I Tracked Every Merlin Aircraft for 180 Days. The Testbeds Just Peaked
Six months of public ADS-B data, no company input. Three certification testbeds, two countries, one synchronized spring peak
Shawarma Capital. May 2026. Every figure in this post is reconstructed from public ADS-B data. No company input, no private feed, no leaked data. This is research synthesis, not investment advice.
Almost everything written about a pre-revenue defense company comes from the company. The press release. The investor deck. The CEO on a podcast. All of it useful, none of it evidence, because a company describing itself is not the same thing as a fact.
This is a different exercise.
For six months I tracked every aircraft I could connect to Merlin Labs using nothing but public ADS-B data. ADS-B is the position signal that aircraft broadcast continuously, several times a second, and that anyone with an antenna or an archive can receive. Seventeen airframes. One hundred and eighty days. Roughly six hundred individual flights, reconstructed one at a time from raw position data. No company input. No NDA. No paid feed. Just the aircraft, telling on themselves.
Here is the finding in one sentence. Merlin's certification fleet flew the busiest stretch of the entire window in March and April 2026, and three separate testbeds, owned by two different Merlin entities, flying in two different countries, all peaked in the same eight weeks.
That is an independent, non-company-sourced read on whether the program is executing. It is the exact signal the thesis in Parts 1 through 8 is priced on, and until now nobody outside the company had measured it. What follows is the full roster, the method, the complete ledger, the testbed identification, and an honest accounting of what the data does and does not prove.
It is long. It is meant to be. Every number is checkable, and the last section tells you how to check it.
Let's go.
I. The One Rule For Reading Flight Data
Before any numbers, the rule that governs all of them.
Flight tempo is direction. It is not a date, and it is not a verdict.
A flight-test program flies hard when it is winning. It also flies hard when it is losing. It flies hard repeating test points a regulator rejected. It flies hard rebuilding a flight-hours case after a setback. Tempo rises for re-tests and for problems exactly as readily as it rises for progress. Position data cannot tell a milestone apart from a do-over.
So this is monitoring intelligence. It is not a trade and it is not a secret. It tells you the testbeds are busy, and it tells you precisely when that changed. It does not tell you the certification will pass. I will hold it that honestly for the entire post, and you should too. Anyone who hands you ADS-B data as a buy signal is selling you something.
II. What I Actually Did
The project had five steps.
One, I assembled the full list of aircraft connected to Merlin across five tiers of relationship, from Merlin-owned to entirely unrelated. Forty-eight airframes entered the net. Seventeen came out the other side as aircraft worth reconstructing in full.
Two, I resolved the ICAO hex code for each aircraft. The hex is the only identifier that matters for tracking, and Section V explains why getting it exactly right is the whole game.
Three, I reconstructed one hundred and eighty days of flight history from the public adsb.lol whole-planet archive, one calendar day at a time, and segmented the raw position traces into discrete flights.
Four, I analyzed the result: monthly tempo, sortie length, intensity, and the synchronized spring ramp.
Five, I cross-checked the one genuinely new public fact the project produced, the identity of the USAF KC-135 autonomy testbed, against the Air Force's own published imagery.
Every number below is reproducible from public sources. Nothing is modeled. Nothing is estimated. Where data is missing, it is reported as missing rather than filled in.
III. The Airframes Merlin Chose To Fly
Before the roster, the hardware, because the choice of airframe is itself a piece of the thesis.
Merlin's autonomy product, the Merlin Pilot, is software that flies an aircraft from takeoff to touchdown. Software does not certify in the abstract. It certifies on a specific airframe, with a specific regulator, on a specific type certificate. So the aircraft Merlin flies are not incidental. They are the certification.
The Cessna 208 Caravan is the civil workhorse of the program. It is a single-engine turboprop, rugged, simple, flown all over the world for cargo and short-haul passenger work. It is the airframe behind Merlin's FAA Supplemental Type Certificate effort and the concurrent New Zealand Civil Aviation Authority path. Five of the seventeen aircraft in this study are Caravans or close Caravan variants. When you see a Caravan flying short repeated sorties, you are watching the civil certification campaign happen.
The C-130J Super Hercules is the defense anchor. It is the four-engine military transport at the center of the $105 million USSOCOM IDIQ. Its autonomy program is pre-first-flight, targeted for late 2026, which means it has no identifiable airframe in this dataset yet. Anyone naming a C-130J tail number today is guessing.
