Wholestory

Last Updated: September 19, 2026

The Record

Latvia’s defence minister said domestically made counter-drone and radar equipment would cover the entire length of its borders with Russia and Belarus by the end of September. With days to go, neither completion nor failure has been stated: the minister’s own account in July still had radar and a command system left to buy and integrate, with negotiations underway. The claim stays open until someone says otherwise. Four new dated claims joined the record. A permanent US Army base in Poland, with a location promised soon. A Ukrainian corps commander’s forecast that Russia’s campaign would culminate in September and that it would take a Fortress Belt settlement by the end of autumn. OpenAI’s chief executive ruling out a 2026 listing. And the EPA administrator’s promise that Americans will see electricity prices fall after his agency struck the 2024 power-plant rules. One was kept. OpenAI said on September 5 that it was past time to define standards for disclosing misalignment incidents; the framework arrived on the 16th, six weeks before the deadline. What it does not contain is a trigger — which the record notes, because a promise kept and a promise that binds are different things.

The Whole Story

Public figures make checkable claims about the future constantly — capability milestones, job numbers, market calls, timelines for war and peace — and almost none are ever revisited, because the outlets that reported them have moved on by the time the due date arrives. Wholestory's stories persist, so its predictions can too. Every forward-looking, falsifiable claim captured in any topic's timeline is tagged with a theme, the observable that will decide it, and a resolve-by date.

When a prediction comes due, a hindsight cycle checks its named observable against primary data and records a verdict — true, false, mixed, or unverified — alongside the original claim, which is never edited or reinterpreted after capture. Verdicts accumulate into each voice's public track record and feed the credibility scores attached to every entity in the paper.

The house never bets. Wholestory publishes no forecasts of its own: trendlines on every page end at the last observed data point, and this board scores other people's calls only. Prevalence here reflects the sample of the conversation our cycles capture, not the conversation itself — loud voices are overrepresented, and the methodology is disclosed in this page's research brief.

Continue Reading →

Every entry below is a falsifiable claim captured verbatim in one of Wholestory's topics, scored when its due date arrives. Wholestory publishes no forecasts of its own, and this board reflects the sample of the conversation our research cycles capture — not the whole conversation.

The Board in Numbers

ThemeOpenDueResolved
bubble1801
geopolitics-war1500
warming-trajectory705
infrastructure-impact701
jobs700
energy-transition401
capability-milestones400
compute-prices200
national-advantage200
regulation-law200
adoption101
safety-incidents001
emissions-pledges100

Total claims 80

Kind 65 forecast · 15 commitment

Next due 2026-09-30 (geopolitics-war) · 2026-11-30 (geopolitics-war) · 2026-11-30 (geopolitics-war) · 2026-11-30 (geopolitics-war) · 2026-11-30 (geopolitics-war)

Americans will see a decrease in electricity prices, but this is just the beginning.?

Decided by: US average retail electricity price for all sectors, as published by the EIA, lower than the same month a year earlier at any point before 31 December 2027 — resolves by 2027-12-31

Theme: energy-transition · Commitment · From climate-change

Context: Said by the EPA administrator announcing his own agency’s repeal of the 2024 power-plant standards. A dated, checkable forecast about a price the administration influences but does not set. EIA publishes the deciding figure monthly.

it's past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models. […] We and the larger AI community do not yet have a clear standard for how to report misalignment that shows up during training, evaluation, and deployment, including examples that don't look like traditional security incidents but could provide insight into AI behavior and future risks. We're working on a framework and will share it in upcoming weeks, and in parallel we're working with dozens of government regulatory agencies worldwide on these issues.

Decided by: OpenAI publishes a framework stating when and how it discloses misalignment incidents arising in training, evaluation or deployment. — resolved (true)

Delivered early. OpenAI published ‘Our framework for reporting model misalignment’ on September 16, six weeks before the deadline, with six incident reports alongside it. The metric asked for a framework stating when and how OpenAI discloses; one exists. What it does not contain is a trigger — see the September 16 entries.

Theme: safety-incidents · Commitment · From state-of-ai/safety

Context: Delivered early. OpenAI published ‘Our framework for reporting model misalignment’ on September 16, six weeks before the deadline, with six incident reports alongside it. The metric asked for a framework stating when and how OpenAI discloses; one exists. What it does not contain is a trigger — see the September 16 entries.

It would be ‘ill-advised’ for OpenAI to go public now, given safety concerns; the company will not list in 2026.?

Decided by: Whether OpenAI completes an initial public offering on a US exchange during calendar 2026 — resolves by 2026-12-31

Theme: bubble · Commitment · From state-of-ai/the-money

Context: Said to Fortune on September 12. His chief financial officer told an August all-hands that OpenAI ‘will be a public company in 2027’, and could go sooner if the business keeps inflecting. OpenAI filed its prospectus confidentially with the SEC in June.

Progress has been made toward a permanent US Army base in Poland, which currently hosts American troops only on a rotational basis: ‘If this happens, the location will be announced very soon. This will be a historic step for our Great U.S./Polish Alliance.’?

Decided by: A specific location for a permanent (not rotational) US Army installation in Poland, named on the record by the US or Polish government, after September 17, 2026 — resolves by 2026-12-31

Theme: geopolitics-war · Commitment · From ukraine-russia-war/nato-relations

Context: Posted to Truth Social on September 17, 2026 and reported by NPR; the post returns no readable text to the public, so the wording rests on NPR's account. No location has been named. ‘Very soon’ is read here as by the end of 2026.

The Russian offensive campaign will culminate in September 2026; Russian forces will shift to mechanized breakthrough tactics by October, reinforcing roughly 200-strong assault units with columns of about 50 armoured vehicles; and Russia will try to seize at least one Fortress Belt settlement by the end of its autumn campaign.?

Decided by: Whether ISW confirms, after 11 September 2026, (a) Russian mechanized assaults of roughly battalion scale on the Fortress Belt axes, and (b) the confirmed seizure of any Fortress Belt settlement — Kostyantynivka, Druzhkivka, Kramatorsk or Slovyansk — resolves by 2026-11-30

Theme: geopolitics-war · Forecast · From ukraine-russia-war/territory-control

Context: A Ukrainian corps commander forecasting an adversary he does not control, said on September 11. He separately said about 1,200 Rubikon drone crews, a third of that unit, are in the Dobropillya direction — a Ukrainian claim about Russian dispositions, not geolocated, and not placed on the map.

We expect the war will continue — Ukraine is counting on US energy and air-defence support this winter, and does not expect an increase in Patriot supplies until next year.?

Decided by: Whether active hostilities between Russia and Ukraine are still under way on 2026-12-31 — no ceasefire or peace agreement in effect between the two governments — as reported by wire services or confirmed by either government. — resolves by 2026-12-31

Theme: geopolitics-war · Forecast · From ukraine-russia-war/peace-negotiations

Context: Unverified — the horizon has not passed. “Into the winter months” is read concretely as hostilities still under way on 31 December 2026, which bounds the claim at both ends. Said in Kyiv on 6 September, hours after the US envoys' first visit (Kyiv Independent).

China is on track to add enough wind, solar, nuclear and hydropower this year to cover electricity demand growth, despite a slowdown in new capacity.?

Decided by: Combined 2026 growth in Chinese generation from wind, solar, nuclear and hydro at least equalling the 2026 growth in total electricity generation, per National Bureau of Statistics and China Electricity Council annual data. — resolves by 2027-03-31

Theme: energy-transition · Forecast · From climate-change

Context: Not yet checkable. It settles on full-year 2026 Chinese generation data: whether the combined increase in wind, solar, nuclear and hydro output at least matches the increase in total electricity generation. Clean capacity additions have slowed this year, and the first quarter went the other way.

