Wealth Signal

Wealth Signal

Edition 037 Monday 19 October 2026 Coverage 2–19 Oct 2026 Core window 12–19 Oct

Anomaly Board

3 breaks

Actions that break the actor’s own prior pattern. Comparison is to that person’s own history, not to a peer group.

01

Steven Schuurman · Elastic N.V. (NYSE: ESTC), co-founder & director

Trades 5 & 12 Oct 2026 · Forms 4 filed 6 & 13 Oct · Form 144 filed 13 Oct

Through his Dutch holding company CMXI B.V. he sold 1,500,000 shares at $90.13 on 5 October and 2,000,000 at $94.15 on 12 October. The reported position went 6,004,600 → 4,504,600 → 2,504,600: 58.3% of the stake in eight calendar days, roughly $323.5m gross. Neither Form 4 checks the Rule 10b5-1 box and neither reports a plan.

Why that breaks his own pattern: the Form 144 filed 13 October lists exactly one sale in the trailing three months — the 5 October block. He went from one disclosed disposal a quarter to two blocks in eight days, the second 33% larger and executed 4.5% higher than the first. The stake crossed below 5% of Elastic’s 104,991,280 shares outstanding on 5 October (5.72% → 4.29%) and now sits at 2.39%.

02

Joseph Lubin · SharpLink, Inc. (NASDAQ: SBET), director

Trade 8 Oct 2026 · Form 4 signed 13 Oct 2026

Consensys Software Inc., which Lubin runs as CEO, sold its entire 5,676,952-share common position in one block at $8.39 — $47.63m, to zero. A 100%-to-nothing disposal by SharpLink’s own strategic advisor is a first.

But the same filing shows the man did not sell a share. Lubin retained 5,154,213 shares directly — a position 91% as large as the one the entity dumped — plus 1,200,000 through Permanent Highest Power Capital LLC (held by the Gradient Ascent Trust, his son a co-trustee), 975,600 through ConsenSys AG and 81,300 through Ethereal Ventures Fund II. CSI also kept all 3,455,019 warrants, struck at $6.15/$6.765/$7.38/$7.995 and running to 30 May 2030. Sold: 43% of what he is deemed to beneficially own, 0% of what he owns outright.

03

Robert C. Pohlad · PepsiCo (NASDAQ: PEP), director

Transaction dated 14 Oct 2026 · Form 4 signed 16 Oct 2026

Reported under transaction code P for 184,646 shares at $128.28 — about $23.7m, carried everywhere as the largest PepsiCo director purchase in years, position up 74.35% to 433,009. PEP rose 2.4% on Friday 16 October on volume roughly 31% above average.

What breaks the pattern is that no shares were bought. Footnote 1 states the shares came out of irrevocable trusts he had established, pulled back under his power to substitute assets of equivalent value, with his revocable trust paying in cash of equal fair market value priced at the average of the day’s high and low. Footnote 2 states that 168,632 of the shares in the “owned after” figure were already his and were merely re-registered into that trust. Net new money into PepsiCo: zero.

Checked and found unremarkable: David Duffield’s $18.7m Workday sale (10b5-1 plan adopted 2 Dec 2025, 35,480,751 Class B untouched); Gabriel Mecklenburg’s Hinge Health sale (plan adopted 12 Jun 2026, 2,720,306 shares retained); Carlos Slim’s Control Empresarial trimming Talos Energy and PBF (continuation of an established cadence); the Silver Lake Dell distributions (fund mechanics, not an individual). No domicile changes, no first-time share pledges, no personal guarantees, no resignation-plus-disposal pairs and no new trust or holdco formations surfaced in this window. Three breaks is what the window produced; the board is not padded.

§ 02

The Log

five person-actions

The PepsiCo purchase that contained no purchaseRobert C. Pohlad · PepsiCo, Inc. · 14 October 2026

Confirmed · primary document Measurement failure Code P No 10b5-1
What the document says
Form 4, transaction date 14 October 2026, code P, four acquisition lines of 84,507 / 45,075 / 45,075 / 9,989 shares, each at $128.28, total 184,646 shares. All four are reported as indirect, held by a revocable trust. Shares owned after: 433,009 indirect, 22,953.5012 direct, 27 through a spouse’s revocable trust. Footnote 1: the shares came from irrevocable trusts the reporting person established, acquired “using his power to substitute assets of equivalent value”, with his revocable trust giving those trusts cash of equal fair market value “based on the average of the day’s high and low prices on October 14, 2026”. Footnote 2: 168,632 of the shares in the total were previously reported as directly owned and were moved into the revocable trust. Signed by Cynthia A. Nastanski as attorney-in-fact.

