades discussed are publicly disclosed under the STOCK Act, available through House financial disclosures and aggregated on Quiver Quantitative, Capitol Trades, and insidersignal.ai.

A former Microsoft executive who sits on the House Intelligence Committee and co-chairs the House AI Commission bought up to $1.1 million in out-of-the-money Microsoft calls 35 days before the company’s most important earnings print of the year. He has run this trade before. His track record on it is 22.93% annually. The filing is legal. The question the filing does not answer is why the timing is always this good.