It took off early morning
Before most of the market was awake on Thursday, August 13, $FRGT started moving. Eight minutes later it printed $5.00, more than double Wednesday's close of $2.42.
By 10am it traded at $1.93, below where it started.
No press release. No SEC filing since July 15. Nothing on the wires.
It started with about $70,000
The three minutes before it moved were nearly dead: 2,752 shares, then 1,322, then 3,138. Then one minute traded 23,416 shares and lifted the price 12%. The minute after that traded 206,802 and took it to $4.37.
That first real minute was roughly $70,000 of stock. Freight Technologies last reported 2,860,626 ordinary shares outstanding, so $70,000 bought close to 1% of the entire company in sixty seconds, and that was enough to set off everything that followed.
The whole company traded four times over
Across the day, more than four times every share in existence changed hands. Our scanner logged 86x relative volume in the ignition minute. Almost all of it came before the market opened:
| Time (ET) | Range | Shares |
|---|---|---|
| Early morning | $2.71 → $5.00 | 7,292,462 |
| Pre-market, later | 3.37 → 2.84 | 1,778,887 |
| Into the bell | 2.80 → 3.16 | 1,814,496 |
| 9:30am open | 2.96 → 2.31 | 434,947 |
| 10:00am | 2.29 → 1.85 | 295,360 |
By the time most people could trade it, the move was already finished.
One caveat on that number, and it cuts in an uncomfortable direction. The share count is a moving target, and it moves one way:
| Date | Ordinary shares outstanding |
|---|---|
| December 31, 2024 | 109,255 |
| December 31, 2025 | 1,660,627 |
| April 20, 2026 | 2,860,626 |
That is a 15-fold increase in 2025, then another 72% in under four months, all after five reverse splits that were supposed to shrink the count. April 20 is the freshest figure the company has published, and quarterly reporting stopped in September 2025, so the real number today is higher than 2,860,626 and nobody outside the company knows by how much.
Our own bot bought this at $4.05 and stopped out 17% lower. It is being tuned to size down on early-morning moves in floats this thin.
Why did it go up?
There was no news, so the answer is mechanical rather than fundamental. In the early hours the only participants are algorithms and a small number of very active traders. The order book is thin enough that a modest amount of money moves the price double digits, and a double-digit move is exactly what automated momentum scanners are built to detect. One buyer wakes the scanners, the scanners wake everyone else, and the move becomes its own reason to buy it.
That is the easy half. The harder and more useful question is why a move up on this stock so reliably becomes a move down, and why it does not come back.
The answer is written down in the filings.
The math that drives everything
On July 14 the company sold 1,200,000 Series C preferred shares for $1.2 million. One sentence from the filing does all the work:
the initial Conversion Price will be equal to the lower of (A) $4.752, and (B) the lowest daily volume-weighted average price of the Ordinary Shares in the seven (7) consecutive trading day period immediately preceding the date of the conversion; in each case subject to the floor price, $0.7764
Simply put: convertible preferred
An investor hands over cash and receives a special share that can be swapped later into ordinary shares, the kind you buy. The only question that matters is: swapped at what price?
The conversion price is the lowest daily average price over the previous seven trading days.
### The incentive runs backwards
At $2.00, a million dollars converts into 500,000 shares. At $1.00, the same million converts into 1,000,000 shares. Every dollar the stock loses hands the holder more of the company for the same money. That is not a claim about anyone's behaviour. It is a description of the contract.
Three of these are outstanding, all from the past eight months:
| Investor | Preferred shares | Cash in | Conversion floor |
|---|---|---|---|
| DIP SPV I, for the JAK Solar purchase (Dec 2025) | 5,500,000 | $5,500,000 | $1.68 |
| Institutional investor (Mar 2026) | 1,000,000 | $1,000,000 | $1.102 |
| Institutional investor (Jul 2026) | 1,200,000 | $1,200,000 | $0.7764 |
| Total | 7,700,000 | $7,700,000 |
What $7.7 million turns into
This is the number that reframes the whole stock, and you do not have to model it, because the company states it outright in its own risk factors:
As of December 31, 2025, we had 1,660,627 ordinary shares outstanding as public float and preferred shares and warrants convertible into up to 32,689,825 additional ordinary shares.
Thirty-two point seven million shares of dilution sitting behind 1.66 million. Nearly twenty times the entire company, disclosed by the issuer, in a document anyone can download.