The KC-135 Stratotanker is the aerial-refueling platform behind the Merlin and Sierra Nevada autonomy data-collection program hosted by the 171st Air Refueling Wing. It is a 1950s-design jet tanker, and the program that flew Merlin instrumentation aboard one is the subject of Section XI.
Three airframe classes. One civil, two defense. The civil one is the one you can actually watch certifying in the flight data, and it is where this study has the most to say.
IV. What "Merlin's Fleet" Actually Is
The first hard problem is definitional. "Merlin's aircraft" is not one list. It is five concentric rings, and only the innermost ring is Merlin-owned. Collapse the rings and you get the company badly wrong, which is exactly what the public commentary does.
Ring 1. Merlin Labs owned. Five aircraft.
N208B, "Big Red," a Cessna 208B Grand Caravan, is the primary US FAA certification testbed, based at Quonset State Airport in Rhode Island. N506DB is a Burton Long-EZ experimental, a small R&D platform out of Mojave, California. ZK-MLN is a Cessna 208B Grand Caravan registered to Merlin Labs NZ Ltd, the New Zealand certification testbed. ZK-MLO is a Cessna 208B Super Cargomaster, also Merlin Labs NZ, a testbed that also flies paid charter and freight. ZK-MLP is a Cessna TU206F Turbo Stationair, lightly used, which never broadcast a clean identifier I could lock onto. It is named here for completeness and excluded from the flight history. I will not report numbers I did not measure.
Five airframes. That is the owned fleet.
I flag the count only because the figure that circulates publicly is fifty-five. Fifty-five is a 2021 press figure for the King Air fleet of a different company, Dynamic Aviation, that Merlin once announced an intent to retrofit with autonomy. Those aircraft are Ring 5 below. They are not Merlin's fleet and never were. A company with five owned aircraft and two more under operation is not a smaller version of a fifty-five-aircraft company. It is a different company, and the data should describe the real one.
Ring 2. Operated, not owned. Two aircraft.
VH-WZY is a Cessna 208B Grand Caravan, Australian-registered, operated by Merlin since January 2026. ZK-JMP is a Cessna 208B Supervan, owned by a skydiving operator and flown under lease by Merlin under the callsign MERLIN 1. As the ledger shows, it is the single busiest aircraft in the entire dataset.
Ring 3. Partner testbed. One aircraft.
N437VN is a Scaled Composites Model 437 "Vanguard," the airframe behind Northrop Grumman's Beacon autonomy testbed program. Merlin is one of six integration partners on Beacon. This aircraft is partly Merlin and mostly not, and it is weighted that way throughout.
Ring 4. The military host pool. Nine KC-135 tankers.
The 171st Air Refueling Wing, Pennsylvania Air National Guard, based at Pittsburgh, is the host unit for the Merlin and Sierra Nevada KC-135 autonomy program. Eight 171st KC-135T jets sit in the host pool, plus one separate KC-135R candidate at MacDill. Read Section X before you let this ring into the Merlin story, because it does not belong there.
Ring 5. The Dynamic Aviation King Airs. Thirty-five aircraft.
These belong to Dynamic Aviation Group of Bridgewater, Virginia, not to Merlin. They are the "fifty-five" press figure. There is no public per-tail confirmation of which were ever converted to autonomy or how many. They fly Dynamic's own ISR and survey business. Their activity is noise. I tracked them only to rule them out, and ruled them out. They appear in no number presented here as a Merlin signal.
The universe is five owned, two operated, one partner, nine military host, thirty-five unrelated. That is the shape of the thing before a single flight is counted.
V. The Unit Of Truth Is The ICAO Hex
You cannot track an aircraft by its tail number, because the tail number is not broadcast. What an aircraft transmits is its ICAO hex, a 24-bit address baked into the transponder. Every flight-tracking service in the world keys on that hex. Track the wrong hex and you have silently tracked the wrong aircraft for six months and learned nothing, and you would never know.
So the hex codes had to be exact. Sixteen of the seventeen I read straight from the raw archive, matched against known registration pairs. The seventeenth, the Northrop Model 437, never showed a clean registration-to-hex pair anywhere in the raw data.