Consensus read

A director from the family that has run Pepsi bottling for three generations put $23.7m of his own money into the stock while it sits near multi-year lows, lifting his stake by three quarters. Insider buying of that size at a depressed multiple is the strongest signal available in the Form 4 universe: directors sell for a hundred reasons and buy for one.

The market agreed and marked the shares up 2.4% the following session on heavy volume.

Contrarian read

Nothing was purchased and the filing carries no information about PepsiCo’s value. This is a basis swap: Pohlad used a grantor-trust substitution power to pull low-basis PepsiCo stock back out of irrevocable trusts and push cash of identical value in. Family-level PepsiCo exposure is unchanged to the share.

Inference, not stated in the document: the purpose is almost certainly IRC §1014. Appreciated stock sitting in the grantor’s own estate gets a basis step-up at death; cash has no built-in gain to waste on one. The swap moves the asset to the pocket where the step-up is worth something.

Evidence that discriminates

The price field. Every genuine open-market trade in today’s log carries a weighted-average price with a stated range footnote — Duffield’s six tranches spanning $185.695–$191.6999, Mecklenburg’s four spanning $101.00–$104.72. Pohlad’s is a single clean $128.28 which the footnote itself defines as the average of the day’s high and low. That is a valuation convention used in trust work, not an execution. No broker fills exactly at the midpoint of the day’s range, four times, on four separate lines.

The line structure. The transaction is split across four sub-amounts against four separate irrevocable trusts. A market order for 184,646 shares is not routed that way; trust-by-trust substitutions are, because each trust is a separate legal counterparty.

The arithmetic behind “+74.35%”. Footnote 2 removes 168,632 shares from the numerator — they were already his, simply re-labelled from direct to trust-held. The percentage compares a post-consolidation trust line against a pre-consolidation one. It is not a measure of anything he did with money.

Note what is not evidence here: the stock rising 2.4% afterwards is consistent with both readings, since the market cannot distinguish them from the headline either. That is the point.

What would falsify me

A Form 4/A, a Form 144, or a broker execution record showing an actual market trade on 14 October 2026. Or PepsiCo’s next proxy showing Pohlad’s aggregate beneficial ownership up by roughly 184,646 shares year-on-year rather than flat. Either would mean real dollars moved and the open-market reading was right.

Pattern vs his own historyI could not retrieve his earlier PepsiCo Form 4 record inside this scan, so I am not claiming this is his first substitution — families that hold through revocable and irrevocable trusts in parallel typically do this repeatedly. What is established by the 14 October filing alone is that it contains no open-market execution, and that at least three outlets reported it as one.

0
3

Eight days, 58% of the stake, no plan — and the price went upSteven Schuurman / CMXI B.V. · Elastic N.V. · 5 & 12 October 2026

Confirmed · primary document Anomaly No 10b5-1 Mechanism unproven
What the documents say
Form 4 filed 6 Oct: sale of 1,500,000 ordinary shares on 5 Oct at $90.13, 4,504,600 owned after, indirect, held of record by CMXI B.V. Form 4 filed 13 Oct: sale of 2,000,000 on 12 Oct at $94.15, 2,504,600 owned after. Both report Schuurman as Director only — the 10% Owner box is not checked on either. Neither marks the Rule 10b5-1(c) box. Both signed by Marielle Reints under power of attorney. Form 144 filed 13 Oct via Goldman Sachs: 2,000,000 shares, aggregate market value $191,040,000, against 104,991,280 shares outstanding. Acquisition history in that notice: 7,290,000 shares on 28 Apr 2015 from JTG Ventures B.V. for €7,290 in total, and 5,186,000 on 17 Dec 2020 in the IXII B.V. merger for €0.