In 2027, we have secured the supply to again double AI revenue to approximately $115 billion. Our demand actually exceeds this outlook, and we will work to improve supply. In 2028, we expect the trajectory of growth to continue. We have line of sight for fiscal 2028 AI semiconductor revenue growth to again double to $230 billion.?

Decided by: Broadcom's reported AI semiconductor revenue for fiscal 2027, as stated in its own fourth-quarter and full-year results. Resolves true if it reaches approximately $115bn (a band of $109bn-$121bn, +/-5%), false if it falls short of $109bn. — resolves by 2027-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Newly recorded; resolves when Broadcom reports fiscal 2027. Not in the results release and not formal guidance — a verbal multi-year outlook the company said it will not update quarterly. An analyst on the same call derived $11-12bn of content per gigawatt from Broadcom's own figures, against the $20-30bn Tan asserted.

Serhiy 'Flash' Beskrestnov, Ukraine's presidential adviser on defence technologies, said Russia's next step is a cheap jet-powered missile — the Geran-5, Banderol, Dan-T, Dan-M and the jet UMPB-5R glide bomb — pushing speeds from the Geran-4's 300-500 km/h up to 500-650 km/h: 'That is effectively a missile, and it will be very hard for an interceptor to work against such a target.'?

Decided by: Documented cruise speeds of Russian jet-powered one-way attack munitions (Geran-5, Banderol, Dan-T/M, UMPB-5R) in Ukrainian Air Force reporting or specialist analytical assessment reaching 500-650 km/h — resolves by 2027-03-31

Theme: geopolitics-war · Forecast · From ukraine-russia-war/tactics

Context: Not yet checkable. The named systems are already in use at lower speeds — ISW records Banderol and S8000 in the September strike packages — but no 500-650 km/h figure has been independently measured.

Russia's political leadership has ordered its forces to reach the administrative borders of Donetsk and Luhansk oblasts by 31 December 2026; the military leadership replied that the order was unrealistic and asked for 31 March 2027 instead. Neither date is achievable given the battlefield situation.?

Decided by: Whether Russian forces control the full administrative territory of both Donetsk and Luhansk oblasts, as assessed by ISW and DeepState control mapping, on 31 December 2026. — resolves by 2026-12-31

Theme: geopolitics-war · Forecast · From ukraine-russia-war/territory-control

Context: Recorded as a prediction resolving 31 December 2026. Palisa reports Russian internal deliberation at second hand and is its sole source. ISW says Russian forces will not meet the deadline: Putin has set 15 Donbas deadlines since 2022 and met none, and at August's Donetsk rate of 2.36 square kilometres a day ISW puts completion in October 2032 — if it is possible at all.

Current monitoring shows that local air quality continues to meet or exceed EPA standards.?

Decided by: SpaceXAI removes all 69 temporary gas turbines from its Southaven site — resolves by 2027-07-31

Theme: infrastructure-impact · Commitment · From state-of-ai/infrastructure/on-site-power

Context: Still unverified. The air-quality claim is the company’s own assessment, with no independent monitoring on the record, against DeSoto County’s ‘F’ ozone grade and the emissions Rep. Pallone cited on 28 July. The removal commitment has partial evidence — 11 of 69 turbines off by 31 August — and resolves by July 2027.

El Niño will strengthen to a very strong event peaking towards the end of 2026, and forecasts put at nearly 100 percent the likelihood it persists through February 2027.?

Decided by: Peak three-month Oceanic Niño Index reaching the very-strong threshold (2.0°C or higher) and continuous El Niño conditions persisting through February 2027, per NOAA CPC / WMO ENSO monitoring. — resolves by 2027-02-28

Theme: warming-trajectory · Forecast · From climate-change

Context: Not yet checkable. It settles once the peak three-month Oceanic Niño Index is known and whether El Niño ran unbroken through February 2027; 'very strong' conventionally means an ONI of 2.0°C or higher. WMO calls this its most confident El Niño outlook ever.

The Russian General Staff has prepared a scenario for the additional mobilisation of 600,000 people across 2026 and 2027 — 300,000 in each year — chiefly to bring units fighting Ukraine up to strength; Putin will most likely decide after the 18–20 September State Duma elections.?

Decided by: A formal Russian mobilisation decision or announced wave totalling approximately 600,000 or more, taken across calendar 2026 and 2027 combined. Bounded at both ends: decisions before 2026-08-27, the date of the claim, do not count, and the count closes 2027-12-31. Distinct from the 2026-08-23 forecast of ~300,000 after the September elections — a single 300,000 announcement resolves that one and not this one. — resolves by 2027-12-31

Theme: geopolitics-war · Forecast · From ukraine-russia-war/home-front

Context: A belligerent's intelligence services briefing on the other side's internal planning, recorded as their claim. The supporting figures they gave — 240,000 Russian personnel lost January–July against 232,000 contract recruits, convict-recruitment categories widened, 15–60 days' training — are part of the claim, not corroboration of it.

Revenue for the third quarter of fiscal 2027 is expected to be $108.0 billion, plus or minus 2 percent, with no Data Center compute revenue from China assumed.?

Decided by: NVIDIA's reported revenue for the third quarter of fiscal 2027, as published in its own results release. Resolves true if reported revenue falls within $105.84bn-$110.16bn (the stated ±2% band), false if outside it. — resolves by 2026-12-15

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Newly recorded; resolves when NVIDIA reports Q3 FY2027, expected in November. Formal guidance, so it is scored as a forecast rather than a commitment. The company met or beat its own guide in the preceding quarter, which reported $96.2bn.

Russia could ask North Korea to provide up to 50,000 additional troops during a potential September meeting between Russian President Vladimir Putin and North Korean leader Kim Jong Un.?

Decided by: A publicly confirmed North Korean commitment to deploy approximately 50,000 or more additional troops to Russia, agreed or announced between 2026-09-01 and 2026-12-31. Confirmation means a statement by either government or a Western intelligence assessment; a Ukrainian intelligence claim alone does not resolve it. — resolves by 2026-12-31

Theme: geopolitics-war · Forecast · From ukraine-russia-war/home-front

Context: Newly recorded; resolves 2026-12-31. Ukraine's military intelligence says 8,500 North Koreans have arrived in Kursk Oblast to hold rear positions and free Russian troops for assaults. Kim Yo Jong called the larger figure groundless on 19 August. A belligerent's claim that its enemy cannot man its own front.

The Kremlin is preparing to mobilise around 300,000 servicemembers after the September 18–20 State Duma elections, in part to reinforce Russian positions in Donetsk and Luhansk oblasts.?

Decided by: A formal Russian mobilisation decision or publicly announced wave of approximately 300,000 or more servicemembers, taken between 2026-09-21 and 2026-12-31. Bounded at both ends: decisions before the Duma elections, or after 2026, do not count. — resolves by 2026-12-31

Theme: geopolitics-war · Forecast · From ukraine-russia-war/home-front

Context: Newly recorded; resolves 2026-12-31. A Ukrainian head of state's intelligence claim about Russian intentions — recorded as a voice, not as established fact. Russia has announced one partial mobilisation since 2022 and has since relied on contract recruitment; ISW carries the claim without endorsing it.

We believe that this is the most active stage at which diplomacy can lead to the end of the war. It's the end of this year and the end of summer 2027.?

Decided by: A negotiated agreement ending the Russia-Ukraine war — a ceasefire or peace settlement signed by both Russia and Ukraine — concluded between 2026-11-01 and 2027-09-30. Bounded at both ends: agreements before the window, or after it, do not count. — resolves by 2027-09-30

Theme: geopolitics-war · Forecast · From ukraine-russia-war/peace-negotiations

Context: Newly recorded; resolves on 2027-09-30. "End of this year" is read as from 1 November 2026 and "end of summer 2027" as 30 September 2027. Said at the National Flag Day ceremony in Kyiv, with the two sides still divided over Ukrainian withdrawal from Donbas.

