Contrarian Codex — Extra newsletter (September)
"Cut what, though? No that's not a rhetorical question, cut what?" — a 25-page extra issue before Mart's month away: the budget arithmetic that leaves rates as the only lever, the missing insurer bid at 5%, a uranium equity shakeout against record term prices, the Saudi pipeline strikes, and news on four Codex holdings (PetroTal, Outcrop, Aldebaran, Merlin).
One-line take: the fiscal framework reduced to one question. (1) Cut what? Social Security, Medicare, Medicaid, veterans' benefits and net interest swallowed ~95% of the ~$4.85tn collected over the first 11 months of the fiscal year, growing ~8.5% against receipts up ~3%; add defense and "the tally clears 110% of receipts." Entitlements and defense won't move with the Iran conflict unresolved and a midterm six weeks out, and cutting either triggers a recession that widens the deficit — so "that leaves, to the surprise of exactly 0 people, rates," and "the only direction that helps is down even if it sparks more inflation." (2) Hikes as stimulus. Net interest already ran ~$1.05tn, +12%, before the hike; with public debt above 100% of GDP "part of every hike comes back as stimulus," and a full Volcker 8% on ~$40tn means ~$3.2tn a year of interest and a $4–5tn deficit, "north of 12% of GDP." Meanwhile diesel cracks set a closing record near $108/bbl, distillates tracked their lowest August since 1951, and employers expect health benefit costs +8.2% in 2027 — medical inflation that turns entitlements into "a hard currency liability." (3) The missing bid. The 10-year hit 5.04% and the 30-year topped 5.42%, and pensions and life insurers never showed up — private placements are ~23% of life insurers' admitted bonds, and selling unmarked private credit to buy 5% Treasuries would force capital-eating losses: "a Mexican standoff between private credit, insurers and the long end." So the Treasury funds itself: doubled long-end buybacks are "the softest possible form of yield curve control," and a buyback that fell short of its own $6bn cap sent the 10-year up 11bp in a day. His base case stays "a gradual grind," with the Bank of England's 2022 gilt U-turn as the template, pressure venting through the dollar, and "financial repression is the lower rates they need, and hard assets remain the durable position on the right side of it." (4) Uranium: sentiment down 12 to 19 ("the top of pessimism") on a ~10% equity drop, while the averaged long-term price (~$96.50) sits above any print in history and spot holds ~$90 on a rising floor; buyers are waiting for October budgets and the NEI Houston seminar. He added to his core uranium positions last week and will add more if the July lows are revisited — "come December I expect this episode to be remembered as yet another shakeout." (5) Oil: drone strikes on the East-West pipeline's pumping stations knocked out Saudi Arabia's ~4.5 mb/d Yanbu escape hatch; Hormuz flows recovered to ~12 mb/d on a 7-day average against ~20 mb/d pre-war; the Chinese restocking bid carried the rally, leaving flat price "extremely convex." (6) Holdings: PetroTal's Block 131 royalty cut ("governments do not hand stranded assets royalty cuts"), Outcrop's 58 Moz AgEq Santa Ana resource ("says more about expectations than about the rocks"), Aldebaran's 513m of 1.01% copper at Altar ("Somebody has this wrong, and I doubt it is the drill core"), and an apology on Merlin Labs, down ~35% in a week to under $2 — still held at 100% on the C-130J program.