Consensus read

A co-founder and sitting director sold 58% of his remaining stake in eight days with no trading plan to hide behind. Plan-free selling means he picked the dates himself. Elastic has run hard into an AI-search narrative that every large-cap infrastructure vendor is now attacking; the person who built the thing is taking the other side while the window is open.

Benzinga headlined it as a “sell strategy”; the shares fell 5.7% in the days around it.

Contrarian read

The trigger is a threshold, not a thesis. The stake entered the window at 5.72% of shares outstanding. The first block took it through the 5% line to 4.29%; the second took it clear to 2.39%.

CMXI B.V. is a Dutch holding company holding a Dutch N.V. Its gain is sheltered by the participation exemption only while the holding qualifies, and 5% is the qualifying line. Section 13(16) of the Dutch corporate income tax act carries the exemption forward for a holding that drops below 5% after having qualified — but as a continuation, with an end. Once you have decided to break 5%, the tax-rational move is to finish the sell-down inside that window and stop optimising on price.

Evidence that discriminates

He sold more, at a higher price, seven days later. A seller acting on adverse private information gets progressively worse fills: the first block teaches the market what is coming. Schuurman’s second block was 33% larger and executed 4.5% above the first. Price-insensitive size-clearing produces that shape; informed selling does not.

Both blocks were advertised in advance. Form 144 notices naming the exact share count went to the SEC ahead of each execution, through a single named broker. Pre-announcing a 2,000,000-share block — about 1.9% of shares outstanding — maximises the concession you pay. That is what a seller who must move size does. It is the opposite of what a seller with an edge does.

The basis makes tax, not price, the dominant variable. The Form 144 puts the acquisition cost at roughly €0.001 per share on the 2015 tranche and zero on the 2020 tranche. When basis is nil, a 4% move in the share price is noise next to whether the gain is exempt or taxed.

Honest limit. I could not confirm from a citable source how long section 13(16) runs. The Meijburg commentary confirms the scheme exists and names the article; the PwC and Taxand country guides do not cover the below-5% case at all. So the tax-clock mechanism is a live hypothesis, not a demonstrated fact, and I am not going to pretend otherwise.

The three live hypotheses, and what separates them

(A) Threshold-driven. Prediction: the residual 2,504,600 shares go in further blocks on a short horizon, at whatever price is available.

(B) Ordinary diversification after a run. Prediction: the selling stops or slows sharply now that the stake is under 2.5% and the concentration problem is solved.

(C) Informed exit. Prediction: ESTC underperforms materially on company-specific news before Elastic’s fiscal Q2 report, traceable to something knowable inside the company on 5 October.

These are genuinely distinguishable within one quarter, which is unusual and worth the watch.

What would falsify me

Hypothesis (C) coming true: ESTC down materially on company-specific bad news before the next results, dated to something that existed on 5 October. Also falsifying, in the other direction: the residual stake sitting untouched for two quarters while the stock trades higher — that would make this ordinary diversification, and the threshold story an invented pattern.

Pattern vs his own history · and a filing discrepancy worth notingThe Form 144 of 13 October lists exactly one sale in the trailing three months, which is the step change. That same table dates the prior sale 2 October at an implied $92.25 ($138,375,000), while the Form 4 dates it 5 October at $90.13 ($135,195,000) — a $3.2m gap between the broker’s notice table and the Section 16 record. The Form 4 is the authoritative document; the 144 entry appears to be an estimate carried forward. Worth knowing that the two filings do not reconcile.

0
3

The entity exited; the man did not sell a shareJoseph Lubin / Consensys Software Inc. · SharpLink, Inc. · 8 October 2026

Confirmed · primary document Anomaly Five ownership lines
What the document says
Form 4 signed 13 October 2026. Transaction 8 October, code S, 5,676,952 shares disposed at $8.39, owned after: 0, indirect, by Consensys Software Inc. Footnote 1: CSI held the shares directly, Lubin is its CEO and may be deemed to control it, and CSI sold all of them in a block trade. Four further ownership lines are reported and untouched: 5,154,213 direct; 1,200,000 through Permanent Highest Power Capital LLC, owned by the Gradient Ascent Trust (Lubin’s son a co-trustee, immediate family the beneficiaries); 975,600 through ConsenSys AG; 81,300 through Ethereal Ventures Fund II L.P. Table II: 3,455,019 common stock purchase warrants held by CSI, struck at $6.15 (1,382,007), $6.765, $7.38 and $7.995 (691,004 each), issued 30 May 2025 under a Strategic Advisor Agreement and expiring 30 May 2030. None transacted.