We estimate that hyperscalers could issue USD 300 billion in investment-grade bonds over the next year alone.?

Decided by: Aggregate US hyperscaler investment-grade bond issuance between 2026-08-20 and 2027-08-20, measured from a market tracker such as SIFMA, LSEG or Bloomberg league tables. Bounded at both ends; issuance before or after the window does not count. — resolves by 2027-08-20

Theme: bubble · Forecast · From state-of-ai/the-money/debt-and-credit

Context: Newly recorded; resolves 2027-08-20. The same note estimates hyperscalers could add $1.5 trillion of debt in total before lease-adjusted leverage reaches the investment-grade industrials average — a capacity estimate, not a dated forecast, and not scored. J.P. Morgan Asset Management sells the credit it is analysing, so the estimate is recorded as an interested party's.

The earth is warmer in 1988 than at any time in the history of instrumental measurements... With about 99 percent confidence we can state that the warming during this time period is a real warming trend... In my opinion, the greenhouse effect has been detected, and it is changing our climate now.±

Decided by: Whether subsequent decades confirmed the 1980s warming Hansen ascribed to the greenhouse effect as a real, sustained human-caused trend, and the magnitude of his intermediate Scenario B trajectory against observed global temperature. — resolved (mixed)

Detection vindicated: the 1980s warming was real and human-caused, as every later assessment confirmed. His intermediate Scenario B ran a little hot — observed warming tracked close to but somewhat below it, about 30% slower over 1970–2016 than Scenario B projected, mainly because the model's 4.2°C sensitivity was high and the halocarbons it assumed kept growing were later cut by the Montreal Protocol (Carbon Brief/Hausfather).

Theme: warming-trajectory · Forecast · From climate-change

Context: Detection vindicated: the 1980s warming was real and human-caused, as every later assessment confirmed. His intermediate Scenario B ran a little hot — observed warming tracked close to but somewhat below it, about 30% slower over 1970–2016 than Scenario B projected, mainly because the model's 4.2°C sensitivity was high and the halocarbons it assumed kept growing were later cut by the Montreal Protocol (Carbon Brief/Hausfather).

Under the IPCC Business-as-Usual (Scenario A) emissions... a rate of increase of global mean temperature during the next century of about 0.3°C per decade (with an uncertainty range of 0.2°C to 0.5°C per decade); this is greater than that seen over the past 10,000 years.±

Decided by: Observed global mean surface temperature trend per decade from 1990 across the major records, against the FAR's 0.3°C/decade central estimate and its 0.2–0.5°C/decade range. — resolved (mixed)

The central estimate ran high: warming since 1990 has averaged about 0.2°C per decade, at the bottom of the stated 0.2–0.5 range and roughly a third below the 0.3 central. The First Assessment Report overestimated the 1970–2016 warming rate by about 17% because its business-as-usual case assumed faster CO₂ and halocarbon growth than occurred — but observed warming stayed inside its stated range (Carbon Brief/Hausfather).

Theme: warming-trajectory · Forecast · From climate-change

Context: The central estimate ran high: warming since 1990 has averaged about 0.2°C per decade, at the bottom of the stated 0.2–0.5 range and roughly a third below the 0.3 central. The First Assessment Report overestimated the 1970–2016 warming rate by about 17% because its business-as-usual case assumed faster CO₂ and halocarbon growth than occurred — but observed warming stayed inside its stated range (Carbon Brief/Hausfather).

For the mid-range IPCC emission scenario, IS92a, the model-projected global mean warming is about 2°C by 2100. This estimate is approximately one third lower than the “best estimate” in 1990.?

Decided by: Global mean surface warming by 2100 under mid-range (IS92a) emissions against the SAR's ~2°C best estimate; interim check = observed 1990-onward warming rate vs the SAR's projected rate. — resolves by 2100-12-31

Theme: warming-trajectory · Forecast · From climate-change

Context: Not due until 2100. On the interim record the report ran cool: after cutting the 1990 estimate by a third — adding aerosol cooling and a lower 2.5°C sensitivity — its projected warming came in about 28% slower than observed over 1990–2016, the series' steepest under-shoot (Carbon Brief/Hausfather).

EU sanctions have already cost Russia dearly, depriving Russia's war machine of over EUR 1 trillion ($1.16 trillion) and for autumn I am putting forward the most far-reaching sanctions listings since the start of the war. Once adopted, they would immediately raise the total number of sanctioned Russian entities by a third.?

Decided by: Whether the EU adopts new Russia sanctions listings — on or after 2026-08-17 and on or before 2026-11-30 — that raise the total number of sanctioned Russian entities by approximately one third, as recorded in the Official Journal of the European Union or the Council's own press releases. Only adoptions dated after the claim count; announcements, Coreper presentations and proposals do not. — resolves by 2026-11-30

Theme: geopolitics-war · Forecast · From ukraine-russia-war/home-front

Context: Unverified — the listings do not yet exist. EU services are preparing roughly 1,600 designations for member-state ambassadors in early September, with adoption targeted for October, so the "by a third" claim is checkable only on adoption. The EUR 1 trillion figure is Kallas's own assertion, not an audited total; "autumn" is read here as on or before 30 November 2026.

The drone shot down over Latvia on August 14 entered from Belarus, and Latvia will keep intercepting and destroying drones that violate its airspace — but it does not yet have enough counter-drone systems or radar along its border, and will deploy domestically produced equipment across its entire frontier with Russia and Belarus by the end of September.?

Decided by: Confirmation from Latvia's Ministry of Defence or National Armed Forces that domestically produced counter-drone and radar equipment is deployed along the full length of Latvia's borders with Russia and Belarus, on or before 30 September 2026. — resolves by 2026-09-30

Theme: geopolitics-war · Commitment · From ukraine-russia-war/nato-relations

Context: Not yet due. Latvia's defence ministry confirmed the August 14 intercept and the reinforcement of eastern-border air defence, but has published no coverage map for the counter-drone and radar rollout. The Belarus entry point is Melnis's own attribution; the armed forces' same-day account named only Russian electromagnetic interference and did not identify the drone.

Given the estimated trend and the volume estimate for October–November of 2007 at less than 9,000 km3, one can project that at this rate it would take only 9 more years or until 2016 ± 3 years to reach a nearly ice-free Arctic Ocean in summer. Regardless of high uncertainty associated with such an estimate, it does provide a lower bound of the time range for projections of seasonal sea ice cover.

Decided by: Whether the Arctic September sea-ice minimum falls below the 1 million sq km “ice-free” threshold in any year through 2019, per NSIDC. — resolved (false)

False. Nothing close has happened: NSIDC put the September 2025 minimum at 4.60m sq km, tenth lowest in the satellite record and still more than four times the 1m sq km “ice-free” threshold. The projection was offered as a hedged lower bound rather than a central estimate — a caveat the coverage of it did not always carry.

Theme: warming-trajectory · Forecast · From climate-change

Context: False. Nothing close has happened: NSIDC put the September 2025 minimum at 4.60m sq km, tenth lowest in the satellite record and still more than four times the 1m sq km “ice-free” threshold. The projection was offered as a hedged lower bound rather than a central estimate — a caveat the coverage of it did not always carry.

The global mean surface temperature change for the period 2016–2035 relative to 1986–2005 will likely be in the range of 0.3°C to 0.7°C (medium confidence).?

Decided by: The observed 2016–2035 mean global surface temperature minus the 1986–2005 mean, across the major surface records (HadCRUT5, NOAA, GISTEMP, Berkeley Earth, ERA5). — resolves by 2036-06-30

Theme: warming-trajectory · Forecast · From climate-change

Context: Not yet due — the window does not close until the 2035 average is in. Observations through 2022 sit in the middle-to-warmer half of the projected range, which Ed Hawkins, revisiting the projection in 2023, judged still reasonable though perhaps cautious at its lower bound.

