Chains. Four questions per entry.
A public log of consequence-chain reasoning. Each entry traces a macro shift through its second-order effects, the problems it forces, and the infrastructure those problems require. The killed entries are often the more useful read; they show where the reasoning broke.
- 01 Live Shift real, buyer carries the loss.
- 02 Unresolved Chain holds, the ownable position is unclear.
- 03 Killed One question broke irreparably.
Methodology
- 01
First-order
What changes immediately?
- 02
Second-order
What follows from that change downstream?
- 03
Forced problem
What new problem becomes mandatory to solve?
- 04
Infrastructure gap
What’s missing that someone has to build?
A chain is LIVE if the macro shift is real, the first two questions resolve cleanly, and the buyer is identifiable. UNRESOLVED if the chain holds but the ownable position or the buyer is genuinely unclear. KILLED when a question breaks, usually one of three ways: the second-order consequence doesn’t follow; the buyer for the missing infrastructure doesn’t exist at any price; or the choke point is already claimed and the only data left has no moat because everyone holds it.
Active dossiers
-
Dossier № 01 India tightens renewable deviation penalties faster than anyone can forecast
Macro shift India's renewable share keeps climbing, concentrated in a few states, and the regulator is squeezing tolerance to match. CERC has moved to narrow the deviation-settlement bands for wind (±15%→±10%) and solar (±10%→±5%) from 2026, and to compute deviation against schedule rather than available capacity.
- 01 First-order
Renewable generators and the industrials buying their power face materially higher penalty exposure for the same forecast error. Deviation has gone from a back-office line to a central commercial term in power-purchase agreements.
- 02 Second-order
To stay inside the band, generators need better forecasting and, more quietly, an independent account of what actually happened in each fifteen-minute block. The only attribution today comes from the load-dispatch centre, which is the same counterparty assessing the penalty.
- 03 Forced problem
The party paying the penalty has no independent way to verify the deviation it is charged for, or to contest a misattribution.
- 04 Infrastructure gap
Substation-edge frequency-and-quality telemetry, time-stamped and exposed by API to the parties carrying the financial risk, not only to the regulator who already has it.
Verdict The one that survives, and not the one I would have guessed. It clears the bar the others fail. The buyer carries the loss directly instead of sitting downstream of someone else's customer, and the data is generated at my own substation edge, privately and time-stamped, rather than pulled from a public feed everyone else also holds. The fork I still have to test is honest: the obvious spend is better forecasting, which is crowded, while the open position is independent attribution, the telemetry a penalised party uses to contest the dispatch centre's account. Whether they pay to contest or just forecast harder is the first thing to validate. But the structure is sound here where it is not anywhere else. Caveat I am keeping in view: central DSM since 2022 links charges to market price rather than raw frequency, so the frequency-linked version is sharper at the state level and for conventional plants. The verification gap holds either way. This is the chain the grid-frequency tracker was built to probe.
- 01
-
Dossier № 02 GNSS spoofing crosses from contested zones into commercial routing
Macro shift SDR-based GNSS spoofing is commodity-grade: sub-$2,000 hardware, city-scale coverage, and it has left the battlefield. In January 2026 thirteen coastal European states and Iceland issued a joint warning over interference in the Baltic and North Sea. Roughly 40% of European air traffic now operates in seriously affected airspace.
- 01 First-order
Aviation and maritime operators stop trusting raw GNSS fixes near contested coasts and ports. Behaviour changes before anything actually fails. Operators reroute, fall back to manual procedures and inertial navigation, and that carries its own cost.
- 02 Second-order
A single receiver cannot tell a coherent multi-constellation spoof from real geometry. Receiver-internal defences hit a ceiling. Trust has to come from outside the receiver.
- 03 Forced problem
Someone has to certify position independently: a source of truth for where a vessel or aircraft actually was at a given timestamp, which the spoofed receiver cannot produce for itself.
- 04 Infrastructure gap
A network of independent, time-synchronised ground sensors dense enough to localise the interference and triangulate a position consensus by time-difference-of-arrival.
Verdict Leaning killed, for a lean team. The technical gap is real, and independent position verification is genuinely needed. I just cannot find the ownable position inside it for someone small. The buyer I would reach, maritime insurers and operators, sits downstream of whoever already owns that relationship, and that is where the value settles. The government is also filling the floor for free, with the EU's interference-monitoring service going live in 2026. What is left for a small team is a government-funded utility or a feature sold to aerospace primes, and neither is the thing I would be building. The gap is real. The position probably is not mine. This is the chain closest to what the rooftop station already does, which is exactly why I had to be careful not to keep it alive out of fondness.
- 01
-
Dossier № 03 Synthetic Earth imagery gets good enough to move markets before anyone can prove a frame is real
Macro shift Open diffusion models trained on public Earth imagery now generate plausible 10-metre scenes. The threat stopped being hypothetical in early 2026. During the Gulf conflict, fabricated satellite imagery, authentic tiles manipulated with generative tools, reached millions of views before anyone identified it.
