Actionable insights — July 2026 monthly newsletter

1. Run the sector exposure as a dial, not a switch — off an explicit overbought/oversold state

1. Accept the premise first: in a structurally bullish but historically volatile sector, the return comes from staying invested — so the variable you manage is how much, not whether.
2. Publish a single sector-state label each month (oversold / moderately oversold / neutral / overbought) and let the cash weighting follow it mechanically, rather than re-litigating the thesis every drawdown.
3. Raise cash into strength (trims taken at named prices) and spend it into weakness — and state the intent before the deployment so the decision isn't made in the moment.
4. Compute the invested percentage honestly: strip out both cash and any physical/defensive holding to see true miner exposure, which is the number that actually drives the drawdown.

• Cash percentages that drift up purely because holdings depreciated (Huhn flags this explicitly) — that is passive de-risking, not a decision, and it needs an active rebalance to become one.

2. Make the stance change explicit and dated — quote last month against this month

1. Restate your prior published stance verbatim before revising it, so the change is auditable and can't be revised in hindsight.
2. Name the two or three conditions that changed. Here: a further double-digit drawdown plus arrival into the seasonal window.
3. Convert the new stance into a deployment schedule with a window, not a single-day decision ("over the next 30–60 days"), which removes the need to call the exact low.
4. Attach a falsifiable price marker for the thesis playing out, so the follow-up is measurable.

• The deployment actually happening — a stated intent to spend cash that is still unspent two letters later is a stance that quietly reverted.

3. Time deployment with seasonality, then treat the seasonal low as a window rather than a date

1. Keep a seasonality chart for the sector and know where the calendar low typically falls; check whether current weakness coincides with it before attributing the selloff to news.
2. Stack the seasonal signal on an independent one (drawdown depth, relative-strength breakdown, RSI) — seasonality alone is too weak to trade.
3. Deploy across the window in tranches rather than at the projected low; the calendar tendency gives you a period, never a day.

• A seasonal low that fails to arrive on schedule while fundamentals are intact — the pattern breaking upward is itself information about the strength of the underlying bid.

4. "Price becomes narrative" — audit whether a bearish story is a fact or a rationalisation of the tape

1. When sentiment deteriorates, list the fundamental facts that changed in the period. If the list is empty and only the price moved, you are reading narrative, not news.
2. Classify each cited cause: is it about the commodity's supply/demand, or about rates, positioning, beta and flows? Only the first bucket is a thesis risk.
3. Treat the sentiment low itself as the setup — "these types of quiet periods accompanied by depressed/subdued sentiment provide fertile setups for positioning."
4. Maintain the falsification list separately, so a genuine break in the structural bid is still recognised when it comes.

• The moment a "narrative" item becomes a fact — e.g. long-term contracting volumes falling, or the LT price rolling over rather than grinding up.

5. Read the long-term contract price as the floor, and know which reporter leads

1. Track spot and long-term prices as separate series. In a market where nearly all volume moves under multi-year contracts, the LT price — not spot — is the producer's economics and the floor under spot.
2. Know the reporting lag between price services and weight the leading one: "TradeTech consistently leads UxC in price reporting by 30–60 days in terms of the LT price trend." Blend the two for a headline number.
3. Read the forward curve for the market's own view (3-year, 5-year forwards) rather than substituting your own forecast.
4. Extend the same reading to the rest of the fuel cycle — conversion and enrichment prices often tighten first and confirm the direction.

• UxC catching up to TradeTech (confirmation), or TradeTech stalling for two consecutive months (the earliest warning the term bid is fading).

6. Use two fixed ratio charts every month — the equity proxy vs the commodity, and vs the index

1. Pick one liquid ETF as the sector-equity proxy and chart it against (a) the underlying commodity and (b) the broad index — the same two ratios, every month, so the series is comparable over years.
2. Ratio vs commodity answers "are equities cheap relative to the thing they produce?"; ratio vs index answers "is this sector-specific or market-wide?"
3. Add RSI to the ratio, not just the price, and treat a fully broken-down ratio at a multi-month low with washed-out RSI as an entry zone.

• The equity/commodity ratio turning up while the commodity is still flat — that is the first evidence positioning has finished, and it usually precedes the price leg.

7. Convert ETF flows into a "mandated selling" number to identify a price-insensitive seller

1. Track share counts outstanding and AUM for each sector ETF monthly — creations and redemptions, not just performance.
2. Translate net redemptions into a dollar figure of forced selling. Those sellers transact regardless of valuation, so the resulting weakness carries no fundamental information.
3. Check the concentration of the largest ETF: heavy top-five weightings mean redemptions land disproportionately on a handful of names, distorting their prices relative to the sector.
4. Once flows stop being negative, the same mechanic reverses — the marginal buyer returns without any change in fundamentals.

