Contrarian Codex — research hub

Stock & name index

▲ Positive

► Neutral / referenced

▼ Negative

Overall thesis

• The dollar is the release valve. Carry this much debt with a tax base wired to asset prices and the rate decision stops being about inflation and becomes one about the currency. Hikes strengthen the dollar and tighten every screw (self-limiting); cuts + bank-rule loosening (banks gorging on Treasuries) is "QE routed through the banking system." Direction isn't in doubt — but being early while levered is the same as being wrong. The 28 August Jackson Hole update makes the mechanism explicit: Warsh "will manage the short rate and let the long end fend for itself, which leaves the release valve precisely where it has been all along with USD written all over it." The market has fully separated the policy rate from the term premium (the whole hawkish repricing happened inside 2 years while 10s and 30s barely moved), Treasury is running the long end alone with a $4bn-per-operation buyback against $2.1tn of financing, and a hike is "mechanically expensive for a Treasury that has parked itself" in bills — a Chair "boxed in by an issuer who needs him not to move."
• Sticky inflation, an energy shock in waves. April was the on-ramp, not the peak: the disinflationary goods crutch is gone, the Hormuz closure baked second-half pass-through into food/freight/fertilizer, and producer prices lead consumer prints. Constructive on real assets; avoid long-duration bonds. By 28 August the print is the 65th consecutive month above the 2% target (core PCE 3.3%, headline 3.7%), the policy rate has sat at 3.50–3.75% for five meetings with three regional presidents dissenting to hike and cuts "not on the menu at all," July's energy component fell 1.5% so the crude pass-through lands in the August/September data instead, and 50% tariffs now run in both directions with Canada — "I struggle to construct a path where you run 50% tariffs in both directions against your largest trading partner and get a disinflationary outcome out the other end."
• Gold as a referendum on monetary credibility. Central banks have overtaken Treasuries as the largest official reserve asset; a record 45% plan to add more and 74% see the dollar's reserve share lower in five years. Corrections (to the high-$3,000s/low-$4,000s) are gifts; the multi-year path runs to $10–15k. PHYS reserves held throughout. Newsletter #126 adds a second, policy-driven engine: the US carries 261,498,926 oz at a statutory $42.2222 fixed in 1973 (~$11bn of book against ~$1.18tn at market), and revaluing it credits the General Account with >$1.1tn "without selling an ounce or auctioning a single bond" — ~$260bn of fiscal headroom per $1,000 of gold price, out of the same account now funding the $4bn buyback operations that fresh bill issuance would otherwise defeat. Roosevelt ran the identical trade in 1934. The regime change is who benefits: a rising gold price was Volcker's enemy as a thermometer on dollar confidence, but "that logic stops holding the moment gold becomes collateral behind $1 trillion of fiscal headroom." Limits stated: no new wealth, under 3% of $40tn of debt, and the orthodox legal reading needs Congress to move the statutory price first — but Bessent's denial and his stated intent to "monetize the asset side of the balance sheet" do not coexist indefinitely.
• Uranium: "slowly, slowly, then all at once" — and we are at the front of the contracting cycle. Term price at a nominal record (~$97/lb TradeTech, an 18-year high) props spot; the existing fleet + life extensions + restarts + construction already leave a gap before the financial/sovereign/military/SMR/hyperscaler buyers queued behind it. Summer-lull weakness is a chance to add to core holdings. Cameco's Q2 call sharpened the mechanism: contracting is still running below replacement rate while the price climbs anyway — every prior time the market reached a point like this it was at the back end of a contracting cycle, whereas this time it is at the front (COO Grant Isaac's read: the path from here is "likely triple digits"), with new layers carrying higher floors and higher ceilings based mid-to-high $90s. The corollary for the producers: today's realized prices are held below spot by ceiling-capped older contract vintages, and Cameco's own sensitivity table (a book flat at $100 spot realizing ~$67/lb in 2026 vs ~$88/lb by 2030) shows the upside is loaded into the back years as those roll off. The 2026 WNA report (14 September) is the hallway confirmation: sellers hold "the strongest position… for many years," utilities now sign market-referenced floors in the $80s and ceilings past $150 they refused 18 months ago, spot has a rising mid-$80s floor bid by traders, utilities and producers, and the 229m lb utility inventory is a 2027 cushion against 52% of the next decade of US demand uncovered. Demand keeps stacking (KHNP's 800,000 SWU tender, the NNSA's ~40m lb US-origin RFI, India contracting 2–3x its stated need, hyperscalers and BHP circling) while the right tail sits in supply (Arrow slipping, Kazatomprom's acid-driven cost blowout). With equity sentiment back near depression he spent another third of the cash on the cornerstones: 10% back into Denison at $2.85 and Energy Fuels to a full position plus Jan-27 $15 calls. The next day's Triangle Investor Interviews conversation (15 September) put the report into one phrase: the market "has properly run out of shock breakers." Mobile inventories are thin and legacy contracts already flexed up, so an Arrow slip to ~16m lb in 2034–35 (against ~29.5m in 2031) "will be a shock." Replacement costs have roughly doubled since the 2021/22 feasibility studies (Cameco: greenfield needs $120), and he puts 85%+ odds on $150 uranium. SMRs will not arrive in volume before the end of the 2030s (Darlington is the gate), but he calls their demand "severely underpriced." The 19 September Investing News interview puts a clock on it: "$150 uranium is inevitable… within 2 years," reached in stair-steps as term drags spot, with ~70% of new contracts market-referenced (floors in the $80s, ceilings $150+ or none) and the sector "beholden to macro" until the dollar weakens. (2026-SEP-19) Three days later the extra newsletter logs the equity shakeout. Sentiment fell 12 points to 19, "the top of pessimism," on a ~10% drop. Meanwhile the averaged long-term price (~$96.50) is above any print in history and spot holds ~$90 on a rising floor while buyers wait for October budgets. He added to core uranium positions and pre-committed to add more at the July lows. (2026-SEP-22)