1. Stocks & names mentioned
The four Codex holdings with news this issue, plus the two uranium names the market section leans on (Arrow and Kazatomprom). Bannerman, Mining Americas, Global Atomic and LibertyStream are named only as having news that will be covered in coming interviews and analysis pieces. Contextual third parties — Duke Energy (Belews Creek ESP), EDF / ASNR, Holtec (Palisades SMR-300s), Westinghouse, KHNP, SNURDC, Uranium One, Ma'aden, Barrick, Saudi Aramco / Bahri, Sinopec, Textron (SkyCourier UX), Sibanye-Stillwater, South32, Nuton (Rio Tinto), and the hyperscaler bond issuers (Amazon, Alphabet, Meta, Oracle) — are covered in the talking points rather than tabled as picks, per this source's convention.
| Ticker | Name | Research | View | What he said | At |
| PTAL | PetroTal | — | Positive | A full position kept "as an oil market hedge," with a re-rating expected "if oil stays elevated." Perupetro cut the Block 131 royalty on qualifying Cushabatay output to a sliding 5% / 9% / 15% from 23.5% (a flat 5% on the untested Noi and Copacabana formations), in exchange for two new wells within two years. At today's ~185 bopd that is only ~$500k a year, but he reads it as "Lima trying to make Peruvian upstream investable again." August output was ~11,700 bopd (YTD ~13,200, ~3% ahead of budget but down from 12,500+ in Q2), so the October drilling restart "has to deliver." Balance sheet: ~$105m cash against ~$37m debt, ~$50/bbl operating margins, north of $40m of quarterly EBITDA on a ~$340m market cap. "The market can keep pricing PetroTal like a stranded asset, but governments do not hand stranded assets royalty cuts." Watch item: the Bretana erosion-control contract, which should be signed "before the river rises again." Cost basis C$0.52, 100% allocated. | read ↗ |
| OCG | Outcrop Silver | — | Positive | The updated Santa Ana resource came in at ~58 Moz AgEq: ~30 Moz Indicated at 519 g/t and ~28 Moz Inferred at 369 g/t, up a little over 50% on the 2023 maiden estimate. He was "a little underwhelmed" on three counts. Indicated grade fell ~15% under a stricter classification, which he calls "the responsible move". The update slipped ~6 months. And the total "lands a long way short" of the ~100 Moz that had been floating around. His key risk is dilution: backed out, the resource is only ~4 Mt on veins often under 2m wide, so "a 519 g/t resource can turn into 350 g/t mill feed awfully fast." The upside: 12 known veins are not yet drilled to resource level, and the stock trades at roughly $2 per in-ground ounce with silver near $66. "Calling a 58 Moz resource with Indicated grades north of 500 g/t a disappointment says more about expectations than about the rocks." Cost basis C$0.17, 75% allocated. | read ↗ |
| ALDE | Aldebaran Resources | — | Positive | Six infill and geotechnical holes at Altar returned 513m of 1.06% CuEq (1.01% Cu) inside ~1,260m of 0.66%, plus other runs past 1,300m. That is "well over double" the 0.42% M&I grade, in holes drilled to convert Inferred to M&I ahead of the PFS. The late-2025 PEA ($2bn after-tax NPV8, ~20% IRR, ~$2.25/lb AISC over the first 20 years) used $4.35 copper and $2,500 gold. With copper ~$6.60 and gold ~$4,400, "that price deck has aged like milk." Yet the stock closed ~C$2.77, ~23% down on the year. He puts that down to the market pricing Argentine risk, ~$1.6bn of capex, a long runway and the Nuton exit. "Copper north of $6.50 a pound, holes carrying more than a kilometer of mineralization, and the stock down on the week? Somebody has this wrong, and I doubt it is the drill core." Cost basis C$0.79, 80% allocated. | read ↗ |