Consensus read

The public face of SharpLink’s ether-treasury strategy is cashing out. His entity sold every common share it held, in one block, to zero. The stock fell 7.3% on the news. Digital-asset treasury vehicles trade on the premium their sponsor’s credibility supports, and the sponsor just took $47.6m off the table.

Contrarian read

What was sold was a fee, not a position. Footnote 5 records that CSI’s warrants were received as compensation under a Strategic Advisor Agreement; the common stock was advisory consideration parked on an operating company’s balance sheet. An operating company monetising a fee-in-kind is treasury management, not a view.

Meanwhile the holding that does express a view — the 5,154,213 shares Lubin owns outright, with no corporate balance sheet, no co-trustee and no fund investors attached — did not move.

Evidence that discriminates

The ratio inside a single filing. The block that sold was 5,676,952 shares. The direct personal holding left in place is 5,154,213 — 91% as large. Across the four untouched lines he retains 7,411,113 shares. He sold 43% of what he is deemed to beneficially own and 0% of what he owns outright. If the thesis had broken, the direct stake is the cheapest, fastest, least encumbered thing to sell, and it is the one thing that did not move.

Concession I have to make. At $8.39 the $7.995 warrant tranche is only 4.9% in the money, so “kept the convexity” is a thinner claim than it first looks — and warrants issued under an advisor agreement may be contractually non-transferable, which would make retention a constraint rather than a choice. That is exactly why I rest the argument on the direct-holding ratio and not on the warrants. Evidence consistent with both readings is worth nothing here.

What stays unknowable. Why CSI needed $47.6m on 8 October is not in the document. An operating-company funding need, a decision to de-risk a balance-sheet position in a correlated asset, and a quiet first step out of a larger stake are all consistent with what is filed.

What would falsify me

A Form 4 within 90 days showing Lubin disposing of any meaningful part of the 5,154,213 direct shares, or CSI exercising and selling the warrant tranches. Either converts this from “an entity monetised a fee” to “the principal is leaving, and started with the pocket that attracted least attention.”

Pattern vs his own historyI could not retrieve his full prior SBET Form 4 record — the per-ticker insider search was unreachable from this scan — so I am not claiming this is his first entity-level disposal. What the 13 October filing establishes on its own is the composition above: one line to zero, four lines untouched, and the untouched lines larger in aggregate than the one that sold.

0
3

“$3.3 billion since IPO” is a cumulative total doing the work of a signalMichael Intrator, Brian Venturo, Brannin McBee · CoreWeave · reported 16–17 October 2026

Multi-source Measurement attack Plan status unverified
What the reporting says
Aggregated from SEC filings and syndicated on 16–17 October: founder sales of $3.3bn since the 28 March 2025 IPO, of which $2.4bn falls in 2026. Intrator (CEO) more than $907m since 1 January 2026, against under $70m in all of 2025. Venturo (Chief Strategy Officer) more than $1.2bn in total, including $734m in Q2 2026. McBee (Chief Development Officer) $1.06bn, described as 40.4% of his undiluted IPO holding. A separate $488m was sold privately in 2023–24, before listing. The company says the founders still own about 20% combined and called the sales “pre-scheduled, fully disclosed, and entirely standard.” The framing is set against a roughly 25% share-price fall and widening credit spreads.

Consensus read

Three founders took $3.3bn out while the stock has fallen 25% and the bonds have come under stress. Pre-scheduled or not, insiders converted billions of paper into cash and public holders are holding the drawdown. The gap between Intrator’s August statement that conviction “has never been stronger” and his $907m of 2026 selling is the tell. This is the first real crack in AI-compute financing.

Contrarian read

The headline number is a 19-month cumulative total being read as a statement about this month, and the comparison does not survive contact with the dates.

A cumulative total is monotonic: it can only rise. “Founders have sold $3.3bn” will be true and larger every month from now on no matter what any of them does next, which makes it structurally incapable of signalling anything about the present.

