Actionable insights — The Power Shortfall's Powerful Solution

1. Walk the supply chain from the demand shock and cross out each link for a stated reason

1. Start from a demand fact you are willing to treat as settled, and say so explicitly rather than smuggling it in — it is the one assumption the whole position rests on.
2. List the small number of technologies or inputs that can physically satisfy it. Keep the list short and exclude anything that cannot deliver at the required scale or duty cycle.
3. For each, check the direction of travel in the relevant market, not just its current share — a large input that is shrinking is a different investment case from a small one that is growing.
4. Then decompose the surviving option into its actual production chain (raw material → processing → conversion → end use) and ask at which link value is concentrated.
5. Cross out each link explicitly, giving the reason: no listed vehicles, no earnings, state-owned, already priced. The eliminations are the argument; a thesis that jumps straight to a ticker has skipped them.
6. Take what survives, and be honest that it survived by elimination rather than by being the best expression.

• A chain walk that stops at the first investable link rather than the most concentrated one — that is the tell of a thesis reverse-engineered from a position already held. Also watch for the settled premise being the actual risk: if the demand shortfall is met by efficiency, demand response, or the AI build-out slowing, every downstream step is sound and the position still loses.

2. Read the aggregate line on a stacked chart before reading the components

1. On any stacked area or composition chart, look first at the top of the stack — the total — and note where it is flat, where it inflects, and how long each regime lasted.
2. A long flat total with violent churn underneath tells you the industry has spent that period replacing capacity rather than adding it: the operating skills, supply chains and permitting throughput built up are for substitution, not expansion.
3. Date the inflection in the total and compare it to the lead time of the assets involved. If demand turns years before supply can respond, the gap is arithmetic rather than forecast.
4. Only then read the components, and read them as shares of a fixed pie during the flat period — a rising band there took share from another band, it did not meet new demand.
5. Ask which band has the capacity to grow in absolute terms when the total starts rising, given build times and current construction pipelines.

• Monthly or preliminary data at the right-hand edge of the series driving the inflection call — this chart's 2025 figures are flagged preliminary. One year of upturn is a data point; confirm it against interconnection queues and load-growth forecasts before treating the inflection as a regime change.

3. Locate the chokepoint, then test it for investability before building a thesis on it

1. For a physical supply chain, compute one country's or company's share at each stage rather than for the industry as a whole. Concentration is almost never uniform along a chain.
2. Find the stage where share is highest and substitution is slowest — long-lived, capital-intensive, licence-gated processing is the usual answer, not extraction.
3. Sanity-check the share definition before using it. Attributable equity, production, and capacity are three different numbers, and joint ventures inflate all of them if read carelessly.
4. Now do the step most theses skip: check whether that stage has any listed pure-play exposure at all. Screen it before writing the thesis, not after.
5. If it does not, do not force a proxy. Either move to the adjacent link that is investable, or accept a broader instrument and say plainly that you are giving up the concentration you identified.
6. Keep the chokepoint on file anyway — it is the reason the broader instrument should rerate, and the trigger to revisit if a listed vehicle ever appears.

• A listed enrichment or conversion vehicle appearing, or Western capacity expansions being funded — that is the event that would let the trade move to the concentrated link. Watch equally for the footnote risk: state-owned shares change through asset sales rather than through market share, as this chart's own ~14%→~13% note shows.

4. Choose the commodity over the operators when the operators do not yet earn anything

1. Once the theme is decided, decide separately how to own it: the physical commodity, the producers, the royalties, or the downstream users. These are different bets with different failure modes.
2. Test the producers on current earnings, not on the commodity price at which they would earn. A sector that is broadly unprofitable at spot is priced on future prices plus execution — you are buying two forecasts, not one.
3. If the thesis is demand-side and forecastable from a construction pipeline, prefer the instrument that expresses only the price: a physically-backed vehicle carries no permitting, cost-overrun, grade, labour or dilution risk.
4. Accept the trade-off explicitly: no dividends, no production growth, no operational leverage, and typically a discount/premium to net asset value that moves against you at the wrong times.
5. Note the conditions under which you would rotate — usually once the price move has made the producers cash-generative, at which point the leverage you gave up becomes the better bet.

• The point at which producers turn cash-generative at prevailing spot — that is the rotation signal out of the physical vehicle and into leverage. In the meantime, watch the trust's discount to NAV: a physical vehicle's tracking of the metal is only as good as its premium/discount behaviour, and in a drawdown that discount widens exactly when you would want to add.

5. Convert a foreign build-out into a benchmarked number rather than a headline count

1. When citing another country's build-out, do not stop at the unit count — convert it to capacity, which is the quantity that consumes fuel.
2. Benchmark that capacity against something the reader already has a feel for: an existing national fleet, current annual demand, or the installed base of the thing being replaced.
3. Separate the domestic build from the export order book. An exporter's pipeline is a claim on future fuel demand located elsewhere, and it is often the larger number.
4. State the exclusions that keep the count honest — pilot-stage or proof-of-concept technology should not be counted as capacity.
5. Use the comparison to size the demand side of the commodity thesis, and check it against the fuel requirement per gigawatt rather than assuming the relationship is linear in units.

• Construction starts and grid connections rather than announcements — reactor pipelines slip by years, and "planning stage" units routinely never break ground. Watch also for the fuel-timing mismatch: a reactor's first core is loaded years after construction starts, so an order book converts into uranium demand on a lag, which is precisely why a physical position has to be sized to survive the wait.