| MRLN | Merlin Labs | QT · SA · STK · FA | Positive | Still a full position despite an open apology. Merlin withdrew its small-aircraft certification application in New Zealand to target large aircraft, and unveiled the autonomous SkyCourier UX concept with Textron. A mid-month lockup expiry then took the stock down ~35% in a week to under $2, a record low (from $7–10 around June's CDR and $17 at the April peak). "I underestimated the downside volatility of a post-SPAC pre-revenue semi-AI play." The equity story is now the C-130J: sole prime on a $105m USSOCOM IDIQ, ~75 SOCOM airframes and ~570 C-130Js worldwide, and ~$3.3m per aircraft per year of piloting cost for a per-tail licence to bite into. A takeoff-to-touchdown flight demo would be "the biggest de-risking event this company has produced" and could take the stock back toward $7–10; meaningful revenue is guided only for 2027. The market cap is under $200m against $184m of Q2 cash (~$28m quarterly burn, runway reportedly into 2028), with the 12% convertible preferred's conversion price ratcheted from $12 to $6.67. "The market is pricing Merlin like the C-130J program already failed. It has not even flown yet." Cost basis $6.05, 100% allocated. | read ↗ |
| NXE | NexGen Energy (Rook I / Arrow) | QT · SA · STK · FA | Neutral | Cited as evidence, not a stance on the stock: a contact who conducted a site visit at Arrow "strengthened my belie[f] that Arrow delivering something like 16 million pounds by 2033-2034 instead of roughly 30 million by 2030-2031 remains a right-tail risk a lot of buyers are woefully unprepared for." | read ↗ |
| KAP | Kazatomprom (KAP: LSE GDR) | SA · STK | Neutral | Cited as evidence, not a stance: Kazatomprom lined up two sizable sales — a spot-term deal with China's SNURDC and a supply contract with Rosatom subsidiary Uranium One. "Russia has never before gone straight to Kazatomprom for a term purchase like this… every pound committed east is a pound Western utilities will not be bidding on in the 2030s." Pricing and volumes are confidential; shareholders vote in early October. | read ↗ |
2. Talking points
The coming month — content every Tuesday and Thursday while he is away
- The third long report this month (after the ~60-page #126 and the 30+ page WNA report); the next regular newsletter is due at the end of October.
- Pipeline while he is away from the desk: updated analysis pieces with valuation models for Mining Americas (the Cerro de Oro permits) and Global Atomic (updated price estimates and financing news), a renewed piece on Western Australian uranium mining, and interviews with Oceanwall's Ben Finegold, Comstock, LibertyStream, Bannerman Energy and Cauldron Energy.
Cut what? The budget lines that can't move
- Social Security, Medicare, Medicaid, veterans' benefits and net interest ran ~$4.6tn against ~$4.85tn of receipts over the first 11 months of the fiscal year: ~95%, growing ~8.5% against receipts up ~3%. "Put military spending on top and the tally clears 110% of receipts," so everything else "runs on borrowed money."
- Through Q3 (TBAC breakdown): Treasury (mostly gross interest) +$120bn / +10%, HHS +$98bn / +7%, SSA +$63bn / +5%, VA +$27bn / +10%, military +$31bn. The "savings" were a 55% drop at Education (a student-loan cost re-estimate, not money that stopped going out) and a 20% drop in "other."
- Luke Gromen's framing: Washington can cut only entitlements and veterans' benefits, defense, or rates. Cutting either of the first two tips the economy into recession, craters receipts and widens the deficit. "If we need a cut, it has to be on deficit spending and if we need a release valve, it's going to have to be the USD."
Defense is off the table, and so are entitlements — which leaves rates
- Fitch's August affirmation named military, interest, Medicare and Social Security as the forces limiting deficit reduction; it sees the general-government deficit at 7.4% of GDP this year and next (from 6.8% in 2025), the highest of any AA sovereign and "a far cry from the 3% by 2028" in Bessent's 3-3-3 pitch.
- "Good luck finding a politician willing to go near Social Security or Medicare roughly 6 weeks out from a midterm, or 6 weeks after one either, or ever." The interest bill is the only line left, "and the only direction that helps is down even if it sparks more inflation."
Hikes as stimulus — why the 1970s model doesn't apply
- Net interest ran ~$1.05tn over 11 months, +12%. CBO noted lower short rates had partly cushioned it, "a cushion the recent hike now flips into reverse."