For that one I computed the hex. The FAA assigns hex codes to US-registered aircraft by a fixed, reversible algorithm, so an N-number converts to its hex deterministically. I implemented that algorithm, and before trusting it I ran it against nine aircraft whose hex codes I had already confirmed from raw data. The first pass failed. A per-digit block offset was wrong, set to plus one when it should have been plus six hundred and one. After the fix, the algorithm reproduced all nine confirmed pairs exactly. Only then did I let it produce the Model 437 code, a53f07. That is why the Model 437 line is allowed to exist in the ledger. It was validated nine times before it was used once.
That is the standard the rest of this analysis is built to. If a number cannot survive that level of check, it is not in here.
VI. The Method: 400 Gigabytes To Surface A Few Megabytes
The data source is the adsb.lol globe_history archive. It is public and openly licensed. It is also brutally inconvenient, and the inconvenience is the reason almost nobody does this.
There is no per-aircraft endpoint. The archive does not let you ask for one plane. It publishes one whole-planet file per day, every aircraft on Earth, roughly two gigabytes compressed and split into four parts. To get one aircraft's track for one day, you download the entire planet, extract the trace files for the seventeen hex codes that matter, and delete the rest.
Then you do that one hundred and seventy-four more times.
On the order of four hundred gigabytes of archive was moved to surface a few megabytes of actual signal. Peak disk footprint stayed near three gigabytes, because each day's planet was deleted before the next was pulled. The run was driven in fifteen-day chunks, with results committed after every chunk, so an interrupted machine never lost more than a few days of progress.
Turning a raw position trace into discrete flights needs one definition: when does a sortie end. The rule here is ground time. More than twenty-five minutes stationary closes one sortie and opens the next. Any airborne segment under two minutes is discarded as noise, a stray ping rather than a flight. Airborne hours are the plain sum of segment durations. Six days in the window had no archive published at all, and those six are reported as gaps, never as zeros, because inventing a zero on a day with no data would quietly corrupt every monthly count downstream.
How the 180-day dataset was built. 181 days in window, 175 with data, 6 gaps, 17 aircraft, 400 GB processed, 597 sorties, 600 airborne hours, 9 of 9 hex pairs validated.
VII. The Finding: The Spring Ramp
Merlin's certification flying was not steady. It ramped, hard, into spring 2026. And it is not one aircraft doing it. Three separate testbeds, owned by two different Merlin entities, flying in two different countries, all posted their heaviest months of the entire window in March and April 2026.
ZK-MLN, the New Zealand certification Caravan, ran 4, 0, 0, 4, 11, 14, 8 sorties from November to May. Dead through December and January. February wakes it. March and April are its two busiest months of the window.
ZK-MLO, the Super Cargomaster, ran 0, 0, 0, 0, 9, 22, 11. Four straight months of zero. Then March takes it from nothing to nine sorties, and April to twenty-two, its single most active month.
N208B, "Big Red," the US certification testbed at Quonset, ran 12, 9, 5, 11, 1, 24, 11. Choppy all winter. Near-grounded in March at a single sortie. Then twenty-four sorties in April, its biggest month by a wide margin.
Three aircraft. Two countries. One eight-week window. Three independent flight-test programs do not all peak in the same eight weeks by accident. Something coordinated the calendar. That is the finding, and it is the kind of result you can only reach with data the company did not hand you.
Three testbeds, one spring ramp. Monthly sorties for N208B, ZK-MLN, and ZK-MLO, with the excluded 171st KC-135 fleet line for contrast. All three testbeds peak in March and April 2026.
VIII. Sortie Length Tells You What Kind Of Flying It Is
A sortie count tells you how often. The average sortie length tells you what they were doing. There is a short group and a long group, and the split is clean.
For a testbed, short sorties mean circuits. Takeoff, pattern, landing, repeat. That is the literal rhythm of certifying an autopilot. You fly the same profile again and again, varying one parameter, logging the result. N208B flew twenty-four April sorties averaging fifty-four minutes each. That is a flight-test schedule, not transport flying.
The long group is transit legs, charter runs, and tanker tracks. Ordinary point-to-point flying. ZK-MLO at seventy-two minutes is the testbed that also earns revenue, which is why it sits between the two groups. The KC-135s span sixty-three to one hundred and three minutes, which is simply what tanker flying looks like.
Intensity points the same way. When these aircraft fly, they tend to fly again the same day. N208B averaged 2.0 flights per active day. VH-WZY 2.5. ZK-JMP 2.9. A program that flies an aircraft two or three times on the days it flies at all is working a schedule, not running errands.
IX. The Full Ledger
No rounding, no summary, no aircraft hidden. Every airframe in the dataset, all 175 days of usable data. The figure below is the complete ledger: sorties, airborne hours, and active days for all seventeen aircraft.