Evidence that discriminates

The time distribution. If founders were selling into the October credit deterioration, the sales would cluster in the weeks before it. They do not. The single largest component — Venturo’s $734m — is dated Q2 2026, months ahead of the October stress. The $488m of pre-listing sales is from 2023–24 and sits outside the stated window entirely.

McBee’s denominator is not a choice variable. “40.4% of his undiluted IPO holding” measures against a fixed historical quantity, not his current stake and not his current wealth. Percentages of a frozen denominator always look larger as time passes; they say nothing about what fraction of his exposure is left today.

The part that survives, and should be kept. Intrator’s $907m in 2026 against under $70m in 2025 is roughly a 13× change in his own run rate. That is a genuine break in one man’s cadence and belongs on the record — but it is dated to 2026 broadly, not to October, and it is being smuggled in under a number that cannot support it.

Honest limit. The reporting does not state that the sales ran under Rule 10b5-1; it relays the company describing them as pre-scheduled and then separately explains what a 10b5-1 plan is. That is a corporate characterisation, not a filing field. I did not independently check the plan boxes on the underlying Forms 4, so treat “pre-scheduled” as unconfirmed.

What would falsify me

Forms 4 showing a material block of founder selling dated inside the October drawdown, or showing the 10b5-1 box unchecked on the large 2026 sales. Either would mean the aggregate was not merely sloppy framing but was diluting a live, discretionary exit — and the consensus read would be the right one.

Pattern vs their own historyIntrator: roughly 13× his 2025 pace. Venturo: one quarter (Q2 2026) carries $734m of his $1.2bn lifetime total, which is itself a concentration worth asking about — but it is a Q2 question, not an October one. The honest headline would have been “CoreWeave founders sold $2.4bn in 2026, most of it by June.”

0
3

The donor list’s interesting column is the asset, not the amountTyler & Cameron Winklevoss, Brandon Lutnick and others · disclosed 15 October 2026

Single source · leaked document Named, dated Measurement attack
What the reporting says
An internal spreadsheet obtained by The Washington Post, reported 15 October 2026, tracks the 67 donations above $10,000 recorded to the White House ballroom fund as of late September 2025 — $333m from private donors. Named individuals include Charles Ergen ($12.5m in Dec 2025 and $12.5m in Feb 2026, anonymity requested), the Adelson family ($25m), Paul Singer ($5m, Oct 2025, through his foundation), Marc Benioff ($1m, Oct 2025, anonymity requested), Kelly Loeffler and Jeff Sprecher ($5m promised), Joseph Popolo Jr. and his wife ($1m from a family fund), Isabela Herrera ($2.5m committed, first payment $835,000 on 14 Oct 2025), and Harold Hamm (under $5m combined with Continental Resources). Two entries are not denominated in dollars given: Tyler and Cameron Winklevoss, bitcoin valued at $5,050,018, and Brandon Lutnick, stock worth over $2.5m.

Consensus read

A straightforward access story. The interest is in who paid, how much, and what each of them wanted; the anonymity requests and the Cabinet and ambassadorial connections are the substance. The dollar figures are the ranking.

Contrarian read

The amounts are the least informative column. Two of the named individuals gave appreciated property rather than cash, and that changes what the number means. A gift of long-held appreciated property to a qualifying recipient does not realise the embedded gain while still being recorded at full fair market value. For an asset with near-zero basis, the giver’s after-tax cost is a fraction of the figure on the list.

Read that way, the list is not rank-ordered by sacrifice. Ergen’s $25m in cash cost him $25m. The Winklevoss entry at $5,050,018 did not cost $5,050,018.

Evidence that discriminates

The odd cents. $5,050,018 is a quantity multiplied by a spot price. It is the fingerprint of an in-kind transfer marked at a particular moment. Cheques are written for round numbers, and every cash entry on the list is round: $25m, $12.5m, $5m, $1m. Lutnick’s entry is explicitly described as stock rather than a dollar amount. The in-kind/cash split inside one leaked list is the discriminating detail — without both forms sitting in the same table you could only speculate.

What this is not evidence of. Nothing here tells you these donors' cost basis, their holding periods, or what deductions if any were claimed. The mechanic is general; the application to these individuals is inference. Singer's entry going “through his foundation” is a third structure again, with its own treatment.