The United States will reduce its emissions by 26–28% below 2005 levels by 2025, and make best efforts to reduce by 28%.?

Decided by: US net greenhouse gas emissions in 2025 versus 2005, per the EPA Inventory of US Greenhouse Gas Emissions and Sinks covering calendar 2025. — resolves by 2027-04-15

Theme: emissions-pledges · Commitment · From climate-change

Context: Not yet checkable. The pledge is written against the national greenhouse-gas inventory, and the edition covering calendar year 2025 is not published until spring 2027.

Arctic sea-ice extent in September 2015 will be 0.98 million square kilometres.

Decided by: Observed September 2015 Arctic sea-ice extent, per the Sea Ice Prediction Network / NSIDC record. — resolved (false)

False by a factor of about 4.7: September 2015 ended with 4.6m sq km of ice, the fourth lowest on record at the time. The forecast was filed formally with the Sea Ice Prediction Network, but the method behind it was never published.

Theme: warming-trajectory · Forecast · From climate-change

Context: False by a factor of about 4.7: September 2015 ended with 4.6m sq km of ice, the fourth lowest on record at the time. The forecast was filed formally with the Sea Ice Prediction Network, but the method behind it was never published.

Renewables will overtake coal to become the world’s number one source of electricity in the early 2030s.

Decided by: The first calendar year in which renewables exceed coal as a share of global electricity generation, per Ember / IEA full-year data. — resolved (false)

False, and in the direction the outlook was criticised for at the time: renewables passed coal in 2025, supplying 33.8% of world electricity against coal’s 33.0%, at least five years before the projected window opened. The figure came from a central scenario that assumed governments would miss many of their own clean-energy targets.

Theme: energy-transition · Forecast · From climate-change

Context: False, and in the direction the outlook was criticised for at the time: renewables passed coal in 2025, supplying 33.8% of world electricity against coal’s 33.0%, at least five years before the projected window opened. The figure came from a central scenario that assumed governments would miss many of their own clean-energy targets.

Next year or the year after that, I think it will be free of ice in summer and by that I mean the central Arctic will be ice-free. You will be able to cross over the north pole by ship.

Decided by: Whether the Arctic September sea-ice minimum falls below the 1 million sq km “ice-free” threshold in 2017 or 2018, per NSIDC. — resolved (false)

False. Nine summers on the central Arctic is still frozen: NSIDC recorded a September 2025 minimum of 4.60m sq km, tenth lowest in the satellite record. When the claim was made, none of the 40 groups filing with the Sea Ice Prediction Network expected even a record low that year.

Theme: warming-trajectory · Forecast · From climate-change

Context: False. Nine summers on the central Arctic is still frozen: NSIDC recorded a September 2025 minimum of 4.60m sq km, tenth lowest in the satellite record. When the claim was made, none of the 40 groups filing with the Sea Ice Prediction Network expected even a record low that year.

Solar will overtake nuclear, wind, hydro, gas and, finally, coal, to become the world’s single-largest source of electricity by 2033.?

Decided by: Whether solar is the largest single source of global electricity generation in 2033, per Ember / IEA full-year data. — resolves by 2034-12-31

Theme: energy-transition · Forecast · From climate-change

Context: Not yet due. Solar passed wind for the first time in 2025 and closed on nuclear, reaching 2,778 TWh — but coal still supplied 10,476 TWh, so the projection requires solar to roughly quadruple within eight years.

The world is very likely to breach the Paris Agreement’s 1.5°C temperature limit this decade.?

Decided by: Whether the assessed long-term human-induced warming level exceeds 1.5°C above 1850–1900 by the end of 2030, per IPCC/WMO assessment rather than any single calendar year. — resolves by 2031-06-30

Theme: warming-trajectory · Forecast · From climate-change

Context: Not yet settled, and the wording admits two readings. A single calendar year above 1.5°C has already happened, in 2024; the Paris limit is defined against a multi-decade average, which is the harder test this is scored against.

Even 2025 projects planned two to three years ago struggled to come online, so 2026 projects face even steeper challenges.?

Decided by: A majority of large data centers scheduled for 2026 come online on their originally announced timelines — resolves by 2026-12-31

Theme: infrastructure-impact · Forecast · From state-of-ai/infrastructure

Context: Still open, and this cycle's record runs its way. Hut 8's '1 GW' Beacon Point campus carries 1,000 MW of secured capacity against 704 MW of signed IT leases, and Southern Company reports a 75 GW pipeline of which 10 GW is contracted. In August ERCOT and data-center representatives said plainly that many of the 1,800-plus projects in the Texas queue will never be built for want of financing or a tenant. Resolves on how much of the 2026 pipeline energizes by year end.

GPU rental pricing is more likely to keep rising than falling through 2026.?

Decided by: H100/GPU cloud rental rates end 2026 higher than they began the year — resolves by 2026-12-31

Theme: compute-prices · Forecast · From state-of-ai/infrastructure

Context: Still open, and no longer clearly running in the predicted direction: the Silicon Data H100 index read $2.53 per GPU-hour on 6 September, below the $2.75 it showed on 27 July; late 2025 was near $1.96. Resolves on where the rate ends 2026.

Graphics memory costs will rise further and substantially increase graphics card costs in the second half of the year, with entry-level cards facing more severe shortages and availability declining further.?

Decided by: Graphics card average selling prices and entry-level scarcity both increase over the second half of 2026 — resolves by 2026-12-31

Theme: compute-prices · Forecast · From state-of-ai/infrastructure

Context: A forward call in the direction the company benefits from, running with the trackers: TrendForce puts server DRAM contract prices up 13 to 18 percent in the third quarter. The wording is transcribed by a trade outlet, not checked against the filing. Resolves on card prices at end-2026.

Is this circular financing? No. OpenAI will pay the lease.?

Decided by: Whether NVIDIA discloses a Trigger Event under the PORTS-Pike residual value guaranties — OpenAI insolvency causing a lease default, or OpenAI's failure to make lease payments, as defined in the 17 August 2026 Form 8-K — in any SEC filing between the first lease commencement (expected 2028) and 31 December 2030 — resolves by 2030-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Not checkable before the leases commence, expected from 2028. Recorded as a forecast rather than a commitment: it is a claim about OpenAI's future conduct, which Nvidia does not control, even though Nvidia is the counterparty that pays if it fails. The same day's 8-K defines Nvidia's obligation entirely by the event Huang says will not occur.

While the exact magnitude and mix of future debt issuance from the hyperscalers is uncertain, our review of management commentary leaves us expecting a growing role for debt financing in the AI buildout in the years ahead.?

Decided by: Whether investment-grade bond issuance by the five largest AI spenders reaches roughly $250 billion in 2026 and $400 billion in 2027, and whether debt-funded capital expenditure passes a third of the total — resolves by 2027-12-31

Theme: bubble · Forecast · From state-of-ai/the-money/debt-and-credit

Context: Forward forecast, checkable against issuance and filed capex for 2026-2027. The 2026 leg is firming: about $194 billion issued in the first half against a $250 billion full-year forecast, and BNY independently expects the full year to exceed $250 billion, with hyperscalers already over 12% of a record $1.5 trillion of US investment-grade issuance to 12 August. The debt-share-of-capex limb remains unmeasured. Goldman's note was read via Yahoo Finance rather than directly.

The SEC's position lets sponsors of data-centre securitisations hold progressively less of their own equity in these deals, and will push the industry toward more of them.?

Decided by: Whether data-centre securitisation issuance in 2027 exceeds its 2026 level, and whether sponsors' retained equity in new deals falls below the 5% risk-retention floor that applied before the SEC staff's position — resolves by 2027-12-31

Theme: bubble · Forecast · From state-of-ai/the-money/debt-and-credit

Context: Forward call made the week the reading became public; not yet checkable. The direction is plausible on the record — data-centre asset-backed issuance rose from $2.4bn in 2020 to $15.5bn in 2025 and is pacing to a 2026 record — but that growth predates the SEC's position, so 2027 issuance and disclosed retained-equity levels in new deals are what settle it.