The Series C tranche this piece opened with is only one floor of that building. Run its $7.7 million through the conversion formula and it alone produces this:
The preferred is worth more shares than the entire company has. Not a fraction of it. Multiples of it. And every step down the price takes makes that multiple larger.
The newest floor, $0.7764, sits below the $1.00 minimum bid Nasdaq requires for a listing. The most recent financing is floored beneath the price at which the company begins failing its own listing rule.
The Series C is not the whole stack
It is worth being precise about how much paper is actually outstanding, because the $7.7 million figure understates it badly. The FY2025 balance sheet lists four separate series of preferred, all of them still outstanding at December 31, 2025:
| Series | Outstanding Dec 31, 2025 | A year earlier |
|---|---|---|
| Series A | 3,354,309 | 1,815,438 |
| Series B | 13,407,566 | 1,262,074 |
| Series C | 5,500,000 | 0 |
| Series seed | 7,020 | 7,020 |
| Total preferred | 22,268,895 |
Series B alone grew more than tenfold in a single year.
And there is debt on top of it. On November 19, 2025 the company sold senior convertible notes with $1,000,000 of face value for $900,000 in cash, a 10% original issue discount, meaning the buyer was above water the moment the ink dried. They accrue 10% interest payable in shares, convert at $0.90 (roughly $4.50 after December's 1-for-5 reverse split) with a floor of $0.14, and carry a 4.99% ownership blocker.
There is also a pre-funded warrant with a nominal exercise price and a five-year term, issued for services rather than cash. A pre-funded warrant struck at nearly nothing is not really a warrant. It is stock waiting for a form to be signed.
One holder already did the whole round trip
This is the part that turns the mechanic from theory into record.
Fetch Compute, Inc. disclosed 648,057 ordinary shares, 9.9% of the company, in a Schedule 13G filed November 26, 2025. That 9.9% is not a coincidence; it is the ownership blocker, held right at the ceiling.
On June 1, 2026 Fetch Compute filed an amendment reporting 0 shares. 0.0%.
Converted to the cap, sold, and gone, inside about six months. The filings do not say at what prices, and we are not claiming they do. But the round trip itself is on the public record, and it is exactly the shape the contract rewards.
Who actually controls this company
The FY2025 beneficial ownership table, calculated on 2,860,626 shares as of April 20, 2026, is the single most important page in the filing:
| Holder | Shares | % of class |
|---|---|---|
| Freight Opportunities, LLC | 27,551,582 | 73.1% |
| DIP SPV I, L.P. | 5,982,813 | 15.9% |
| All directors and officers | roughly 1,505 | about 0.05% |
Those top two are not separate parties. The filing's own footnote says Freight Opportunities LLC is wholly owned by ATW Opportunities Master Fund, L.P., with voting and investment decisions made by Antonio Ruiz-Gimenez and Kerry Propper at 1 Pennsylvania Plaza, New York. That is the same firm and the same two people behind DIP SPV I.
Together, on an as-converted basis, that is 33,534,395 shares, about 89% of the company.
The people running it hold about five hundredths of one percent. The CEO's own line in that table reads 42 shares.
### What this means in one sentence
When you buy $FRGT you are not buying a slice of a freight company. You are buying a slice of the roughly 11% that one structured-finance firm does not already have the right to own.
The $20 million memecoin facility
One more item, because it is in the same filing and it says something about how capital gets used here.
On April 29, 2025 the company established a $20 million convertible note facility "exclusively for the purchase of Official Trump tokens." $1.5 million was drawn, and on May 27, 2025 it converted into 387,305 Series A4 preferred shares.
A company with $346,718 of cash and a going concern warning arranged a twenty million dollar facility to buy a memecoin. That is not an interpretation. It is the sentence in the 20-F.
One caveat on the table: some of this has already converted, so the remaining Series C balance is smaller than $7.7 million. The tranche-by-tranche detail is not broken out, but the effect is visible in the share count, which grew 72% between December 31, 2025 and April 20, 2026.
The cycle
Here is the part worth understanding, because it is not a one-time event. It is a loop, and every turn of it leaves the shareholder with less.
Convert low. The holder swaps preferred into ordinary shares at the lowest seven-day average price. Cheaper stock means more shares for the same money.
The share count stays small. This is the piece people miss. Five reverse splits in under three years have repeatedly crushed the share count back down. A company with 1.66 million shares is extraordinarily cheap to move: as Thursday showed, about $70,000 was enough. A thin float is not just a risk on the way down. It is what makes the spikes possible in the first place.
Sell into strength. Any rally, from any cause, is the exit. It does not have to be engineered to be useful. It just has to happen.