The full ledger. Every aircraft tracked, 175 days of data, November 20 2025 to May 19 2026. ZK-JMP leads with 181 sorties, fleet total 597 sorties and 600 airborne hours.
ZK-JMP is the workhorse. One hundred and eighty-one sorties, almost one hundred and thirty-seven airborne hours, sixty-two active days. It is the busiest aircraft in the fleet by a factor of more than two over the next aircraft. Its behavior in the spring is its own story, in Section XII.
The two zeros are reported as zeros, not quietly dropped. N506DB, the Long-EZ R&D platform, logged no captured flights across the whole window. 63-8020, the MacDill KC-135 candidate, was added to the run late and covers only the tail of the window, and it logged none either. A platform that did not fly is a fact, and facts that argue against an exciting narrative get the same ink here as the ones that support it.
X. The KC-135 Ramp I Am Deliberately Not Counting
Here is the part a weaker version of this post would lead with, and it would be wrong to.
The nine 171st-wing KC-135 tankers also ramped. Steeply.
The KC-135 line climbs from eight sorties a month to seventy-five. It is the most dramatic line in the entire dataset. It is also not a Merlin signal, and folding it into the Merlin story would be the single most misleading thing anyone could do with these numbers.
Tanker wings ramp on their own cycle. Deployments, exercises, currency requirements, seasonal training blocks. A KC-135 wing flying harder in late winter tells you about that wing, not about one autonomy program riding along on a handful of its jets. The two KC-135s actually tied to Merlin by photographic evidence, 59-1460 and 58-0084, sit mid-pack at thirty and nineteen sorties against a host-fleet range of eighteen to thirty. Flight tracking cannot pick the autonomy testbed out of the pack. The data does not have the resolution, and pretending otherwise would be a lie of exactly the kind this post exists to avoid.
So the KC-135 ramp is real, it is dramatic, and it is excluded from the Merlin read on purpose. I am telling you it exists, and I am telling you to ignore it. That discipline is the whole point of doing this with data instead of vibes.
XI. The Military Testbed: 59-1460, "Classic Iron"
If flight tracking cannot identify the KC-135 testbed, how is it identified at all? Photographs. And the evidence is worth walking in full, because this is the one place the project produced a genuinely new public fact.
The aircraft is USAF KC-135T serial 59-1460, nicknamed "Classic Iron," ICAO hex ae0596, a 171st Air Refueling Wing tanker out of Pittsburgh. Three independent threads converge on it.
One, the Air Force's own imagery. The official DVIDS gallery documenting the program, titled "Next-Gen Autonomy: 171st Fuels Future of Self-Flying Tankers," shot by the 171st wing on August 1, 2024, contains six images. Five show the Merlin avionics data-acquisition rig installed inside a KC-135 cabin. The sixth, DVIDS image 8630294, is an exterior ramp shot. Two people carry that same instrumentation cart directly alongside a KC-135 whose nose is painted "1460," with a boarding airstair deployed at that aircraft's forward door. The Air Force left the tail number out of every single caption. The photograph did not.
Two, the registry. "1460" resolves to USAF serial 59-1460, an independently verified 171st-wing KC-135T.
Three, the archive. Forensic analysis of fifteen days of ADS-B archives across the July-to-August 2024 test campaign window confirms 59-1460 operating out of Pittsburgh on four days inside that window.
The precise, defensible claim is this. The Air Force's own published program photo set shows the test instrumentation at the door of KC-135 59-1460. That is an evidence-based identification, not an Air Force-stated confirmation, and the wording matters. A second tail, 58-0084 "Ghost Raiders: Final Contact," appears in a separate Merlin-supplied program photo and is the second KC-135 tied to the program by imagery.
That is as far as honest analysis goes. The program ran a first data-collection flight in late July 2024, flew three sorties, and the instrumentation was removed around August 1, 2024. Sierra Nevada Corporation is the integration partner, Merlin supplies the autonomy, and GE Aerospace joined the effort in September 2025. None of that requires guessing. The C-130J program, by contrast, has no identifiable airframe yet, and that gap is reported as a gap.
XII. Two Movements Worth Watching
Two aircraft did something in the spring the monthly numbers do not explain on their own.