- Volcker's hikes worked on bank-lending-driven inflation with low public debt. Today inflation is fiscally driven and debt is above 100% of GDP, so raising rates "blow[s] out interest expense, which for retirees parked in money market funds amounts to a raise, so part of every hike comes back as stimulus" (Lyn Alden's point).
- A full Volcker 8% on ~$40tn is ~$3.2tn a year of interest; add ~$4tn of entitlement and veterans' payments against a bit over $5tn of receipts, then defense and everything else, and the deficit reaches $4–5tn, "north of 12% of GDP."
Diesel and medical costs — the inflation hikes can't reach
- US diesel cracks set a closing record near $108/bbl in September, with distillate inventories tracking their lowest August since 1951 and retail diesel past its June-2022 record at ~$6/gal, while WTI sat in the low $90s. In 2022 the crack ran $64–73 with WTI at $110–119.
- The Beige Book had retailers citing transport and fuel costs; the K-shape shows in value-end price sensitivity against solid high-end spending. "Where crack spreads trade 3 to 6 months from now will do more to the inflation path than 25 versus 50bp from the Fed ever could."
- Marsh expects employer health-benefit costs per worker +8.2% in 2027 (biggest since 2003; Aon 9.5%). Unfunded entitlements are owed in knee replacements and insulin, which reprice automatically, "which makes a huge chunk of the federal budget behave like a hard currency liability."
The missing insurer bid — a Mexican standoff with private credit
- The 10-year printed 5.04% and the 30-year pushed above 5.42% ahead of the Fed meeting, ~19-year highs, and the pension and life-insurer bid "never really showed up."
- His explanation: private placements rose to ~23% of US life insurers' admitted bonds in 2025 from ~18% in 2021, and estimates put private credit at 11–16% of industry assets. Selling it to rotate into 5% Treasuries would force the marks, and an insurer facing a capital hole sells its most liquid asset — Treasuries. "No yield on the 30-year is high enough to pull that bid back in."
- Private-credit gating is "closer to the deal investors signed up for than a bank run," but solvency problems are showing at the margins. Regulators can loosen bank leverage rules and a central bank can create reserves; insurers have to sell something to buy something.
Who funds $2tn at 5%? Increasingly the Treasury itself
- Foreign central-bank holdings have gone "roughly nowhere for more than a decade while debt held by the public nearly tripled," and much of the remaining foreign long-end bid is hedge funds who stay only while volatility is low.
- Treasury said in mid-August it would at least double long-end buybacks from $2bn per operation (early September to early November). A new $6bn maximum disappointed a market braced for $7–10bn, and the next operation bought $5.19bn against $10.5bn offered. That was only the third shortfall in 53 long-end operations since 2024; the 10-year rose 11bp to 4.95%. Bessent had said he is "the house" now, "a line that aged about as well as milk left in the sun."
- Stepping in this early risks a Streisand effect. Buying duration with bills and TGA cash is "the softest possible form of yield curve control, closer to an Operation Twist run out of the Treasury." His MOVE framework flags a long end driven by term premium and supply "as a Treasury credibility problem."
Hyperscalers compete for the same capital — and AI eats the tax base
- Amazon, Alphabet, Meta and Oracle issued ~$194bn of bonds in 2026 through early July, +79% on all of 2025; Goldman expects ~$250bn this year and $400bn in 2027, lending weight to Warsh blaming part of the long-end climb on competition for capital. S&P 500 capex is +25% y/y, and Warren Pies notes the Fed had never cut into a capex boom until the past year.
- ~70% of federal receipts (~$3.4tn of $4.85tn) come from paycheck withholding; individual income plus payroll taxes are ~87%. "Anyone pitching AI as the productivity miracle that saves the budget should explain how the budget survives the payroll cuts that miracle requires."
China and Japan — the creditor has an option the debtor doesn't
- China's 10-year near 1.7%, ~330bp below Treasuries, is the conspicuous absentee from the bond rout — Beijing "ate its bitterness early." Where capital moves freely, yields drag each other around.
