Toby Neugebauer sells 7.66m Fermi shares into a 25% two-day collapse — while demanding the board sell the company at full market value
| Date | Shares | Wtd avg | Range | Proceeds |
|---|---|---|---|---|
| Mon 3 Nov 2026 | 2,563,800 | $2.0391 | $2.000–$2.085 | ≈$5.23m |
| Tue 4 Nov 2026 | 94,000 | $1.9948 | $1.990–$2.005 | ≈$0.19m |
| Wed 5 Nov 2026 | 5,000,000 | $1.5206 | $1.440–$1.590 | ≈$7.60m |
| Three sessions | 7,657,800 | $1.700 | $1.440–$2.085 | ≈$13.02m |
Selling entities on the 3 November tranche: the Melissa A. Neugebauer 2020 Trust and Vicksburg Investments Management LLC. He is trustee of the one and managing member of the other, and disclaims beneficial ownership except to the extent of his pecuniary interest. On Friday 6 and into Thursday 7 November the stock fell about 15.9% from a $1.82 close to $1.53, low $1.44, on 41.4m shares against a 16.2m average — roughly 155% above normal volume. Market capitalisation around $1.01bn.
The backdrop is a live control fight. His DFAN14A campaign seeks to replace the board and run a "75-day plan" that he says must include "a full-market-value sale or strategic partnership"; he has solicited agent designations on a GREEN card to call a special meeting, proposed a slate alongside his appointed directors Miles Everson and Larry Kellerman, and has obtained expedited discovery in the related litigation. Separately, plaintiffs' firms have announced a securities class action over an alleged $150m anchor-tenant exit.
Forms 4, Fermi Inc., reporting person Neugebauer Toby R., transactions 3, 4 and 5 Nov 2026, filed 5 Nov 2026 · Schedule 13G dated 14 Nov 2025: Toby R. Neugebauer 139,016,035 sh; Vicksburg Investments Management LLC 44,656,376 sh; Melissa A. Neugebauer 2020 Trust 94,359,659 sh; combined 278,032,070 · DFAN14A, Fermi Inc., filed 20 May 2026 and subsequent · SEC no-action correspondence, Vicksburg / Fermi Rule 14a-8, 18 Sep 2026
A disgruntled co-founder finally capitulates. He has fought the board all year, the anchor-tenant story has blown up into a securities class action, the stock is at its 52-week low, and now he is selling. Read the $13m as his verdict on Fermi's value — the insider who knows the asset best is monetising it at any price, and the 16% drop on heavy volume is the market agreeing with him.
These sales contain no view on Fermi at all. They are a cash raise against a deadline, and the trade pattern proves it: he nearly doubled his size on the single worst day, at the bottom of the day's range, while publicly demanding a full-market-value sale of the same company. Nobody who wants a premium exit prints five million shares at the 52-week low in front of prospective buyers unless the money is needed before the campaign resolves.
- The escalation profile — this is the crux. 3 Nov: 2,563,800 shares at $2.0391. 5 Nov: 5,000,000 shares at $1.5206. Size +95%; weighted price −25.4%. A seller acting on a view is price-elastic: as the price falls below where he thought it should trade, he sells less, or waits. A seller acting on a need is quantity-inelastic: the share count is set by the cash required, and a falling price forces him to sell more, not less. The observed pattern is the second one and is inconsistent with the first. This fact does not merely fit the contrarian read — the consensus read actively predicts the opposite.
- The size is wrong for an exit and right for a raise. 7,657,800 shares against the 278,032,070 reported across the three filers in the November 2025 13G is about 2.75%. He retained roughly 97% of a position he is simultaneously arguing is undervalued. Capitulation does not stop at 2.75%.
- The timing cost is enormous and he paid it anyway. An activist soliciting other shareholders' support hands the incumbent board its single best rebuttal the moment he sells — our largest holder is dumping stock. A campaigner with any discretion over timing eats that cost only if he cannot avoid it. The 7 November volume spike at 155% of average is the cost landing.
- His own base rate. His prior disclosed Form 4 sale was reported at roughly $1.28m. Three sessions in November ran about 10× that figure. This is not a cadence; it is a discontinuity.
(A) Campaign funding. Inference Dissident proxy contests are funded personally and commonly run $5–15m through a special meeting, a slate and litigation. $13.02m raised in three sessions sits inside that band. Discriminator: the definitive proxy materials must disclose estimated solicitation expenses and who bears them. A figure in that range, borne personally, makes (A) the leading explanation; a figure in the low single-digit millions kills it.
(B) Collateral or margin. A trust and an LLC are the classic vehicles for pledged stock, and a lender-driven liquidation reproduces the escalate-into-weakness pattern exactly. Discriminator: a Form 4 footnote or 13D/A disclosing pledged shares, or a broker appearing as transferee. I found none, which is evidence against (B) but not conclusive — pledges by a non-officer holder are not always surfaced on a Form 4.
(C) Genuine capitulation. The ordinary reading. Discriminator: it predicts continued selling at lower prices and a quiet stand-down of the campaign. If the next filing presses the special meeting while the Form 4s go silent, (C) is dead.
I am not resolving this. What the evidence does establish is that the price of these sales carries no information about Fermi's value, because the seller was not optimising for price.
A Form 4 footnote showing these sales ran under a Rule 10b5-1 plan adopted before the decline. That would make the escalation mechanical rather than deadline-driven, and my "cash on a clock" framing would be wrong about the agent even if right that the price is uninformative. More decisively: definitive proxy materials disclosing solicitation costs borne by a third party or in the low single-digit millions would kill hypothesis (A) outright. And if he sells another large tranche at $1.20 next week with no deadline in sight, the simple capitulation reading wins and I was over-reading a three-day tape.
Prior disclosed sale ≈$1.28m; this week ≈$13.02m. Combined family stake 278,032,070 shares at the November 2025 13G.
Measurement warning: the remaining-stake figures circulating for these filings are internally inconsistent — one account reports 84,813,010 shares held after the 5 November sale while also reporting 42,077,276 after the 3 November sale, which cannot both be right alongside a 7.66m total disposal. These are per-filer and per-entity lines being quoted as though they were one position. Treat any "his stake is now X" number from this week as unreliable until read off the filing itself. The transaction data — dates, counts, weighted prices — is consistent across three independent reports and is what the argument above rests on.