ZK-JMP, the leased freight Caravan and the busiest aircraft in the fleet, flew fifty-two sorties in March and then one in April. The workhorse went dark in the exact month the certification testbeds peaked. One reasonable hypothesis is that the fleet shifted weight off revenue freight flying and onto test flying. The data is consistent with that reading. It does not prove it, and an aircraft can go quiet for maintenance, scheduling, or contract reasons that have nothing to do with the test program.
VH-WZY, the Caravan Merlin took over in January 2026, flew twenty-one sorties in its first month and nine in its second, then went silent for three months. Either it returns to the line or that arrangement stalled. Right now it is an open question, and it is flagged as one rather than guessed at.
XIII. Tying The Flight Data To The Thesis
I have spent twelve sections being careful. This one too.
The Merlin thesis, laid out across Parts 1 through 8, is priced on certification. The FAA Supplemental Type Certificate for autonomy on the Cessna 208 Caravan, on its Stage of Involvement path. The concurrent New Zealand CAA effort, which is why ZK-MLN and ZK-MLO exist as a separate testbed pair on the other side of the world. The $105 million USSOCOM IDIQ on the C-130J, whose preliminary design review cleared on March 5, 2026, with a first flight targeted for late 2026. Every one of those is a milestone that has to be flown to be reached.
For context, here is the program bench the thesis rests on.
What the flight data adds to that bench is independent corroboration that the program is, in fact, executing. Not company-sourced. Not a press release. Three certification testbeds, across two regulatory jurisdictions, flew their heaviest stretch of a 180-day window in March and April 2026, in the short-sortie circuit pattern that matches autonomy certification work.
If Merlin were quietly stalled, that is not the signature you would expect in six months of position broadcasts. You would expect flat or falling tempo and idle certification testbeds. Instead the data shows a synchronized peak across two countries.
That is corroboration. It is not proof. The flight data raises or lowers your confidence that the company is doing the work it says it is doing. It does not settle the bull or bear case, and it does not put a date on anything.
It does bear, mildly, on the kill criteria the thesis is governed by. Part 8 re-checked the seven criteria that would break the position and found zero of them tripped. The first criterion is the C-130J critical design review failing or the program being descoped. A certification fleet flying its hardest stretch of the observable window is weak evidence against a quiet descope. Weak, because tempo cannot tell a milestone from a re-test, but it is evidence, and it points the right way. The second criterion is a named customer issuing a stop-work order. Nothing in six months of flight data looks like a stand-down. The testbeds did the opposite of standing down.
XIV. Read It From Every Side
A single reading of a dataset is how you fool yourself. So here are four, including the ones that hurt.
The bull read. Three independent testbeds, two countries, two operators, all flying their heaviest of the window in the same eight weeks, at sortie lengths that match circuit work. That is a certification program accelerating into its milestones.
The bear read. Tempo is not progress. A program flies hardest when it is fixing problems, repeating failed test points, or rebuilding a flight-hours case a regulator pushed back on. The same March-April spike is fully consistent with a program behind schedule and catching up. The data cannot tell a re-test from a milestone, and an honest reader has to hold that.
The noise read. Six months is one window, not a trend. Winter weather alone suppresses flying from November through January, so part of the spring climb is seasonal thaw rather than program intent. Strip the season out and the true signal is smaller than the raw monthly counts make it look.
The discipline read. The 171st KC-135 ramp is steeper than anything on the Merlin testbeds, and it means nothing here. Putting it in the Merlin story would produce the most dramatic chart in this post, and it would be false. Leaving it out is the point.
What survives all four reads: the certification testbeds are demonstrably busier than at any other point in the window, and the timing across three aircraft is synchronized in a way coincidence does not explain easily. Everything past that sentence is interpretation, and you now hold the same numbers I do.
XV. The Catalyst Clock
The flight data is one input into a thesis that has hard structural dates ahead of it. The Section 7(b) VWAP measurement window runs across August and September 2026. The insider lockup expires September 16, alongside S-1 effectiveness. The conversion-price reset window opens in mid-October. The C-130J first flight is targeted for late 2026.
Note the timing. The next flight-data window, June, lands well before those structural dates. That is the point of watching it. The certification tempo is the cleanest read available before the capital-structure events arrive, and it is the read least contaminated by them. The lockup, the VWAP reset, and the S-1 effectiveness are share-count mechanics. The flight tempo is the operating business. Watching the second one tells you whether the first set is arriving into a company that is executing or a company that has stalled.
XVI. What I Am Watching Next
What turns this from a snapshot into a signal is the next window. June is the first clean test of whether the spring ramp was a step change or a one-off test block.