No AI-chip export policy to China can be “permissive, implementable, enforceable, and protective of US national security” at once.?

Decided by: Whether the January 2026 case-by-case export framework proves enforceable (its 50%-cap, end-use, and third-party-lab conditions hold) or leaks/expands toward more capable chips — resolves by 2027-12-31

Theme: national-advantage · Forecast · From state-of-ai/international-competition

Context: Enforceability claim still open before the 2027-12-31 horizon, but the agency itself has now recorded a gap: BIS's FY2025 annual report to Congress says that "in certain areas, there remains a lack of alignment between U.S. controls and those of partners and allies," naming semiconductor manufacturing equipment as a particular concern. The August compute-rental review — into conduct the rules may not reach at all — is the next milestone.

With CO₂ emissions still increasing, keeping global warming below 1.5°C is no longer plausible; the remaining carbon budget for 1.5°C — about 170 billion tonnes of CO₂ — will be gone before 2030 at the current emission rate.?

Decided by: Whether cumulative global CO₂ emissions (fossil + land-use, per the annual Global Carbon Budget) from 1 Jan 2025 exceed 170 Gt CO₂ before 1 Jan 2030 — i.e. the stated 1.5°C budget is exhausted this decade. — resolves by 2030-01-01

Theme: warming-trajectory · Forecast · From climate-change

Context: Recorded as a falsifiable claim about the near future; resolvable against cumulative emissions accounting before 2030. Spoken by Prof Pierre Friedlingstein presenting the Global Carbon Budget and attributed here to the institution.

Both solar and wind are each expected to overtake nuclear power in global electricity generation in 2026.?

Decided by: Whether 2026 global solar generation AND 2026 global wind generation each exceed 2026 global nuclear generation, per Ember / Energy Institute full-year data. — resolves by 2027-04-30

Theme: energy-transition · Forecast · From climate-change

Context: A forward-looking energy-system forecast recorded when made; resolvable from full-year 2026 generation data in 2027.

2026 will finish around 1.51°C above pre-industrial levels — the second-warmest calendar year on record behind 2024 — with a 35% chance of setting a new annual record and a 63% chance of exceeding 1.5°C.?

Decided by: 2026 calendar-year global mean surface temperature anomaly vs 1850–1900, six-dataset mean (NASA, NOAA, HadCRUT5, Berkeley Earth, ERA5, JRA-3Q): central sub-claim resolves if 2026 lands 1.45–1.57°C and second-warmest; the record sub-claim resolves on whether 2026 exceeds 2024. — resolves by 2027-02-28

Theme: warming-trajectory · Forecast · From climate-change

Context: A forward-looking projection recorded when made; resolvable once full-year 2026 temperature data close in early 2027.

2027 will be around 1.7°C above pre-industrial levels, comfortably setting a new record for the warmest year, as the developing El Niño has its largest impact that year.?

Decided by: 2027 calendar-year global mean surface temperature anomaly vs 1850–1900, six-dataset mean: resolves true if 2027 exceeds the prior record (2024, ~1.52°C) and lands near 1.7°C. — resolves by 2028-02-28

Theme: warming-trajectory · Forecast · From climate-change

Context: A forward-looking projection recorded when made; resolvable once full-year 2027 temperature data close in early 2028.

To be clear, the $100 billion ARR in December is not a question mark. That's what we would achieve if we basically did nothing.?

Decided by: SpaceX's annualised recurring revenue (ARR) reaching $100 billion, as disclosed by the company in or around December 2026 — resolves by 2026-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: A forward claim about SpaceX's own revenue, recorded as a forecast (the weaker inference about a speaker forecasting a company he controls). It resolves when SpaceX next discloses an annualised-recurring-revenue figure at or after December 2026: at least $100 billion confirms it. Musk's own framing conditions the number on closing the $60 billion Cursor acquisition, so a resolution should note whether that deal closed. Unverified at capture.

Named investors called AI-era markets a bubble: Ray Dalio saw 'classic signs' of one reminiscent of 1929 and 2000 and agreed with Jeremy Grantham that this is 'the biggest investment bubble in American history'; Goldman Sachs's Peter Oppenheimer partly dissented — 'there does not appear to be a valuation bubble … there may be an earnings bubble' — while Apollo's Torsten Slok declared 'the 60/40 portfolio is broken.'?

Decided by: Whether the AI-era equity-issuance surge marks a market top comparable to 1929 and 2000 — tracked against the IPO calendar (SpaceX's post-IPO share price, and the pricing and aftermarket performance of the Anthropic and OpenAI roughly $1 trillion listings) rather than a fixed deadline — resolves by 2027-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: The speakers set no fixed date; this standing bubble call is tracked against its named indicator, the equity-issuance and IPO surge — SpaceX (public and trading below its first-trade price), Anthropic and OpenAI (both reported eyeing near-$1 trillion listings). The 2027-12-31 resolveBy is a charitable concrete reading, not the speakers' own. Unverified at capture.

Ukraine will jointly produce with Raytheon one of the most critical air defense systems — interceptors for the Patriot system?

Decided by: Whether a signed licence or production contract for Ukrainian manufacture of Patriot-family interceptors or their components is confirmed by RTX/Raytheon, the US State Department or the US Department of Defense, and whether any such interceptor or component is manufactured in Ukraine — evidenced by a company or US government release, or a Ukrainian government production announcement corroborated by a wire service. — resolves by 2028-07-31

Theme: geopolitics-war · Commitment · From ukraine-russia-war/weapons-procurement

Context: Unverified — a forward-looking claim, recorded as a prediction running to July 2028. No contract, licence, value or date has been published by either company or by the US government, and Raytheon's own president of land and air defence systems said at Farnborough on 20 July that the licensing process 'has just started'. Trump's 31 July withdrawal of the enabling licence pledge makes the affirmative outcome materially less likely without resolving it.

A Supreme Court decision on AI training and copyright is essentially inevitable, with some publishers settling but enough litigating to force resolution of the novel question.?

Decided by: A U.S. Supreme Court merits decision squarely resolving whether unlicensed use of copyrighted works to train generative AI is copyright infringement or fair use. — resolves by 2031-12-31

Theme: regulation-law · Forecast · From state-of-ai/law-and-governance/ai-copyright

Context: A forward-looking legal prediction made at the NYT filing (resolveBy 2031). As of this cycle no US Supreme Court merits decision on AI-training copyright has issued; the district-court record is still developing (Thomson Reuters/ROSS, Bartz, Kadrey). The 31 July Munich ruling against Suno is a foreign first-instance decision on German and EU law and is not evidence either way for the US Supreme Court question this prediction turns on. Not due; unresolved.

In most cases it will likely be illegal to copy copyrighted works to train generative AI without permission, because AI trained on those works can generate countless competing works and thereby dilute the market for the originals — so companies will generally need to pay for training rights.?

Decided by: The prevailing outcome in US courts of AI-training fair-use disputes: whether unlicensed copying of copyrighted works to train generative AI is generally held to be infringement (requiring licensing) rather than fair use. — resolves by 2030-12-31

Theme: regulation-law · Forecast · From state-of-ai/law-and-governance/ai-copyright

Context: Dicta in the Kadrey opinion articulating a ‘market dilution’ theory and predicting that unlicensed training will generally be held to require a licence in US courts (resolveBy 2030). As of this cycle US case law is split (contrast Bartz) and no appellate resolution has issued. The 31 July Munich ruling against Suno reaches that result, but under German and EU law, not US fair use — a foreign decision is not evidence for the specifically-US question this prediction turns on. Not due; unresolved.

