Skip to main content
NEW BOARD: THE WORLD AFTER AI

Forecast boards

Dated calls, scored when they resolve. Open a board for the sources, refute notes, and full argument.

Browse boards
Browse boards
Lead PDF
Lead PDF

Current boards

55 calls across 10 boards. Each row gives the claim, resolution date, and probability we put on the dated version.

AI Campus Power-Claim Diligence PDF
31-page analyst dossier

AI campus power-claim diligence

Our analyst team turns announced AI campus power claims into a full primary-source dossier: verify, contest, or demote the energization story before capital treats it as real.

31-page dossier12 project rows
LeadTalen/Amazon Susquehanna is the flagship contested row, not clean proof.DocketER24-2172
CheckA claimed power-secured campus needs MW, counterparty, interconnection path, and permit status.SourcePrimary
Read dossierOpen PDF
Above-Ground Bottlenecks: The Real Critical Minerals Crisis PDF
critical minerals

Above-Ground Bottlenecks: The Real Critical Minerals Crisis

The board converges on a structural shift away from geology as the binding constraint on critical minerals supply, toward inelastic human capital shortages (P1, P3), regulatory waste disposal caps (P2), and...

4 callsPDF board
P1The US critical minerals boom stalls at the drill core: a shortage of Qualified Persons freezes project financing by 2031-2032-06-30
P2Rare earth oxide processing outside China will be capped by 2038 due to insolvable radioactive waste disposal permitting.-2038-06-30
Show 2 more calls
P3Western rare earth processing capacity will stall by 2034 due to a collapse in the pipeline of extractive metallurgists.-2034-12-31
P4Recycled Content Mandates Break the Battery Mining Boom-2035-01-01
Open board PDF
Where the constraint moves next: six inelastic inputs the market has not priced PDF
cross-domain: critical materials, energy, biotech, AI-for-science,...

Where the constraint moves next: six inelastic inputs the market has not priced

Every call here runs on one mechanism. A downstream system scales fast while one upstream input is fixed by physics, chemistry, demography, or treaty, so the rent migrates onto the input that cannot answer a...

6 callsPDF board
P1Hafnium is a byproduct-pinned trickle, now hostage to three inelastic buyers at once: nuclear control rods, advanced-node HfO2 gate dielectric, and single-crystal turbine-blade superalloys. China cut exports 90 percent in H2 2025 and the price tripled from $4,365 to roughly $13,115/kg by April 2026, but we think the long-horizon lock is the byproduct ratio itself, not the export headline.70%2032-12-31
P2Through 2034 the binding constraint on whole-mammalian-brain connectomes is not the electron microscope or the segmentation model. It is the human-proofread, error-audited gold reference set. Generative dense-segmentation auto-traces neurons faster every year but cannot self-certify, so verified ground-truth wiring becomes the scarce, slow-to-grow asset, with proofreading commanding the majority of every whole-brain project budget.46%2034-12-31
Show 4 more calls
P3By 2035 sulfuric acid, set by involuntary fossil-fuel desulfurization that shrinks as the world decarbonizes, becomes the explicitly cited binding gate on at least one HPAL battery-nickel or phosphate-fertilizer expansion, with a sustained real price step-change that outlasts the 2026 acute shocks rather than reverting with them.46%2035-12-31
P4US transmission and HVDC buildout in the 2030s binds on the narrow EHV energizing-and-commissioning craft, not the broadly-priced generic electrician shortage. The binding input is the journeyman and commissioning lead who can live-line and energize EHV lines and converter stations, and roughly half the utility line/substation cohort retires by 2030.46%2033-12-31
P5By 2034 the binding constraint on commercial DT fusion migrates one layer upstream of tritium onto blanket-grade enriched lithium-6, whose only proven industrial process (mercury COLEX) is treaty-banned and whose only current producers are Russia and China. The unpriced chokepoint is qualified, Minamata-compliant Li-6 enrichment capacity outside that duopoly, not plasma, magnets, or tritium.46%2034-12-31
P6By 2033 the binding constraint on the developed-world red-cell supply flips from donation willingness to donor-cohort biology, and O-negative red cells specifically enter chronic structural shortage, with at least one large national blood service (NHSBT, US Red Cross/ABC, or Japanese Red Cross) declaring recurring demographically-driven O-neg emergency or restricted-allocation status that is not attributable to a one-off event.46%2033-12-31
Open board PDF
The long-horizon tier PDF
cross-sector / 2032-2033

The long-horizon tier

Across all six the rent sits one layer below where capital is looking, on an input that cannot be expanded with money on the relevant timescale because the limit is an element, a reservoir, a human career, or...