The price comes back down. Selling plus a fading move returns the stock to where it was, or lower, which resets the seven-day average lower for the next conversion.
Repeat. The 9.99% blocker in the filings guarantees it. The position legally cannot be exited in one go, so it must be converted and sold in rounds.
Each loop needs the price low to create the shares and high to sell them, and the structure gets both, because those two things happen at different moments. Nobody has to force it. A stock this thin generates spikes on its own.
The reverse splits are the receipt
Simply put: a reverse split
The company swaps your ten shares for one worth ten times as much. Your stake is unchanged; nothing is created or destroyed. Companies do it for one reason above all: the price has fallen so low the exchange threatens to delist them.
March 2023 (1-for-10), February 2024 (1-for-10), September 2024 (1-for-25), May 2025 (1-for-4), December 2025 (1-for-5). Five in under three years, each following the price back down to the same place.
Multiplied together, one share today equals 50,000 shares from early 2023.
Here is the most human way to see what that did. In their March 2026 filings, the officers and directors report what they personally own:
| Ordinary shares held | |
|---|---|
| Javier Selgas, CEO | 1 |
| Marc Urbach, director | 1 |
| Nicholas Adler, director | 1 |
| Leilei Nie, director | 600 |
| Jose Gonzalez Leon, director | 0 |
Those figures are from the March 2026 Form 3s. The FY2025 ownership table, struck a month later on 2,860,626 shares, is no kinder: Javier Selgas 42 shares, Donald Quinby 800, Leilei Nie 600, and single digits for most of the rest. Board and management together hold roughly 1,505 shares of 2,860,626, about five hundredths of one percent. Whatever else that says, almost nobody steering this company loses meaningfully if the share price keeps falling.
How traders are actually trading this
None of the above is secret. It is in filings anyone can pull, and the people who trade $FRGT for a living have read them. That shared knowledge is itself a force on the price, and it explains the shape of Thursday better than any theory about who bought at dawn.
The number everyone is watching
A conversion on Thursday would have priced at the lowest daily average of the previous seven sessions:
| Session | Daily VWAP |
|---|---|
| Aug 5 | $3.7309 |
| Aug 6 | $2.9026 |
| Aug 7 | $2.7841 |
| Aug 10 | $2.7006 |
| Aug 11 | $2.5214 |
| Aug 12 | $2.6697 |
$2.5214. That is the effective cost of newly created shares right now. The stock printed $5.00, which is 98% above it.
### Why their break-even IS the conversion price
They paid $1.00 per preferred share, for $1.00 of stated value. On conversion, that $1.00 buys `$1.00 ÷ conversion price` ordinary shares. At $2.5214 that is 0.3966 shares, so their cost works out to exactly $2.5214 per ordinary share.
Read that again, because it is the whole thing. Their break-even is not a price they locked in months ago. It is whatever the conversion price is on the day they convert. An ordinary investor who bought in December at $3.47 is deeply underwater. These holders are never underwater on an old entry, because they do not have one.
There is one exception worth noting, and it runs the other way: the November 2025 notes were bought at $0.90 on the dollar, so that holder's break-even sits 10% below face before a single share moves.
And note what that does to behaviour. A holder at $2.5214 is profitable at anything above $2.5214. They are not waiting for a target, because there is no target. Any green is green. That is why supply shows up on every rally instead of at some specific level, and why the rally does not need to reach a round number to be sold.
The break-even follows the price down
This is the mechanic that answers "why doesn't it recover," and it is the one most people miss.
The conversion price is not fixed. It re-prices to the lowest average of the last seven trading days, so it ratchets down as the stock does. A seller whose cost basis chases the price lower can keep selling profitably at levels that would have been a loss a week earlier.
And note which direction that ratchet turns. The formula takes the lowest of the seven days, not the average. Thursday's spike pushed that day's VWAP to $3.7184, the highest reading in the whole window, and it changed nothing at all, because a maximum cannot move a minimum. A rally cannot raise their cost. Only a new low can lower it.
Look at where that stands today. The stock trades near $2.05, roughly 19% below the $2.5214 conversion price. Right now converting is unattractive. But if the stock simply sits here, the high VWAPs from early August roll out of the seven-day window and the conversion price resets toward $2.05 on its own, with nobody doing anything.
Time works for that side of the trade and against yours. Patience costs them nothing.
Why the chase reverses so fast
Everyone in the name knows the above, and the result is a crowd with no intention of holding.