We have clear intelligence that Russia is preparing a new and quite significant wave of mobilization for the fall. Currently, the Russians are losing more troops on the front than they are recruiting.?

Decided by: Whether Russia announces a formal wave of mobilization or involuntary reserve call-ups — a presidential decree or an equivalent published instrument, not a recruitment drive or an end-strength adjustment — on or before 2026-11-30, as reported by wire services or published on pravo.gov.ru. — resolves by 2026-11-30

Theme: geopolitics-war · Forecast · From ukraine-russia-war/home-front

Context: Unverified — the horizon has not passed. "The fall" is read as on or before 30 November 2026, which fixes the resolveBy. No mobilisation decree has issued; the 27 July decree raising authorised end strength to 2,426,130 is a different instrument and does not resolve it. ISW expects Putin to wait until after the mid-September Duma elections.

Trilateral negotiations between Ukraine, the United States and Russia will be held in the autumn of 2026 — though if the timing were up to me they would happen in July.?

Decided by: Whether a negotiating session with official delegations from all three of Ukraine, the United States and Russia convenes on or before 2026-11-30, as reported by wire services or confirmed by any of the three governments. — resolves by 2026-11-30

Theme: geopolitics-war · Forecast · From ukraine-russia-war/peace-negotiations

Context: Unverified — the horizon has not passed (“autumn 2026” read as on or before 30 November). The metric has no lower bound, so three rounds held before the claim — Abu Dhabi twice, Geneva in February — satisfy it literally; the resolving cycle counts only sessions convening after 23 July 2026, and says so.

Electricity bills for American families will actually come down. They're going to have a lot of electricity left over, and they'll put that into the grid.?

Decided by: Average U.S. residential electricity bills fall from their 2026 level — resolves by 2028-12-31

Theme: infrastructure-impact · Forecast · From state-of-ai/infrastructure

Context: Not yet checkable on a 2028 horizon, and the evidence runs the other way: ICF projects monthly bills up 15 to 40 percent by 2030, and PJM’s market monitor attributes $6.3bn of one auction’s $16.4bn in capacity charges to data centres. Resolves on EIA residential price data.

Data-centre load left in PJM's capacity auctions has increased capacity costs by $29.4 billion over the last four auctions, and that number will accelerate in future auctions if PJM succeeds in increasing the maximum price in the auctions.?

Decided by: Capacity charges in PJM auctions attributable to data-centre demand, as calculated by Monitoring Analytics, for auctions held after July 2026 compared with the four preceding auctions. — resolves by 2027-12-31

Theme: infrastructure-impact · Forecast · From state-of-ai/infrastructure/grid-and-ratepayers

Context: Unverified. The $29.4bn is now published with its method — $29,368,774,465 across four auctions, in the monitor's Q2 2026 State of the Market Report — but it remains the monitor's own calculation, uncorroborated. The forward half, acceleration under a higher price cap, awaits the December 2026 backstop auction.

Sell China the fourth-best chip so “their developers get addicted to the American technology stack.”?

Decided by: Whether selling China throttled U.S. chips (H20-class) keeps Chinese AI developers dependent on the American stack, versus accelerating adoption of domestic alternatives (e.g. Huawei Ascend) — resolves by 2027-12-31

Theme: national-advantage · Forecast · From state-of-ai/international-competition

Context: Forecast still open before the 2027-12-31 horizon, but the early evidence cuts against it: by mid-2026 the reported dependency ran partly the other way, with nearly 200 U.S. startup founders and 20-plus firms (Nvidia, Microsoft, Meta, Palantir) urging the administration not to cut off the Chinese open-weight models smaller American firms rely on (Politico 2026-07-23; CNBC 2026-07-25). Revisit against adoption data on both sides.

I hereby announce that in the next five years, China will provide developing countries with 5,000 opportunities in AI training and seminar programs; develop international AI application cooperation centers with the Association of Southeast Asian Nations, the League of Arab States, the African Union, the Community of Latin American and Caribbean States, the Shanghai Cooperation Organization, and BRICS; and enable 30 countries to use the AI-powered meteorological warning system, or MAZU, to safeguard homes around the world.?

Decided by: Whether, by July 2031, Chinese ministry records and recipient-government announcements show 5,000 AI training and seminar places delivered to developing countries, AI application cooperation centres established with all six named regional blocs, and the MAZU meteorological warning system in operational use in 30 countries — resolves by 2031-07-17

Theme: adoption · Commitment · From state-of-ai/international-competition

Context: A five-year delivery commitment made on July 17, 2026; none of the three deliverables is checkable yet. The address supplies no budget, start date or delivery mechanism and names no body responsible for any of them, so the count has to be checked against Chinese ministry records and recipient governments rather than against the pledge itself. The earliest checkable milestones are the establishment of any of the six named cooperation centres and the first MAZU deployments outside China.

Our internal belief is that by March of 2028 we may have a significant fraction of our research being done by AI systems in tandem with our own researchers.?

Decided by: Whether, by 31 March 2028, OpenAI or a credible independent account describes a substantial and named share of OpenAI's research work as being performed by AI systems working alongside its researchers. OpenAI defines no threshold for 'a significant fraction', so the charitable reading applied at resolution is a share OpenAI itself characterises as significant, in a dated public statement, with the work identified — not a capability demonstration and not a further projection. — resolves by 2028-03-31

Theme: capability-milestones · Forecast · From state-of-ai/leading-models

Context: Not yet resolvable; the date is 2028. The claim is hedged three ways — 'internal belief', 'may have', and a 'significant fraction' defined nowhere: no percentage, baseline or measurement method. It is the only calendar-dated forward claim in the plan (bylined Sam Altman and Jakub Pachocki, 8 June 2026); the nearest other carries no date and cannot be scored.

AI's $200B question is now AI's $600B question.?

Decided by: Annualized AI end-user revenue measured against the ~$600B implied by Nvidia run-rate capex — does the revenue gap close or persist — resolves by 2027-06-30

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Multi-year framing whose horizon runs to mid-2027; still unverified. The gap Cahn described has widened, not closed: Goldman now puts the five largest spenders near $750 billion of 2026 capital spending against filed revenue an order of magnitude smaller — OpenAI's audited 2025 revenue was $13.07 billion, and Microsoft's firmest disclosure is Azure passing $100 billion for its fiscal year, a segment that is not AI-only.

Understating depreciation by extending useful life of assets artificially boosts earnings - one of the more common frauds of the modern era.?

Decided by: Whether hyperscaler earnings are shown (via restatements, SEC action, or realized impairments) to have been overstated by understated depreciation — Burry's specific ~$176B / Oracle ~27% / Meta ~21% by 2028 — resolves by 2028-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Forward, falsifiable accounting claim; resolves via later filings. Its first real test arrived when Microsoft extended its data-centre useful lives from 15 to 25 years (effective FY2027) — but on buildings, not the GPUs Burry said turn over every two to three years, and Microsoft calls the effect a 'minimal benefit to FY27 operating income.' FactSet separately reports used H100/A100 prices holding up, consistent with the hyperscalers' longer schedules. Burry's own figures remain untested.

Even as we invest to meet growing demand, full fiscal year operating margins should be down less than a point. In addition, we expect to remain free cash flow positive in FY27.?

Decided by: Whether Microsoft reports positive free cash flow for its fiscal year ending 30 June 2027, on its own definition, while capital expenditure grows year over year — resolves by 2027-08-31

Theme: bubble · Commitment · From state-of-ai/the-money

Context: Forward guidance from a company officer, resolvable against Microsoft's filed fiscal-2027 cash-flow statements. It starts from a wide margin: free cash flow was $19.64 billion in the quarter just reported, against Alphabet's negative $5.855 billion and Meta's $784 million. The test is not whether it stays positive but by how much, given guided capital expenditure of around $175 billion for the year.