• Oil & the Hormuz restart is no light switch. First barrels out are stockpile, not fresh flow; Qatar's destroyed LNG trains are a multi-year hole. The old "supply always shows up" / "the strait can't be closed" assumptions are permanently re-rated. When the conflict re-escalated in mid-July (Part 24 — a second US-Iran strike wave, Tehran declaring Hormuz "closed"), Brent still held just under $79: the market prices a rhythm of skirmishes with windows of calm, not a real closure — and Iran, with ~90% of its exports through Kharg Island, would be sanctioning itself to shut it. The revoked oil-sales waiver is a floor, not a breakout. Oil positions kept for the asymmetric upside. By month six (11 August) the strait is still effectively shut beyond minor transports; Iran and Oman have near-finalised new lane coordinates, which "does not reopen the waterway," and with Washington's concession list against Trump's counter-demand for compensation the sides are further apart than a week earlier — while the market increasingly trades the Strait as a headline rather than a wall, leaving another oil run as one of only two things that would break the equity momentum. By 28 August the risk premium has deflated on an IRGC-claimed revenue-sharing deal with Oman that both foreign ministries describe only as an interim transit framework, with Iran and the western naval coordination group designating different corridors, Trump claiming the mines are gone and 10m barrels transited in a day against an IMO head saying the strait is not open and the mines unconfirmed, and the EIA seeing Middle East output below pre-conflict levels until early 2027 — "the risk premium has deflated on a framework nobody has signed, that would not reopen the waterway by itself, and that depends on concessions nobody has agreed to make… that optionality is being sold far too cheaply here," with shipping volumes and diplomacy the confirmation he wants before adding.
• Copper deficit underneath a policy-whipped tape. Real, building structural deficit (Grasberg, Chilean declines, treatment charges on the floor) traded short-term as a US-tariff / Hormuz-sulfur derivative.
• AI capex is the market's load-bearing wall — and its fault line. ~86% of forward-margin improvement traces to two sectors (semis + energy); the index can grow into its multiple if the capex cycle holds, but the same chart maps where it breaks. Watch the rate-of-change in hyperscaler capex guidance and the dollar. The 11 August update names the mechanism by which it breaks: the buildout "has become a credit product" — Nvidia's $500bn private-capital financing with compute pledged as collateral and NVDA backstopping up to 25%, Goldman the only bank in the consortium and prospective lead bookrunner on the public debt to come, Jensen Huang calling compute "an investable infrastructure asset," data-center securitizations cleared by the SEC to skip post-2008 risk-retention ($2.4bn issued in 2020 → $15.5bn last year, on pace for another record) and ICE paying $6bn for a corporate-bond venue. Demand is still roaring (Microsoft's record ~half-trillion single-session add on Azure +43%) — but the funding side has been re-plumbed into credit, so the complex now breaks on credit conditions rather than AI demand alone.
• Direct AI/critical-materials exposure beyond the commodities. The "Codex 'other'" basket reaches into the AI/defense/critical-minerals build: Merlin Labs (flight autonomy), Comstock (solar recycling + data-center land/power), LibertyStream (oilfield-brine lithium), ASP Isotopes (Si-28 + helium), Westwater (graphite anode), Energy Fuels (rare-earth mine-to-magnet). These are the highest-variance names in the book and he sizes them accordingly — the late-July adds (LIB at C$0.80, LODE at $2.91, plus a small speculative Dec-26 $5 LODE call) all came with the risks stated before the thesis: pre-cash-flow companies that "need the cash to continue," first-of-a-kind facilities where "things can (and likely will) go wrong," and a repeated instruction to size to your own risk tolerance. He added to LIB again on 17 August in the $0.60s, treating the further ~25% drawdown as an unexplained sell-off — "no clear reason … beyond investors being afraid of further dilution (either for capex spending or the expected US uplisting, both of which were already expected)" — with the first facility due into production this year the milestone the position now hangs on.

The product — Contrarian Codex

The trade types below are drawn from Mart's own descriptions across the issues archived here (esp. the 2026-MAY-29 and 2026-JUN-26 newsletters and the standalone analysis pieces).

How it serves retail investors:

• Institutional-depth commodity research down-market. Fuel-cycle, drill-result and feasibility-study detail (uranium term pricing, NPV/IRR, AISC, grade-thickness) translated for non-specialists at newsletter prices.
• A transparent model portfolio with cost basis. Every position carries an entry price and an allocation figure, and sells are flagged with the realised profit — so a subscriber can see the actual book, not just narrative.
• Explicit, falsifiable price targets. The deep dives publish bear/base/bull numbers and the assumptions behind them, then revisit them on news.
• Risk-first cash framework. The recurring advice is the "sleeping level" — size the cash cushion so no single headline forces a bad decision — which is exactly where retail typically does the opposite.

Issues & reports

One dated page per issue/report — each has its stock table, talking points and the source material. Newest first. (Written sources — no video: "PDF" opens the original report; Discord members-channel notes link to the post plus an archived transcript.)

For personal study — not investment advice. Source material © Contrarian Codex / "Mart".