The buying is a chase, not an investment. Nobody in the pre-market believes the freight business re-rated overnight, because nothing happened. They are buying because it is moving, and they know the exit is crowded, so the plan is to be early to leave. The moment it stalls, that plan executes all at once. There is no bid underneath from people who wanted to own it, because those people read the same filings and stayed out.
The tape shows exactly this. The peak came eight minutes in. The rest of the day was lower highs on shrinking participation:
Volume fell 98% from the opening hour to the afternoon. The chase is loud and the exit is silent. Once the sellers are done, the stock does not crash further; it just goes quiet at a lower price and sits there until the next spike.
That is the honest shape of this trade: fast up, faster down, then dead. It does not recover on its own. It requires new buying large enough to start the pattern over.
Why this is hard to hold
Set the day trading aside. If you bought $FRGT to hold, four things work against you regardless of what the business does.
1. Your slice shrinks whether or not you do anything. Every conversion creates new shares. More shares, same company, smaller piece for you.
2. You cannot see how bad it is right now. The company stopped filing quarterly reports in September 2025. The freshest share count it has published is 2,860,626, as of April 20, 2026, and that was already 72% above where the year ended. Four months of conversions since then are simply not disclosed.
3. You cannot see whether insiders are selling. That reporting stopped at the same time.
4. The auditor has doubts the company survives the year.
Simply put: a going concern warning
The outside auditor formally stating there is substantial doubt the company lasts twelve months. The most serious thing an audit report can say short of refusing to sign.
The two numbers behind it: $346,718 of cash at year end against a $7,901,777 net loss for the year. Weeks of runway, not quarters. That pressure explains everything above. This company is not choosing its financings from a menu.
Who holds the preferred?
The Schedule 13G names them. DIP SPV I, L.P., a British Virgin Islands entity, is wholly owned by ATW Fund I, L.P., managed by ATW Partners LLC of 1 Pennsylvania Plaza, New York. The control persons are Kerry Propper and Antonio Ruiz-Gimenez. ATW is a structured-finance firm; this style of small-cap convertible is its business.
Note what that means for the JAK Solar deal: ATW's entity sold the company an asset and was paid in the very paper that dilutes it, 5,500,000 preferred shares with a $5,500,000 stated value.
The same filing contains the mechanic that forces the loop to repeat:
The Holding Company's Preferred Stock is subject to a blocker which prevents the Holding Company from exercising and/or converting its Preferred Stock into Shares to the extent that... the Holding Company, together with its affiliates would beneficially own in excess of 9.99% of the Shares outstanding
That 9.99% cap is why a $5.5 million position reports only 184,309 shares. It also means the position cannot be exited in one go. Convert to the cap, sell down, convert again, for as long as it takes.
What else could explain it
A stock that doubles before dawn on no news invites the obvious suspicion, and we went looking for ordinary explanations instead. Several of them hold up well, and the honest reading is that none of what follows requires anyone to have arranged anything.
Five other stocks did the same thing that morning. This is the strongest of them, and we went looking for it specifically. Of 216 names with real pre-market volume on Thursday, six ran more than 50% before the bell:
XHG ran 1,783% and has fallen further than $FRGT since. Five of the six collapsed off their pre-market high, and they share nothing but size. Whatever moved $FRGT was moving a lot of small caps at once, which is a market condition, not a plan.
The timing does not work. Conversions do not settle in minutes. Nobody watches a price tick at dawn and converts into it. A holder selling that morning would have needed shares already in hand.
Nobody can see who traded. No public record attributes pre-market trades to participants. Only regulators and exchanges see order-level data. We cannot, and neither can anyone else writing about this stock.
We can also rule out two popular explanations outright. It was not a short squeeze: 25,737 shares were reported short at the last settlement, two tenths of one percent of Thursday's volume, with nothing there to squeeze. It was not a social media pump: $FRGT never surfaced on our tracker of the thirty most-discussed tickers in the days around the move.
What is left is narrower and more useful than a theory about any one morning: the contract rewards precisely the price pattern that keeps occurring. That is true whether or not a single person ever acts on it, and it is the reason the pattern outlives any individual spike.
What is genuinely working
None of the above makes the business fake. It is not, and this is the part the tape ignores.
The operating loss has improved three years running, while revenue slipped only 5%. They are cutting cost faster than they are losing business, which is the correct order for a turnaround.
| Year | Operating loss |
|---|---|
| 2023 | -$8,261,198 |
| 2024 | -$6,490,251 |
| 2025 | -$5,894,065 |
The revenue is real. $13,062,534 in 2025 from actual freight brokerage. Trucks move, invoices get paid. This is not a concept with a pipeline of letters of intent.