What would falsify me

The recipient entity’s own filings, or a corrected list, recording the Winklevoss and Lutnick contributions as cash — or showing the donors liquidated the bitcoin and stock first and gave the proceeds. Either makes the amount, not the asset, the whole story and leaves the consensus read intact.

Pattern vs their own historyNot assessable. I have no comparable prior donation record for these individuals in a form that would let me measure cadence, and this rests on a single leaked internal document reported by one outlet. Treated as single-source throughout.

0
3
§ 03

Where Consensus Is Right

tested and not overturned

Ryan Cohen’s GameStop buying is exactly what it looks likeRyan Cohen · GameStop Corp. · 13 October 2026

Confirmed · primary document Ordinary reading survives

Form 4 (accession 0000921895-26-002784), transaction date 13 October 2026: two purchase lines, 67,130 shares at $26.75 and 777,000 at $27.00, $22,774,728 in total, taking him to 42,492,652 shares. That follows roughly 700,000 shares for about $17.1m on 2 October and September blocks worth over $57m — close to $97m in six weeks, at a rising average cost, with the stock near 52-week highs.

What I tested. Whether the “P” code was concealing something structural the way Pohlad’s did. It is not. The filing carries no footnotes at all. Ownership form is D — direct, with no entity, no trust, no LLC in the nature-of-indirect-ownership field. Table II is empty: no warrants, no options, no convertibles, no prepaid forwards. The Rule 10b5-1 box is unchecked and no plan is referenced. I also checked for a pledge, which would appear as a footnote and does not, and for a 10% ownership crossing that would change his reporting obligations — he is around 9.1% including warrants and the 10% Owner box is unchecked.

Why the ordinary reading survives. A man spending his own after-tax money, in his own name, at escalating prices — $24.41 average on 2 October rising to $27.00 eleven days later — with no plan to shelter behind and full Section 16 short-swing exposure for six months afterwards, is doing the only thing on this page that cannot be re-explained by tax or structure. There is no pocket this moves between. The cash left his balance sheet and the shares arrived at market price.

The contrarian angle I went looking for would have been float or index engineering, or collateral staging ahead of a borrow. Neither is supported by anything in the document, and inventing one would be worse than reporting nothing. The consensus is that Cohen believes GameStop is worth more than $27. That is the correct reading, and the complete absence of footnotes is what makes it correct.

0
2
§ 04

What Repeats

one mechanic, five filings

The “% of holdings changed” column is measuring a reporting line, not a person. It appears in every insider screener and in most of the coverage downstream of them, and in three of today’s five filings it produces a number that is false about the human being.

Screener ΔOwn against the position actually held. Sources: Forms 4 filed 13–16 October 2026; percentages in the last column computed from the filings’ own “owned following transaction” fields across all reported lines.
PersonIssuerScreener ΔOwnActually sold / boughtMechanic
PohladPEP+74.35%0.0%Trust substitution + re-registration
MecklenburgHNGE−100%−6.9%Class B → Class A conversion
DuffieldWDAY−49%−0.28%Class B → Class A conversion
SchuurmanESTC−44%−44.4%Single-class, single entity — accurate
CohenGME+2%+2.0%Direct, one line — accurate

Mecklenburg, Hinge Health. The screener reads −100% — a founder clearing out. He converted 200,000 Class B into Class A on 12 October and sold exactly 200,000 Class A in four tranches, so the Class A line went to zero by construction. He still holds 1,644,004 Class B directly, 383,592 through a Family Trust and 692,710 through a GRAT: 2,720,306 shares. He sold about 6.9% of his position, under a Rule 10b5-1 plan adopted 12 June 2026.

Duffield, Workday. The screener reads −49%. Identical mechanic: 99,835 Class B converted, the same 99,835 Class A sold in six tranches under a plan adopted 2 December 2025. Class B remaining: 35,480,751. He sold 0.28% of his position — a rounding error presented as a halving.

The common fault. ΔOwn is computed from the “shares beneficially owned following transaction” field of the single security line the transaction touched. Dual-class founders and trust-holding families almost never transact on the line that holds their wealth: they convert into the liquid class and sell that, or they move assets between their own trusts. The consequence is systematic rather than random — any screen ranked on that column will preferentially surface the people whose actual exposure changed least, because the mechanics that shrink a reporting line to zero are exactly the mechanics used by people who are not leaving.