6 callsPDF board
L1The grid runs out of people before it runs out of steel60%2032-12-31
L2Helium-4 is the resource that leaves and does not come back52%2032-12-31
Show 4 more calls
L3Phosphorus is essential, substitute-free, and geographically captured50%2033-12-31
L4The robot demand fixed by biology, not by adoption curves45%2033-12-31
L5If carbon storage scales at all, the rent moves to the hole in the ground40%2032-12-31
L6The energy wall forces a change of substrate, not just better chips35%2033-12-31
Open board PDF
Where chip-supply rent moves next: eight inelastic nodes, six we'd back PDF
semiconductors and the chip supply chain (logic foundry, memory and...

Where chip-supply rent moves next: eight inelastic nodes, six we'd back

All six calls share one shape. Capital, tools, and wafer capacity can all be bought on a capex clock, but the real ceiling sits a layer or two below the priced story, at an input whose supply grows on a...

6 callsPDF board
P1The ceiling on the AI-compute wafer build-out is not ASML, EUV source power, or fab shells. It is Zeiss SMT ion-beam-figuring throughput for EUV/High-NA projection mirror sets at one site in Oberkochen, Germany.52%2032-12-31
P2Senior yield and process-integration engineers with multi-node sub-5nm ramp experience are the input you cannot clone for ex-Taiwan leading-edge output through the 2030s. Capital and wafer capacity are fundable; a decade of hands-on ramp knowledge is not, and the rent moves to whoever holds that headcount.58%2033-12-31
Show 4 more calls
P3Ruthenium becomes the gating input for advanced-logic interconnects as copper hits its resistivity wall at sub-2nm pitches. Supply is hard-capped at roughly 30 tonnes per year as a platinum byproduct, with no primary mine to develop.40%2035-12-31
P4The constraint on the US/EU fab buildout is not capital or tools but the tiny, slow-to-certify pool of UHP process-piping welders and hookup fitters. Their scarcity makes the announced 2023-2030 leading-edge capacity schedule structurally undeliverable.52%2031-12-31
P5Enriched Si-28 isotope-separation cascade capacity becomes the gating substrate input for silicon-spin qubits and, if adopted, 3D-logic thermal relief. World supply sits at tens to low hundreds of kg/yr against a potential tons-per-year demand cliff by the early 2030s.34%2034-12-31
P6Multibeam mask-write throughput (IMS Nanofabrication, sole production vendor) becomes the gate on how many distinct High-NA AI-chip designs the industry can tape out per year by 2031.38%2031-12-31
Open board PDF
Where the constraint migrates in biotech: finite atoms and human donors that capital cannot replicate PDF
biotech and human health (drug modalities and delivery,...

Where the constraint migrates in biotech: finite atoms and human donors that capital cannot replicate

We make the same structural argument six times. In each case a loud, well-funded boom is paying for the visible layer (curie count, fractionation plants, targeting ligands, reconditioning devices) while the...

6 callsPDF board
P1Isotopically pure (Ac-227-free) Ac-225 from the finite Th-229/U-233 stockpile is the physics-locked binding constraint on targeted-alpha therapy through the 2030s, not raw curie count57%2032-12-31
P2By 2034, the binding constraint on the Ac-225 targeted-alpha-therapy boom is not the accelerator conversion step but the finite, non-manufacturable Ra-226 target feedstock that every accelerator route consumes, and the hot-cell capacity to fabricate, irradiate, and recover it.44%2034-12-31
Show 4 more calls
P3The binding constraint on the RLT boom migrates off the reactor onto kilogram-scale enriched stable-isotope feedstock (Yb-176, Zn-68, Gd-160) concentrated in Russian gas-centrifuge cascades, with Western rebuilds too gram-scale and too isotope-narrow to close the gap across the full therapeutic basket by 2032.38%2032-12-31
P4By 2035 the binding constraint on IVIG/SCIG supply is the human source-plasma donor base, a demographically and geographically concentrated input with no recombinant substitute; structural rationing or allocation tightens in at least two major wealthy markets, and the US share of world plasma collection stays above ~65 percent even as global IG demand grows materially.40%2035-12-31
P5Donor-derived human platelet lysate is the hidden, non-synthesizable feedstock that becomes the binding constraint on industrial-scale allogeneic cell therapy as the field migrates off FBS onto a supplement that can only be made from human blood donors -- a donor pool that is already chronically short for clinical transfusion.38%2033-12-31
P6GLP-1 mass adoption structurally degrades the deceased-donor organ supply: by 2033 the binding constraint in solid-organ transplantation becomes the vanishing pool of young, low-comorbidity SCD organs, forcing value to migrate onto organ reconditioning (normothermic machine perfusion) and xeno/engineered organ infrastructure rather than donor recruitment.38%2033-12-31
Open board PDF
Where the Space Economy's Rent Goes Next: Six Physical Walls Cheap Launch Cannot Move PDF
space and the space economy (launch vehicles and propulsion,...