The balance sheet is not underwater. $12.59M of assets against $7.26M of liabilities, leaving $5.33M of equity.
A secured lender is gone. The Capital Foundry facility was repaid in full and terminated. Secured lenders rank ahead of shareholders when things go wrong, so removing one genuinely improves everyone below.
The product is being adopted. Fleet Rocket reached 92 integrated GPS providers, up 70% since October 2025, and added native CFDI stamping and invoice management. CFDI is Mexico's mandatory electronic invoicing standard, which makes that a specific, defensible feature in exactly the US-Mexico cross-border lane this company runs.
The path out
There is a way through, and it deserves stating properly rather than waving away. Five things, in order of leverage.
1. Retire the preferred rather than convert it. Buying it back or refinancing it with non-dilutive money breaks the loop in one move. Nothing else on this list matters as much, and it is entirely within management's control if they can find the money.
2. Stop the revenue slide. $25.9M, $17.1M, $13.7M, $13.1M. Cost cuts have a floor; growth does not. A single flat year changes the whole story.
3. Raise once, cleanly. A straight loan, a strategic partner, a revenue share, anything without a floating conversion price. The problem was never the amount raised. It is the structure.
4. Turn the AI work into margin. They report 15x domestic and 5x cross-border efficiency gains. If that lands as software-like gross margin on the same revenue, this becomes a genuinely different company. The technology exists and is being adopted; it has to reach the P&L.
5. Hold $1.00. Everything above needs a listing to matter.
The next checkpoint is August 17, when the company is expected to report after the close. One caveat: as a foreign private issuer it is no longer required to report quarterly, so treat the date as expected rather than guaranteed. Three numbers will tell you which way this is going: cash on hand, whether revenue stopped falling, and the share count.
If the share count has barely moved and cash is up on a non-convertible raise, the story genuinely changed. If the share count jumped and there is another preferred deal in the notes, it did not.
The bottom line
Both halves of this are true at once. The operating business is measurably improving. The capital structure is measurably more dangerous. They are not contradictions; they are the same story, because the second is what pays for the first.
If you trade $FRGT, be clear about what you are trading. Not a freight company: a very small number of shares against a financing structure worth more shares than the whole float, whose holders do better the lower it goes, run by a board that owns almost none of it.
Thursday's spike was ordinary. Four other stocks did it too. What is not ordinary is what happens to $FRGT afterward, over and over, and that part is written down in filings anyone can read.
References
SEC filings (Freight Technologies, Inc., CIK 0001687542)
- Form 6-K, July 15, 2026: the $1.2M Series C, the conversion formula, the side-letter floors
- Form 20-F, FY2025, filed May 14, 2026: going concern, all five reverse splits, Nasdaq history, the foreign-issuer change, JAK Solar risk factors
- Schedule 13G, January 2026: DIP SPV I, ATW Fund I, ATW Partners LLC, and the 9.99% conversion blocker
- Form 6-K, November 20, 2025: the $1,000,000 senior convertible notes sold for $900,000, the Series A4 cancellation, and the pre-funded warrant to Fetch Compute
- Schedule 13G, November 26, 2025: Fetch Compute reports 648,057 shares, 9.9%
- Schedule 13G/A, June 1, 2026: Fetch Compute reports zero
- Complete EDGAR filing history
Insider share counts are from the Form 3 filings of March 19, 2026 and the beneficial ownership table in the FY2025 Form 20-F, calculated on 2,860,626 shares as of April 20, 2026. The 32,689,825 dilution figure, the four preferred series, the $20 million token facility and the Freight Opportunities / ATW footnote are all from that same Form 20-F. Financials are from the company's own XBRL data as filed.
Market data
Intraday prices, volume, daily VWAPs, hourly volumes, relative-volume readings and the pre-market comparison table are from The Desperate Trader platform for August 2026. Short interest is as of the July 31, 2026 settlement. Share counts are as reported: 109,255 at December 31, 2024 and 1,660,627 at December 31, 2025 from the balance sheet, and 2,860,626 at April 20, 2026 from the beneficial ownership table. None reflect conversions after those dates. Share-creation figures apply the stated $7,700,000 to the conversion formula and do not adjust for preferred already converted, so they are an upper bound on that tranche and, given the other three series, a considerable understatement of the total.
This is research and commentary, not investment advice, and not an allegation of wrongdoing by any person or firm named.