- Japan held $1.12tn of Treasuries in June with its own 10-year near 30-year highs. As a creditor it can pull capital home; the US has "no domestic pot of money to repatriate." If Japan has a problem Bessent can't fix, the selling lands on Treasuries and the dollar.
Base case: a gradual grind, and the BoE 2022 template
- "A long process punctuated by mini crises, technical explanations and all that jazz." The Bank of England's 2022 gilt U-turn (from preparing to sell to buying long gilts within days) is the template for a smaller-balance-sheet Fed chair ending up buying bonds "for technical reasons."
- Gromen rates a nonlinear air pocket as more likely than most do; given insurers' balance sheets Mart won't rule it out, but "they have stepped in and kicked the proverbial can every time the long end threatened to break" and a midterm makes that likelier, so pressure keeps venting through the dollar.
- The market's comfort stories (Fed credibility regained, an AI productivity miracle, stablecoins absorbing bills) "carry just enough truth to stay believable." His conclusion: "Financial repression is the lower rates they need, and hard assets remain the durable position on the right side of it."
Uranium sentiment: 19, back to the top of pessimism
- A ~10% drop gave back the post-newsletter rebound and dragged sentiment down 12 points to 19, though it remains above the July bottom.
- "I believe this presents an opportunity and that is why I added to my core positions in uranium last week and if we fall more to revisit the July lows (or worse) I will add more."
WNA in one paragraph: sellers hold the strongest hand in years
- Utilities now sign market-referenced contracts with floors in the $80s and ceilings past $150; volume and timing dominate procurement with price "trailing well behind." "You cannot load a balance sheet into a core."
- Term contracting ran ~110m lb (2024) and ~116m lb (2025) against requirements near 200m lb. The 229m lb of US/EU utility inventory came from post-Ukraine double-buying and upflexed legacy contracts and "precious few of those pounds will ever come back to market." EIA: 52% of the next decade's US requirements uncovered, ~186m lb unfilled through 2035.
Spot stalemate, record term
- ~4m lb traded spot across just under 30 deals from late August to early September, lifting spot from ~$87.50 to ~$90; since London only two deals printed in five sessions. Buyers are waiting for October budgets and the NEI fuel seminar in Houston while sellers won't soften offers. Late in the year he struggles to see the stalemate lasting.
- The bid keeps appearing in the high $80s, "a dynamic floor that is rising together with the consistent marching upwards of the term price." Term: 3 deals for just over 5m lb this month, the year's total past 42m lb across 32 transactions (excluding 40–45m lb of Indian demand). TradeTech held $97 and UxC rose $2 to $96, and the ~$96.50 average is above the old $95 high of 2007–08. "A record term price on volumes that still fall miles short of replacement rate."
Demand stack: Kazatomprom's eastern deals, KHNP's bundle
- Kazatomprom's spot-term deal with SNURDC and its supply contract with Uranium One — Russia buying direct from Kazatomprom for the first time as far as he can tell.
- KHNP's 800,000 SWU tender (~26–31m lb of feed) will take offers that bundle uranium as long as nothing is Russian, with a standalone uranium tender to follow if bundling comes up empty.
- Equities are falling on fatigue while spot barely flinches: "Nothing in the fundamentals justifies the move down… come December I expect this episode to be remembered as yet another shakeout, the sector's favorite pastime."
Nuclear headlines: France past 60, Duke's ESP, NRC rewrite, Holtec, NEI survey
- France's ASNR reportedly sees no major technical obstacle to running 32 of EDF's 900 MW reactors past 60 years. Pair that with the WNA finding no age-related capacity-factor decline and ~29 GW that cautious models retired in the late 2030s–2040s stays on: ~13–15m lb a year at 450–500k lb/GW.