Two of today’s five are accurate, and the thing they have in common is simplicity: one share class, one holding line, one entity. Schuurman’s sale really was 44% of the line because the line is the whole position. Cohen’s +2% is right because he holds directly in one class. The screener is not broken; it is measuring a field that stops meaning what you think it means the moment a structure appears above the shares.

§ 05

Transferable To You

Ontario · private company · honest verdicts

1 · The grantor-trust basis swap (Pohlad)

Does not transfer
0
2

This is the sharpest mechanic on today’s page and it is almost perfectly non-transferable to Canada. It needs two American things at once: a grantor trust whose settlor can swap assets in and out without tax consequence, and a basis step-up at death that makes it worth doing. Canada has neither. The rule here is the opposite — a deemed disposition at fair market value immediately before death under s.70(5) of the Income Tax Act, plus the 21-year deemed disposition that most family trusts face. Pulling appreciated assets back into a Canadian estate makes the tax worse, not better.

The Canadian analogue to the goal — keeping future growth out of your own hands — is the estate freeze: exchange your common shares of the operating company for fixed-value preferred shares and let new common shares be issued to a family trust, so growth from that date accrues outside your estate. The freeze is a structure to put in place before value accumulates, not after a liquidity event, which makes the timing question the live one rather than the mechanic itself.

2 · Giving appreciated securities in kind rather than cash

Transfers — and is better here
0
2

This is the one mechanic on the page that works better in Canada than in the United States. A gift of publicly listed securities in kind to a registered charity takes the capital gains inclusion rate on that gain to zero under s.38(a.1), while the donation receipt is still issued for full fair market value. Sell first and donate the proceeds and you lose it: half the gain becomes taxable.

Done by a corporation rather than personally, it compounds. The entire capital gain is added to the capital dividend account instead of half of it, and the CDA balance can later be paid out to you as a tax-free capital dividend. On a security with a $50,000 fair market value and a $10,000 cost base, a cash sale adds $20,000 to the CDA; an in-kind gift adds $40,000.

FIRST STEP THIS WEEK · Check whether you or your holdco hold any publicly listed security at a material unrealised gain, and whether you intend charitable giving before 31 December. If both are true, the gift should move broker-to-broker in kind. Confirm the sequence with your accountant before anything is sold.

3 · Selling through an interposed holding company

Transfers — but it is already your default
0
2

Every analysed seller today sold through something: CMXI B.V. for Schuurman, Consensys Software Inc. for Lubin, a revocable trust for Pohlad, a Family Trust and a GRAT for Mecklenburg, a 1988 revocable living trust for Duffield. Cohen alone held direct.

For an Ontario private company the structural question is not whether to have a holdco — that is ordinary — but whether your operating company’s shares will qualify as QSBC shares when you eventually sell them. The lifetime capital gains exemption is $1,275,000 per person for 2026, indexed, up from $1,250,000 in 2025. It is most often lost not by failing the 24-month holding test but by the 50% and 90% active-business asset tests, which get broken quietly by retained cash and passive investments accumulating inside the opco. The failure is silent until the moment you need the exemption.

FIRST STEP THIS WEEK · Ask your accountant for a current QSBC purification read: specifically what proportion of opco assets are non-active today, and what it would take to move surplus cash up to a holdco without tripping the 24-month test.

4 · Pre-announced blocks and plan discipline (Schuurman, Duffield, Mecklenburg)

Does not transfer
0
3

Rule 10b5-1, Form 144 and blackout windows exist because there is a public market and an information asymmetry to police. A private Ontario company has neither, and importing the machinery would be theatre. The transferable residue is weaker and more general than it sounds on the page: deciding the size and timing of a disposal in writing, in advance, before you have a view on price — which is worth something at a shareholder-agreement level, and nothing at a securities-law level.

5 · Threshold-driven rather than price-driven selling (Schuurman)

Partially transfers
0
3

If the tax reading of the Elastic sales is right, the general lesson is the useful part: large disposals are often triggered by an ownership percentage or a date, not by a price. Canadian structures have the same kind of hard edges, and they are the ones worth knowing in advance rather than discovering:

  • The QSBC active-asset tests above, which are measured at a moment and over 24 months.
  • The 21-year deemed disposition for most family trusts — a date certain, set when the trust was created.
  • The associated-corporation rules, which make a group share a single small business deduction.
FIRST STEP THIS WEEK · Establish whether any date-certain threshold applies to your current structure inside the next 24 months, and write the dates down. If a family trust exists, its 21-year date is the single most important number in the structure.