• The first month of net creations after a redemption run — the flow turn typically precedes the relative-strength turn.

8. Gauge how late the bull market is with a small-cap vs large-cap AUM rate-of-change chart

1. Index the AUM of a large-cap sector ETF and a junior/small-cap one to a common base date and chart the rate of change, not the level.
2. The premise: in genuinely speculative late-cycle phases, retail money chases the juniors, so the small-cap line crosses above and stays above the large-cap line.
3. While the small-cap line remains persistently below, the speculative phase has not begun — which is a bullish read on remaining runway, not a bearish read on the juniors.

• A sustained crossover of the junior line above the large-cap line — on this framework that is the signal to start reducing, not adding.

9. Price a physical trust off its discount to NAV — and convert the discount into an implied commodity price

1. For any closed-end physical holding, compute the discount/premium to NAV monthly and translate it into the commodity price the market is implicitly paying: spot × (1 − discount).
2. Compare that implied price with the actual spot price. A wide discount means buying the physical commodity below the market price, with the discount itself as a second source of return if it narrows.
3. Track the discount's change as a sentiment gauge — a discount widening several points in a month is capitulation, not valuation.
4. Remember the second-order effect: a trust at a discount cannot issue units accretively, so its buying — a real source of spot demand — switches off exactly when sentiment is worst.
5. Applied to an operating vehicle, the same arithmetic tells you whether a buyback is the cheapest way for the company to acquire more of the commodity.

• The discount narrowing back toward NAV: that both re-rates the units and restarts the trust's spot-market buying, a double catalyst.

10. Stress-test whether your thesis actually needs the narrative it is being traded with

1. Identify the thematic basket your holdings have been swept into, and be explicit about the transmission mechanism — here, "Wall Street has created customized 'factor baskets' around the AI thematic which can include uranium equities."
2. Separate the price linkage (correlation via flows) from the demand linkage (does the volume you need actually depend on that theme?).
3. Find a historical dry run for the shock and check what it did over 6–12 months, not one day.
4. Size the theme-independent demand: if the underlying is short of supply for the existing installed base plus already-committed growth, the theme is upside, not the thesis.
5. Look for the offsetting driver the bearish story ignores — here, Chinese power buildout as the mirror image of any US AI slowdown.

• Evidence the linkage has become fundamental rather than flow-driven — e.g. utilities deferring contracting because of data-centre cancellations, which would move it from the price bucket to the demand bucket.

11. Read a producer's incentives, not just its costs, to forecast supply

1. Get the actual cost trajectory of the marginal/swing producer over a decade, not a single year — that is what sets the price floor for the whole market.
2. Then read the fiscal regime: tax structures tiered by output, price-triggered overlays and royalties change what the producer wants to do, independently of what it can afford to do.
3. Ask whether the incentive is self-reinforcing or self-correcting. A tax that penalises volume above a threshold makes restraint more valuable as prices rise — the opposite of the usual "high prices cure high prices" reflex.
4. Cross-check the conclusion against a competitor's public commentary before accepting it.

• Kazatomprom guiding production up into a higher price — that would break the alignment and is the single most important supply-side falsifier of this framework.

12. Size positions so a blow-up is survivable — then triage it on capital, customers and clarity

1. Underweight the positions where a single operational input can stop the business, and reserve overweights for the assets with balance sheet and jurisdiction on their side.
2. When one breaks, quantify the remaining exposure before deciding anything: what does a further 50% decline actually cost the portfolio?
3. Mark it "no new money" rather than reflexively averaging down, and let three named conditions govern the decision — adequacy of capital going forward, flexibility from offtake customers on delayed delivery, and the coherency of the company's next disclosure.
4. Say it plainly: naming a blow-up a blow-up preserves the credibility to be believed on the other positions.

• The terms of the rescue financing and whether offtake customers grant delayed delivery — those two, not the drill results, determine whether anything is left for existing shareholders.

13. Pre-commit to a volatility playbook so the drawdown can't make the decision for you

1. Set the sector allocation deliberately small enough that its volatility is tolerable — "remember…this sector will hit above its weight."
2. Deploy in tranches (new subscribers are told roughly one third of the intended allocation at a time, spread over weeks or months), never in one go.
3. No short-dated call options to time a rally, and no margin — leverage is what forces you out at the low and forfeits the multi-year horizon the thesis requires.
4. Cap single-name concentration against published target weightings, and re-read the underlying thesis during drawdowns rather than the price.
5. Bound the horizon honestly: "we have never suggested that we are headed into a 20-year bull market… we continue to be quite confident that we have entered into a 2–4+ year bull market from here."

• Your own behaviour, not the tape: needing to check prices daily, or wanting to add outside the tranche schedule, both indicate the allocation is too large.