- Duke filed its Belews Creek early site permit (up to 2,250 MW of advanced reactors) only at the end of 2025 and the NRC hearing is already scheduled — the same Duke running the RFP. The NRC plans a broad risk-informed rewrite of its rulebook, and Holtec got the go-ahead to start ground works at Palisades for two SMR-300s (680 MW) before holding a Limited Work Authorization.
- NEI's chief nuclear officers now pencil in ~28 GW of new US capacity by 2040 (from ~23 GW last year), with interest in new build up 54%. "I will believe in the gigawatts when I see concrete poured," but anything above zero matters when zero is what is priced in.
Saudi Jabal Sayid and Korea's reactor list
- Saudi Arabia unveiled ~110m tonnes of ore at Jabal Sayid with heavy rare earths and "promising" uranium, but no grade: at 0.01% that is ~24m lb before recovery losses, at 0.05% ~120m lb. It goes "in the drawer marked headlines with big numbers attached" until a grade appears.
- Seoul's $350bn trade-deal pledge: after shipbuilding's $150bn, up to 8 large US reactors at ~$15bn each (6 reportedly Westinghouse, 2 Korean designs). That is ~9 GW, another 4–5m lb a year plus initial cores. "Anyone still framing nuclear demand growth as a China and India story alone should take a closer look at the global shopping list."
Oil: the East-West pipeline goes down
- Brent came within a whisker of $110 intraday on the Saudi pipeline news, then slid to close the week just under $104 (WTI ~$100) and was battling the pivotal $100 mark as he wrote. ULSD touched ~$5.25/gal, which still puts the diesel crack above $100/bbl, with distillates ~13% below the 5-year average.
- Drone strikes on pumping stations knocked out the line feeding ~4.5 mb/d through Yanbu. Riyadh blamed drones from Iraq, other reports the Houthis. This came a week after Bessent called Hormuz "a worthless piece of water within 2 years." The IEA already had Saudi crude supply near 6 mb/d in August, a 30-year-plus low.
- A bypass could restore about half of Yanbu's throughput within a month, with full repairs in 4–6 weeks. Yanbu stocks are down 7m+ barrels to ~9m, so Riyadh has pivoted back to Ras Tanura and Juaymah through Hormuz, with ship-to-ship handoffs off Sohar and Fujairah.
Hormuz flows are rising — still far from pre-war
- Satellite imagery puts Saudi crude through the strait at ~2.8 mb/d over six days against ~700k in August. Hormuz flows reached nearly 12 mb/d on a 7-day average by mid-month, the fastest since the June MOU, with virtually no Iranian crude and Iraq the biggest contributor (Kpler). That is against ~20 mb/d pre-war. "A rising line is a rising line and it has certainly helped."
- Some days show as few as 4 visible commercial transits against ~125 a day pre-war, with dark transits and unattributed Gulf of Oman ship-to-ship transfers filling the gap.
China's restocking bid carried the rally — flat price is convex
- Chinese crude imports rose to ~8.9 mb/d in August from a decade-low ~7.1 mb/d in June (still ~23% below last year). Fuel exports rose 29% m/m, premiums for Congo's Djeno reached $20 over Brent, and the teapots are being priced out. Sinopec's chairman thinks Chinese oil demand likely peaked in 2025, and the IEA estimates EVs displaced ~1.5 mb/d in Q2.
- Global observed inventories are down ~500m barrels since the war began (~95m in August), and the SPR is near its lowest since the early 1980s. A small Chinese pullback against accelerating Hormuz flows could knock prices hard; a strike on Abqaiq or Ras Tanura would do the reverse. He recommends his Rory Johnston interview. "Be quick on your feet if you're trading oil."
Codex holdings — news flow
- Bannerman, Mining Americas, Global Atomic and LibertyStream have news that will be covered in interviews and standalone pieces over the coming month.
- PetroTal: Block 131 royalty cut, a full position as an oil hedge. Outcrop: 58 Moz AgEq, "underwhelmed" but not writing off 100 Moz. Aldebaran: Altar infill at double the M&I grade while the stock falls. Merlin: an apology, a sub-$2 record low, and the C-130J flight demo as the turnaround trigger.