Heavy capital spending relative to revenue will lead to declining, and in some cases negative, free cash flow and will hurt leverage ratios, to the extent these expenditures are debt-financed.?

Decided by: Hyperscaler free cash flow declining and in some cases turning negative, with leverage ratios deteriorating where the AI capex is debt-financed — resolves by 2027-12-31

Theme: bubble · Forecast · From state-of-ai/the-money/debt-and-credit

Context: Forward, conditional call, now partly observable. Free cash flow has turned down across the group — Alphabet negative $5.855 billion in Q2 2026, Meta down to $784 million, Amazon's trailing-twelve-month at $1.2 billion — and the debt limb has begun at Meta, whose long-term debt rose to $83.664 billion in six months as cash flow collapsed. But leverage has not yet deteriorated on Moody's measure and no rating action has followed. Resolves on fiscal-2026 and 2027 filings.

We expect that this will reduce our total corporate workforce as we get efficiency gains from using AI extensively across the company.?

Decided by: Direction of Amazon's total corporate headcount through ~2028 and whether Amazon credibly attributes reductions to AI efficiency. — resolves by 2028-12-31

Theme: jobs · Commitment · From state-of-ai/societal-impact

Context: Not due (horizon 2028). Amazon has cut corporate roles substantially — 14,000 in October 2025, about 16,000 in January 2026 — so the headcount half is tracking. The causal half is being disowned by the speaker: Jassy ruled out AI in October 2025 ("It's culture… you end up with a lot more layers"), and Beth Galetti framed the January round as removing bureaucracy. Reduction confirmed; AI attribution withdrawn.

The most likely (by far) outcome is for the status quo to continue and for the best open models to lag the best closed models by 6-9months.?

Decided by: Lag in months between the best open-weight and best closed model at equivalent score on the Artificial Analysis Intelligence Index (or successor independent index) — resolves by 2026-12-31

Theme: capability-milestones · Forecast · From state-of-ai/leading-models

Context: Still open — resolves at end-2026 — but current readings run against it. On the Artificial Analysis index the best open model (Kimi K3, 60) trails the closed leader (Claude Opus 5, 63) by weeks, not the forecast 6-9 months. Caveats: the metric is a year-end state; Lambert warned public benchmarks may overstate open closeness; and the UK AI Security Institute puts open cyber capability 4-7 months behind. Recheck at resolution.

Google will replenish more water than it consumes across its data-center sites by 2030, and U.S. data centers in aggregate use less than 1% of the nation’s water.±

Decided by: Google replenishes more freshwater than it consumes across its data-center sites — resolved (mixed)

Google's replenishment pledge is a forward commitment (unresolved). The aggregate "less than 1%" framing is broadly consistent with national estimates but sits in tension with primary local records — e.g. Google's own quintupled draw at The Dalles reaching ~40% of that city's supply — where impacts concentrate. True in aggregate, misleading locally. This cycle adds a second problem with the aggregate framing rather than the local one: the standard operational metric, Water Usage Effectiveness, counts neither the potable-versus-reclaimed distinction nor the water consumed generating the electricity, and Lawrence Berkeley National Laboratory puts that indirect footprint at nearly 800 billion litres for 2023 against 66 billion litres of direct on-site water — roughly twelve times larger. LBNL separately states its own water model may run low, and that correcting one cooling assumption would lift modelled hyperscale water intensity by about a quarter. Verdict held at mixed.

Theme: infrastructure-impact · Commitment · From state-of-ai/infrastructure

Context: Google's replenishment pledge is a forward commitment (unresolved). The aggregate "less than 1%" framing is broadly consistent with national estimates but sits in tension with primary local records — e.g. Google's own quintupled draw at The Dalles reaching ~40% of that city's supply — where impacts concentrate. True in aggregate, misleading locally. This cycle adds a second problem with the aggregate framing rather than the local one: the standard operational metric, Water Usage Effectiveness, counts neither the potable-versus-reclaimed distinction nor the water consumed generating the electricity, and Lawrence Berkeley National Laboratory puts that indirect footprint at nearly 800 billion litres for 2023 against 66 billion litres of direct on-site water — roughly twelve times larger. LBNL separately states its own water model may run low, and that correcting one cooling assumption would lift modelled hyperscale water intensity by about a quarter. Verdict held at mixed.

Electricity rates will not go up for residents because of this project. Georgia families will not subsidize this project.?

Decided by: Georgia residential electricity rates rise in a way Georgia Public Service Commission filings attribute to the Effingham County data center — resolves by 2032-12-31

Theme: infrastructure-impact · Forecast · From state-of-ai/infrastructure

Context: Not yet checkable: the power contract is unapproved and the first phase is not due until 2028. Georgia PSC large-load rules from January 2025 are meant to bar passing such costs to existing customers; whether they do is the test. Resolves on Georgia Power rate filings through 2032.

You are the first leader I have welcomed in person since taking office, and that is no accident. It is intended to send a very clear message: we are with Ukraine 100%, I am personally with you 100%, and I will honor every commitment this country has made to Ukraine in full.?

Decided by: Published UK government (MoD/HM Treasury) figures for military support to Ukraine covering the 2026-27 and 2027-28 financial years, cross-checked against the Kiel Institute Ukraine Support Tracker: the pledge holds if UK military support is sustained at £3bn a year or above and the 2024 UK-Ukraine Agreement on Security Co-operation remains in force. — resolves by 2027-07-31

Theme: geopolitics-war · Commitment · From ukraine-russia-war/nato-relations

Context: A pledge made on Burnham's first week in office and his first meeting with a foreign leader, recorded here as a forward-looking commitment rather than a checkable past fact. The concrete commitment it inherits is published by his own government in the same release: £3 billion a year in military support to Ukraine 'for as long as it takes', within total UK support of £25 billion since the full-scale invasion. Nothing yet contradicts it — the visit itself, and the transfer of the Stone Cloak intellectual property, are consistent with continuity. Note that Burnham's predecessor's commitments included the January 2024 UK-Ukraine Agreement on Security Co-operation and the 100 Year Partnership, but not a position on the timing of Ukraine's NATO accession, which he did not address.

It's kind of like having your parents co-sign on your first mortgage?

Decided by: Whether Nvidia's vendor-financed investments in customers (OpenAI, CoreWeave, Lambda, others) unwind badly, as 1990s telecom vendor financing did — resolves by 2028-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Forward characterization; not yet resolvable, since no unwind has occurred. The structures have instead expanded: Bloomberg counts more than $540 billion of Nvidia customer and partner deals announced in 2026, and Google has agreed to backstop Anthropic's leases in what amounts to a $35 billion arrangement. Whether any of it unwinds badly stays open, with the horizon running to 2028.

So between us, we're going to do half a trillion dollars' worth of business.?

Decided by: Whether roughly $500 billion of two-way business between Nvidia and SK Group is realised in Nvidia's filed revenue and component purchases and SK Group's disclosures — resolves by 2030-12-31

Theme: bubble · Commitment · From state-of-ai/the-money

Context: Forward claim about a multi-year partnership announced days before this cycle. Note the disclosure form: it is a gross two-way figure that nets Nvidia's own memory purchases against SK Group's system purchases, so it is not $500 billion of Nvidia revenue. Resolves against Nvidia's filed revenue and SK Group's disclosures.

AI is likely to replace half of all white-collar workers in the U.S.?

Decided by: Share of U.S. white-collar employment replaced by AI by 2030, vs Farley's ~50%. — resolves by 2030-12-31

Theme: jobs · Forecast · From state-of-ai/societal-impact

Context: Not due (charitable ~5-year reading, resolveBy 2030). Evidence runs against it from the speaker's own company: Bloomberg reported in June 2026 that Ford has been rehiring quality inspectors after AI fell short, and Farley told NewsNation in July there is "no replacement" for experienced engineers. That is a partial retreat by the same speaker, not a resolution — the claim is about U.S. white-collar employment by 2030.