- 01 First-order
Anyone who consumes imagery as evidence (newsrooms, insurers, commodity desks, analysts) loses the ability to trust an unverified frame.
- 02 Second-order
The two obvious fixes both leak. Detection is a losing race, since generators improve continuously and classifiers always lag. Provenance signing (C2PA) is downstream of whoever owns the satellite, can be stripped by a screenshot, and can be forged with a technically valid manifest.
- 03 Forced problem
You cannot prove a frame is real from the frame alone. Trust has to come from something independent of the image: corroborating captures, a physical witness, cross-source agreement at the moment of capture.
- 04 Infrastructure gap
An independent verification layer for raw imagery: cross-source, multi-pass, possibly running over a federation of ground stations that saw the same scene.
Verdict The threat is proven now, but it surfaced in the disinformation arena, not in the commercial-imagery market this chain names as the buyer. Insurers and traders have not been forced to pay for verification yet. And the defensible position, independent ground-truth capture, is itself downstream of whoever owns the sensors, the same leak that runs through this whole page. Still stuck on whether the buy-side moves before a commercial-grade incident rather than a viral one, except the clock is visibly running now.
- 01
Killed dossiers
Chains that didn't hold. The reason is the point of the entry.
-
Dossier № 04 Power, not compute, becomes the thing that decides where AI gets built
Macro shift The binding constraint on AI infrastructure has moved from chips and capital to the physical electrical layer. US interconnection queues run four to seven years, and roughly 2,300 GW sits in queue, more than the country's entire installed capacity. High-voltage transformer lead times have gone from about two years to five. Electrical equipment is under 10% of data-center cost and effectively 100% of the bottleneck.
- 01 First-order
Hyperscalers with $650B+ in committed capex cannot energise it on schedule. A data-center announcement is a claim on power, not power itself.
- 02 Second-order
Site selection, procurement and timeline now turn on who can find and secure deliverable power fastest, a question answered today with scattered filings and tribal knowledge.
- 03 Forced problem
Developers need to know, faster than the queue reveals it, where power can actually be delivered and when.
- 04 Infrastructure gap
A live intelligence layer over interconnection queues, equipment lead times, and substation headroom.
Killed This is the most certain macro on the whole register, and I still cannot build it. Five things kill it for a small team. The intelligence wedge is crowding fast, with grid digital twins, dynamic line rating and queue-analytics shops already in it. The buyers are sophisticated enough to build it in-house. The deepest part of the bottleneck, transformers and behind-the-meter generation, is capital-heavy and not a software play. It is mostly US-shaped, which I am not. And it is the chain furthest from anything I have an edge in. A real shift that belongs to someone else. I marked this unresolved in an earlier pass because it was current and made the page look plugged-in, which was protecting the page instead of testing the chain. So I am killing it.
- 01
-
Dossier № 05 Private 'air-traffic control for space'
Macro shift Mega-constellations make collision avoidance a daily operational problem, and the intuitive answer is a neutral commercial coordinator, an ATC for orbit.
- 01 First-order
Operators need conjunction warnings and a shared picture of who is where.
- 02 Second-order
That picture requires a global sensor network of radar, optical and RF, plus continuous orbit determination.
- 03 Forced problem
A trusted, independent source of orbital truth, which no single operator can certify for the others.
- 04 Infrastructure gap
The coordination layer that sits on top of it and de-conflicts the constellations.
Killed The choke point is already claimed. LeoLabs has spent years building proprietary global radar, with US-government bookings up roughly 180% since 2024, and Slingshot, ExoAnalytic, COMSPOC, Kayhan and India's own Digantara are inside the same position. More fatally, the US government now provides a free baseline through TraCSS, as civil space-traffic coordination moves to the Department of Commerce, which caps the commercial floor. A lean team cannot build a global sensor network, and the public-data layer, the orbital elements my own tracker runs on, has no moat, because everyone holds the same elements. The gap here is capital-intensive hardware that is already being filled. Where it stays interesting is not the coordination layer but narrow attribution, tying an observed maneuver to an actor, and even that leaks unless the underlying capture is private. It folds into the standing note rather than standing on its own.
- 01
-
Dossier № 06 Residential storage attaches to Indian rooftop solar
Macro shift Rooftop solar among Indian middle-class households roughly doubles as panel prices fall, and surplus is exported to the grid under net metering.
- 01 First-order
Households install 3 to 5 kW of rooftop solar and export surplus under net metering at favourable tariffs.
- 02 Second-order
As discoms resist peak-rate buy-back and net-metering tariffs decline, homeowners look at on-site storage to time-shift their own consumption.
- 03 Forced problem
Battery vendors pitch residential 5 to 15 kWh systems on a five-to-seven-year payback.
- 04 Infrastructure gap
An integrated solar-plus-storage installer network with financing.
Killed The payback math does not survive contact with state tariff structures, which are uniformly hostile to residential storage and push real payback past twelve years in most states. The first buyer that actually pays is industrial, not residential. Folded into the grid chain (Dossier 01), where the party with both the money and the urgency actually sits.
- 01