3. In plain English
PTAL — PetroTal Positive
PetroTal pumps oil in the Peruvian Amazon. The market prices it as if its assets are at risk of being stranded, mostly because a river threatens its main field and Peru has been a hard place to operate. Peru has now cut the royalty (the government's share of every barrel) on part of a second field from 23.5% to as low as 5%, as long as the company drills two new wells there.
Mart is honest that the money is small today — about half a million dollars a year at current output. His point is the signal: "governments do not hand stranded assets royalty cuts." With ~$105m of cash, ~$37m of debt and more than $40m of quarterly operating profit at today's oil prices, he keeps it as a full position and as insurance in case oil stays high. The things to watch are the October drilling restart and a signed contract for the riverbank-protection work before the water rises again.
OCG — Outcrop Silver Positive
Outcrop is exploring high-grade silver veins in Colombia. Its updated resource — the official estimate of metal in the ground — grew about 50% to ~58 million ounces, but the grade of the best-defined part fell and the total came in well short of the ~100 million ounces some investors hoped for.
Mart's main caution is about what happens when the rock is actually mined. These veins are often under two metres wide, and mining a narrow vein inevitably takes waste rock with it ("dilution"), which can drop a 519 g/t resource to something like 350 g/t at the mill. Against that, a dozen known veins have not been drilled to resource level yet, and the stock values each ounce in the ground at about $2 while silver trades near $66. His verdict: calling this a disappointment "says more about expectations than about the rocks."
ALDE — Aldebaran Resources Positive
Aldebaran owns 80% of Altar, a very large copper deposit in Argentina. Its latest holes found more than half a kilometre of rock at about 1% copper, over double the average grade of the deposit's best-defined resource. These holes were drilled to upgrade lower-confidence tonnes before the next engineering study, so good results there matter directly for the study.
The company's last economic study assumed $4.35 copper and $2,500 gold. Today copper is about $6.60 and gold about $4,400, so the real economics look much better than the published numbers. Yet the stock is down about 23% this year. Mart accepts the risks the market is pricing (Argentina, ~$1.6bn to build, a long wait to first copper, a partner walking away last year), but his conclusion is blunt: "Somebody has this wrong, and I doubt it is the drill core."
MRLN — Merlin Labs Positive
Merlin builds software that lets aircraft fly themselves. After a strong start the stock collapsed to under $2, first on a change of certification plans and then when early investors were allowed to sell (a "lockup expiry"). Mart apologises openly. He had warned it was an all-or-nothing bet to be sized small, but says he underestimated how violently a young, pre-revenue company that listed through a SPAC can fall.
He is still holding. The company is valued at under $200m while holding $184m of cash, so the market is putting almost no value on the technology. The whole story now hangs on one program: a US special-operations contract to let C-130J cargo planes fly with fewer crew. A successful full flight demonstration would be the biggest proof point the company has had, and he thinks that plus production orders could take the stock back toward $7–10. Until then, "the market is pricing Merlin like the C-130J program already failed. It has not even flown yet."
Analysis distilled from the Contrarian Codex extra newsletter edition (cover dated 21-09-2026; subscriber PDF linked above, not reproduced here). Cost bases and "% allocated" figures are the author's own published portfolio figures. Contextual third parties named in the analysis sections (Duke Energy, EDF / ASNR, Holtec, Westinghouse, KHNP, SNURDC, Uranium One, Ma'aden, Barrick, Saudi Aramco, Bahri, Sinopec, Textron, Sibanye-Stillwater, South32, Nuton / Rio Tinto, Amazon, Alphabet, Meta, Oracle, Fitch, TBAC, Marsh, Aon, Kpler) appear in the talking points rather than the table. For personal study — not investment advice. Source material © Contrarian Codex / "Mart".