I could easily see 30% of [~26,000 back-office roles] getting replaced by AI and automation over a five-year period.?

Decided by: Number of IBM back-office/HR roles left unbackfilled and attributable to AI/automation over 2023–2028, per IBM disclosure or credible reporting (Krishna's ~7,800 figure). — resolves by 2028-12-31

Theme: jobs · Forecast · From state-of-ai/societal-impact

Context: A dated claim about IBM's own workforce, resolvable by 2028. As of this cycle IBM has not published an audited count of back-office roles left unbackfilled specifically because of AI, so it stays unverified. Tagged as a prediction.

BT will cut up to 55,000 jobs by 2030, with about 10,000 roles replaced by AI and new technology.?

Decided by: BT total headcount reduction by 2030 and the ~10,000 share credibly attributable to AI/automation, per BT disclosure or reporting. — resolves by 2030-12-31

Theme: jobs · Commitment · From state-of-ai/societal-impact

Context: A 2030 workforce forecast; the ~10,000 AI-attributed share is the disputed part (Barford: framing over new substance). Tagged as a prediction, resolvable by 2030.

By 2030, AI and other trends will create 170 million jobs and displace 92 million — a net gain of 78 million.?

Decided by: Net global job change 2025→2030 vs WEF's +78M / 22% churn, per ILO and national labor statistics. — resolves by 2030-12-31

Theme: jobs · Forecast · From state-of-ai/societal-impact

Context: A 2030 macro forecast blending AI with demographic and green-transition trends, so it is not an AI-only claim. Tagged as a prediction; resolvable against net-employment statistics by 2030.

Global banks will cut as many as 200,000 jobs over the next three to five years as AI encroaches on human tasks.?

Decided by: Net headcount change across major global banks 2025→2030 attributable to AI, vs the projected ~200,000 (~3%) reduction. — resolves by 2030-01-09

Theme: jobs · Forecast · From state-of-ai/societal-impact

Context: Based on surveyed executive expectations, not committed plans. Tagged as a prediction, resolvable by 2030.

AI could wipe out half of all entry-level white-collar jobs and spike unemployment to 10–20% in the next one to five years.?

Decided by: Share of U.S. entry-level white-collar jobs eliminated and the U.S. unemployment rate by 2030, vs Amodei's ~50% and 10–20%. — resolves by 2030-05-28

Theme: jobs · Forecast · From state-of-ai/societal-impact

Context: Not due (1–5-year horizon from May 2025). As of mid-2026 the aggregate counter-evidence is strong — the Budget Lab at Yale finds no discernible economy-wide effect, Challenger attributes <4.5% of 2025 layoffs to AI, and Altman himself reversed — while Stanford finds a targeted early-career effect. Tagged as a prediction.

my basic prediction is that AI-enabled biology and medicine will allow us to compress the progress that human biologists would have achieved over the next 50-100 years into 5-10 years.?

Decided by: Documented biomedical R&D progress (major disease cures/therapeutics, longevity gains) attributable to AI equivalent to 50–100 years of prior-rate advancement, achieved within 5–10 years of ‘powerful AI.’ — resolves by 2035-12-31

Theme: capability-milestones · Forecast · From state-of-ai/safety

Context: From Amodei’s essay ‘Machines of Loving Grace.’ The clock is defined to start at the arrival of ‘powerful AI’ (which Amodei does not fix to a firm date), so the resolve-by below reads the ‘within 5-10 years’ window against the essay’s October 2024 framing; the metric is loosely operationalised and may resolve mixed.

If the ~7-month doubling of the task-completion horizon holds, AI agents will be able to complete many software/engineering tasks that currently take humans days or weeks — reaching roughly month-long autonomous tasks — within the decade.?

Decided by: The 50%-reliability task-completion time-horizon of the frontier model on METR’s public time-horizons tracker reaching ~1 month (≈160 human work-hours). — resolves by 2032-12-31

Theme: capability-milestones · Forecast · From state-of-ai/safety

Context: METR’s own extrapolation from its measured six-year doubling trend; not yet resolvable. METR notes a 10x error in the absolute measurement would shift arrival estimates by only ~2 years, and that the SWE-Bench-Verified subset doubled even faster (under 3 months).

of course we're in an AI bubble?

Decided by: Whether AI-linked equity valuations undergo a sharp correction, and on what horizon (Gelsinger: several years out; Marks: not yet) — resolves by 2029-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Forward call; not yet resolvable. Recorded alongside opposing views (Marks, Fink) so the conversation's calibration can be scored later.

That is capitalism. We're going to have some big winners and we're going to have some big losers ... but if you have a diversified portfolio, you're going to be fine.?

Decided by: Whether AI capex proves 'well spent' — hyperscalers emerge as clear winners with no broad bust — resolves by 2028-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Boom-side forward call; not yet resolvable.

OpenAI is on track to generate more than $20 billion in annualized revenue run rate this year, with plans to grow to hundreds of billions in sales by 2030.±

Decided by: OpenAI reported (GAAP) annual revenue reaching 'hundreds of billions' by 2030 — resolved (mixed)

Still mixed. The exit-month run rate did cross $20 billion, but audited 2025 GAAP revenue was $13.07 billion, so the framing overstates realised revenue. The 2030 limb remains open and is now trending: the run rate passed $40 billion by August 2026, roughly double the end-2025 pace.

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Still mixed. The exit-month run rate did cross $20 billion, but audited 2025 GAAP revenue was $13.07 billion, so the framing overstates realised revenue. The 2030 limb remains open and is now trending: the run rate passed $40 billion by August 2026, roughly double the end-2025 pace.

There will be no federal bailout for AI. The U.S. has at least 5 major frontier model companies. If one fails, others will take its place.?

Decided by: Whether any federal bailout, guarantee or backstop is extended to an AI or frontier-model company — resolves by 2028-12-31

Theme: bubble · Forecast · From state-of-ai/the-money

Context: Forward commitment; no such backstop had occurred as of this cycle. Not yet resolvable.

There is no way hyperscaler AI infrastructure investments will generate returns at current spending levels.?

Decided by: Hyperscaler AI-infrastructure capex at 2025 spending levels fails to generate adequate financial returns — resolves by 2027-12-31

Theme: infrastructure-impact · Forecast · From state-of-ai/infrastructure

Context: Forward-looking claim by IBM CEO Arvind Krishna about industry-wide capex returns; unresolved. Tagged as a prediction for The Record.

Global data-center capital spending will exceed $1 trillion in 2026.?

Decided by: Global data-center capital expenditure reaches or exceeds $1 trillion in calendar 2026 — resolves by 2026-12-31

Theme: infrastructure-impact · Forecast · From state-of-ai/infrastructure

Context: Forward projection by Dell’Oro Group; unresolved. Tagged as a prediction for The Record.

The U.S. is doubling down on restricting sales of chips to China and purchases from China, but models like Qwen 3 that are state-of-the-art and open […] will undoubtedly be used domestically. It reflects the reality that businesses are both building their own tools [as well as] buying off the shelf via closed-model companies like Anthropic and OpenAI.

Decided by: Documented US enterprise adoption of Chinese open-weight models (hosting-platform traffic shares, credible adoption reporting) — resolved (true)

Confirmed by mid-2026: CNBC reported Western companies increasingly adopting Chinese open-weight models on cost grounds, and OpenRouter token-traffic data showed GLM-5.2 usage climbing faster post-launch than any predecessor — Chinese open models are in production US use at scale.

Theme: adoption · Forecast · From state-of-ai/leading-models

Context: Confirmed by mid-2026: CNBC reported Western companies increasingly adopting Chinese open-weight models on cost grounds, and OpenRouter token-traffic data showed GLM-5.2 usage climbing faster post-launch than any predecessor — Chinese open models are in production US use at scale.