Where the Space Economy's Rent Goes Next: Six Physical Walls Cheap Launch Cannot Move

We see the same trade six times, and the market misreads it the same way each time. The funded, visible bottleneck in space is always transport: launch cost, cadence, landers, GPUs. Cheap reusable launch is...

6 callsPDF board
P1Orbital compute does not hit a launch-cost wall or a power wall first. It hits a radiator-area wall. We think the binding constraint on space-based AI compute through the 2030s is deployable, micrometeoroid-survivable heat-rejection AREA per launched kilogram, governed by Stefan-Boltzmann T-to-the-fourth and a hard ~few-kg-per-square-meter panel-plus-fluid-loop mass floor that does not ride a learning curve. Multi-gigawatt orbital data centers are physically uncoverable this decade, and the rent goes to whoever owns large-area deployable-radiator and two-phase loop technology, not to whoever owns the GPUs or the launch.62%2034-12-31
P2Every kilowatt-class non-polar lunar surface asset through 2040 is gated by the 354-hour lunar night. We think the binding constraint is night-survival energy-storage mass per continuous watt, which forces fission surface power as the only viable route at useful power levels, and that route is single-sourced and schedule-fragile.70%2038-12-31
Show 4 more calls
P3By 2034, deep-space-band large-aperture antenna-hours are the binding supply constraint on missions past GEO, with demand already exceeding DSN capacity at peak and the deficit widening faster than new aperture can be fielded.62%2034-06-30
P4By 2035, the binding constraint on operating proliferated LEO is not tracking sensors or autonomous-maneuver software but the accuracy of thermospheric neutral-density forecasts, and the inelastic input gating that accuracy is real-time in-situ upper-atmosphere driver data from a tiny, non-redundant, government-funded sensor base with no commercial supply.52%2035-12-31
P5By 2035, iridium DOP-26 clad-vent-set fabrication at ORNL, not plutonium-238 fuel rate, becomes the named gating constraint on US deep-space and lunar-surface nuclear power38%2035-12-31
P6By 2038, the binding constraint on proliferated-LEO throughput is the stratosphere's finite alumina-absorption capacity, forcing a regulated migration to non-aluminum demisable structures as the scarce, rent-capturing input.52%2038-12-31
Open board PDF
The World After AI: intelligence gets permission to act PDF
post-AI world / next big thing after AI

The World After AI: intelligence gets permission to act

The post-AI world is not defined by intelligence becoming scarce. It is defined by intelligence becoming cheap enough that the bottleneck moves to permission, coupling, and verification. The winning layer is...

8 callsPDF board
P1Agent authority rails become the transaction network after AI.49%2031-12-31
P2Liability moves to the evidence recorder.51%2031-12-31
Show 6 more calls
P3Autonomous labs bottleneck on standards before autonomy.48%2031-12-31
P4Materials discovery becomes a qualification-data business.42%2035-12-31
P5Programmable-biology access becomes a licensed screening layer.46%2030-12-31
P6Human causal perturbation atlases become the drug-discovery scarce asset.34%2033-12-31
P7Grid stability replaces megawatts as the AI load gate.38%2033-12-31
P8AI campuses become heat-sink and water-rights assets.41%2032-12-31
Open board PDF
Where the Rent Moves Next: Four Binding Constraints Capital Has Not Yet Found PDF
any area, wide open across all industries

Where the Rent Moves Next: Four Binding Constraints Capital Has Not Yet Found

We see the same shift play out across all four calls, in four different physical systems. Capital and headlines crowd into the fast, fundable, capital-intensive layer (cable plants, copper mines, fusion...