This edition is information, not financial, tax, accounting or legal advice, and nothing here is a recommendation to buy, sell or hold any security. Tax provisions are described in general terms and their application depends entirely on facts not stated here. Anything structural or tax-related should be confirmed with a licensed Canadian accountant and, for trusts, freezes and share reorganisations, a tax lawyer, before any step is taken.

§ 06

Sources

primary documents first
  1. SEC Form 4 — Robert C. Pohlad, PepsiCo, Inc.Transaction 14 Oct 2026 · signed 16 Oct 2026 · retrieved 19 Oct 2026
  2. “PepsiCo Shares Climb 2.4% Following Insider Buying Activity” — Defense World17 Oct 2026
  3. “Independent Director of PepsiCo Picks Up 74% More Stock” — Simply Wall StOctober 2026
  4. SEC Form 4 (EDGAR XML) — Steven Schuurman / CMXI B.V., Elastic N.V.Transaction 12 Oct 2026 · filed 13 Oct 2026
  5. SEC Form 4 — Steven Schuurman / CMXI B.V., Elastic N.V.Transaction 5 Oct 2026 · filed 6 Oct 2026
  6. SEC Form 144 — CMXI B.V. / Elastic N.V., 2,000,000 shares via Goldman SachsFiled 13 Oct 2026
  7. “Steven Schuurman Implements A Sell Strategy” — BenzingaOctober 2026
  8. “Scheme for participations falling below the 5% threshold” (section 13(16) Vpb 1969) — Meijburg & CoRetrieved 19 Oct 2026 · confirms the scheme and the article; does not state the term
  9. Netherlands — Corporate income determination (participation exemption, 5% threshold) — PwC Worldwide Tax SummariesRetrieved 19 Oct 2026
  10. SEC Form 4 — Joseph Michael Lubin / Consensys Software Inc., SharpLink, Inc.Transaction 8 Oct 2026 · signed 13 Oct 2026
  11. “Sharplink director Joseph Lubin’s entity sells $47.6m in stock” — Investing.comOctober 2026
  12. “SharpLink Gaming Stock Price Down 7.3% on Insider Selling” — Defense World15 Oct 2026
  13. “CoreWeave’s Founders Have Sold $3.3 Billion of Stock Since Its IPO” — 24/7 Wall St.17 Oct 2026
  14. “CoreWeave Founders Sold $3.3 Billion in Stock Since IPO” — Seoul Economic Daily16 Oct 2026
  15. “Secret documents reveal donors behind $330 million” — The Washington Post, via The Spokesman-Review15 Oct 2026 · single source · leaked internal spreadsheet
  16. SEC Form 4 (EDGAR XML) — Ryan Cohen, GameStop Corp.Transaction 13 Oct 2026 · accession 0000921895-26-002784
  17. “Ryan Cohen Spends $97 Million Buying Company Shares” — Foreign Policy Journal14 Oct 2026
  18. SEC Form 4 — Gabriel M.I. Mecklenburg, Hinge Health, Inc.Transaction 12 Oct 2026 · 10b5-1 plan adopted 12 Jun 2026
  19. SEC Form 4 — David A. Duffield, Workday, Inc.Transaction 9 Oct 2026 · 10b5-1 plan adopted 2 Dec 2025
  20. Top Insider Sales of the Week — OpenInsiderRetrieved 19 Oct 2026 · source of the ΔOwn figures in §04
  21. Top Insider Purchases of the Week — OpenInsiderRetrieved 19 Oct 2026
  22. Lifetime capital gains exemption 2026 — $1.275M on QSBC shares — Money.caRetrieved 19 Oct 2026
  23. Gift of publicly traded securities by a corporation — inclusion rate and CDA treatment — Abundance CanadaRetrieved 19 Oct 2026
  24. Donating appreciated publicly traded securities to charity in Canada — TaxPageRetrieved 19 Oct 2026