If ZK-MLN, ZK-MLO, and N208B all hold double-digit sorties through June and July, the spring ramp was a genuine gear change and the program has moved to a higher baseline. If they sink back toward winter levels, March and April were a discrete test block and not a new normal.
Then the two open questions. Does VH-WZY come back from its three-month silence. Does ZK-JMP stay dark, which would confirm a shift of fleet weight from revenue toward test flying, or return to freight.
And the highest-value watch item of all, any new tail. A previously unseen aircraft at Quonset or in New Zealand would mean the fleet is physically growing. Merlin's management has publicly described a strategy of becoming a defense "neoprime" through M&A. A new airframe appearing on ADS-B is the earliest possible read on whether that strategy is moving from talk to metal, and it is exactly what this method is built to surface.
What turns a snapshot into a signal. Combined testbed sorties per month, peaking at 60 in April, with June flagged as the first clean test, and three watch items: VH-WZY, ZK-JMP, and any new tail.
XVII. How To Check My Work
This is the part that separates analysis from assertion. Everything above is reproducible, and here is the recipe.
The aircraft list is the five rings in Section IV. The hex codes are stated next to every aircraft in the ledger. The data source is the adsb.lol globe_history archive, public and openly licensed, one whole-planet file per calendar day. The window is November 20, 2025 to May 19, 2026. The segmentation rule is a twenty-five-minute ground gap to split sorties and a two-minute floor to discard noise. Run that against the archive for those seventeen hex codes and you will reconstruct the same 597 sorties and 600 airborne hours, because nothing here is modeled.
The testbed identification rests on USAF DVIDS image 8630294, which is public domain and viewable by anyone. The registries are the FAA aircraft registry and the New Zealand Civil Aviation Authority register, both public. The company financials, contract values, and capital-structure dates referenced in Sections XIII and XV trace to SEC EDGAR filings under Merlin's filing entity and to the earlier parts of this series.
If a number in this study is wrong, it is wrong in a way you can find, because the inputs are all public and the method is all stated. That is the standard. Disagree with the interpretation all you like. The interpretation is in Section XIV and it is labeled as interpretation. The numbers are not opinion.
XVIII. The Bottom Line
Six months. Seventeen aircraft. Roughly six hundred sorties and six hundred airborne hours, segmented one flight at a time from about four hundred gigabytes of public archive. One computed hex code, validated nine times before it was used once. One military testbed identified from the Air Force's own photograph. One ramp the data shows clearly and one ramp the data shows clearly and that you must throw away anyway.
The finding stands on its own. Merlin's certification testbeds flew their busiest stretch of the observable window in March and April 2026, three aircraft across two countries peaking in the same eight weeks, in the flight pattern that matches certifying an autopilot. That is independent evidence the program is executing, gathered without a single word from the company.
It is not proof the certification passes. It is not a date. It is not a trade. It is monitoring intelligence, and the next window, June, is what tells you whether the spring ramp was a gear change or a test block. Same method, same sources, next update when the June archives are complete.
Part 8 covered the Q1 print, the burn, and the seven kill criteria. This post covered the airplanes.
Disclaimer and position disclosure. This post is research synthesis, not investment advice. The author holds no disclosed position for the purposes of this post. You should not buy or sell securities based on anything written here. I am not a registered investment advisor. I do not owe you a fiduciary duty. My conclusions could be wrong in ways I have not anticipated. Every figure is reconstructed from public data and stated sources, and where data was missing it is reported as missing. Projections and interpretations are model outputs, not guarantees. Microcaps are illiquid and can lose 50% or more in a single session. Do your own due diligence. Verify every number against the primary sources named above.
More in this series
The Gate Cleared and the Tape Forgot
Merlin Intelligence - Live OSINT Tool
Alyeska Is Underwater. None of the Thesis-Breakers Have Hit
MRLN's $105M Ceiling Is Not the Revenue, It Is the Authorization
Alyeska Wrote Both Checks and the Market Missed It
The UAE Deal Changes Everything Merlin Was Supposed To Be
MRLN: Part 4 - The Board, the Preferred Supernova, and the Catalyst Nobody
$MRLN Part 3: The DAWG, the $54.6B, and Why Merlin Is Positioned to Win the
Holy moly. Talk about using alternative data to verify your thesis. This is Richard Feynman at its finest!
Impressive work, thank you for sharing