4 callsPDF board
P1By 2033 the binding constraint on Western copper has moved off the mine onto qualified non-China smelting and refining capacity. A sustained zero-to-negative treatment-charge regime forces ex-China smelters into structural contraction through the late 2020s, with at least two further Western or allied smelter closures or curtailments and a formal allied policy intervention to protect what smelting remains.55%2033-12-31
P2The grid and AI-datacenter buildout gets gated through the 2030s not by transformers, steel, or switchgear, but by the certified people who specify, set, and energize substations. Settings-grade protection-and-control relay engineers and journeyman transmission linemen cannot be minted fast enough, and the pipeline to make more of them is itself broken.46%2033-12-31
Show 2 more calls
P3Lithium-6 enrichment, not the reactor, is the licensed chokepoint of the Western fusion fuel cycle. By 2034 a Western D-T pilot publicly names enriched Li-6 supply as the binding constraint, because the only scaled route (mercury COLEX) is banned under Minamata and live capacity sits only in Russia and China.42%2034-12-31
P4By 2032 the binding grid-connection constraint has moved downstream from cable manufacturing to marine installation, with the certified high-voltage offshore jointing crew as the deepest lock. At least three sanctioned HVDC links slip commissioning citing installation-vessel or jointing-crew availability (distinct from cable supply), and installation-slot lead times for HVDC campaigns exceed cable-plant lead times.46%2032-12-31
Open board PDF
After AI: where the constraint moves when intelligence leaves the screen PDF
what comes next after AI

After AI: where the constraint moves when intelligence leaves the screen

The shared shift is that AI stops being mainly a model market and becomes a physical deployment market. The obvious layers, GPUs, frontier labs, generic agents, and benchmark wins, are no longer where the edge...

6 callsPDF board
P1The AI frontier moves from model access to firm-power siting.52%2028-12-31
P2Physical AI's bottleneck is certified deployment, not robot bodies.46%2028-12-31
Show 4 more calls
P3Autonomous science shifts the bottleneck from model discovery to experimental throughput.44%2029-12-31
P4The consumer AI interface moves to always-on edge devices, gated by thermals, battery, sensors, and privacy.43%2028-12-31
P5Biomanufacturing's bottleneck is scale-up, not AI organism design.41%2030-12-31
P6Agentic AI's scarce layer becomes authority, auditability, and rollback.48%2028-06-30
Open board PDF
Three Commodity Reversals Where the Fix Is Already Paid For PDF
any area, wide open across all industries

Three Commodity Reversals Where the Fix Is Already Paid For

Our house view across this board is simple. In three separate commodity chains, the supply response the bulls are betting against has already been set off by a specific, dated event, and the market has not...

3 callsPDF board
P1Chinese polysilicon stays below cash cost through 2027 because SAMR banned the supply cut the bulls priced70%2027-12-31
P2DDR5 reverses by end-2027: a representative mainstream module (16Gb-class DDR5) falls at least 35% from its H1-2026 spot/contract peak by 2027-12-3145%2027-12-31
Show 1 more call
P3A nickel surplus headline hides a battery-grade nickel sulfate squeeze: by 2027-06-30 the sulfate premium over LME 3-month nickel (Ni-contained) widens past $4,500/t as the Indonesian acid shortage curtails MHP41%2027-06-30
Open board PDF
Analyst-team dossier

AI campus power-claim diligence

Built from public dockets and official records as a 31-page buyer-facing report on what counts as power secured.

Read the dossier

Not investment advice. These are dated forward calls; they get graded when the resolution date arrives.

Work with us

Point the engine at your mandate

The chat gives you our standing read in seconds. A commission points the whole research team and the leak-free engine at one question you actually trade on: where the binding constraint sits, whether it is already priced, and the dated signal to watch.

Built for funds and macro desks who want the call before it is consensus. You get the full reasoning and the line that would prove it wrong, so you can size it yourself.

Try the chat first →
Thesis
A dated thesis, not a take
One sub-industry, the binding constraint named, a probability with an interval, and the date it resolves.
Evidence
The data layer behind it
Research, patents, trade flows, and a live price channel, so the call rests on a constraint you can actually measure.
Watch
A kill-criterion you can watch
Every call names what would make it wrong and the signal to watch for it, so you can track the position as it plays out.

Mapping the future world

Read the public record, or commission a mandate-specific map.

See the record
Work with us
An independent forecasting instrument.
©2026 Vaticinus. All rights reserved.
Close

Work with us

Tell us what you want the engine pointed at. We read every note ourselves and reply to set up a call with the ones we can move on. Or email [email protected].

Got it. Now pick a time.

Your note is in. Grab a 30-minute call below, or we will reply from [email protected].

Something went wrong while submitting the form.
close