Query

Saved analyses are materialised by the nightly export from the semantic layer; each shows its latest value, the observations behind it and the formula. The console runs read-only SQL against the same store; the Ask button in the header puts a model in front of it, with every number checked against the record it cites.

Saved analyses

METR horizon doubling times, both windows and both reliability levels
How fast is the time horizon doubling, and does the answer depend on the window?
108 daysas of 2026-03-05(25 obs)
obs:0d21fe65 obs:19631991 obs:2325ad87 obs:261c1f51 obs:2eb7c897 obs:3466aea4+19 more
OLS on ln(horizon); rows beyond the 16h suite ceiling excluded; CI from the slope standard error.
Formula
```
python: fits.loglinear
```

Constant-hazard misfit

Do the 50% and 80% horizons sit where a constant task-length hazard predicts?

2.04×as of 2026-03-05(2 obs)

obs:11a8bbd2 obs:858c619b

Two points per model on METR's suite; the constant-hazard model is a null, not a claim.

Formula

SELECT a.as_of_date, (a.value_numeric / b.value_numeric) / 3.1063 AS value, [a.id, b.id] AS obs_ids
FROM observations a JOIN observations b ON a.subject = b.subject AND a.as_of_date = b.as_of_date
WHERE a.series_key LIKE 'metr.%.horizon_50.pt' AND b.series_key LIKE 'metr.%.horizon_80.pt' AND NOT coalesce(a.disputed, false)

Margin stack by layer

Where in the stack does filed operating income bulge?

57.3%as of 2026-06-30(8 obs)compute_semis

obs:288c79d6 obs:815ceb0b obs:87068884 obs:89c26630 obs:a3832632 obs:d271f3bf+2 more

Fiscal quarters mapped to the calendar quarter of their period end (NVIDIA's late-July quarter counts as Q3); a filer's unreported fiscal Q4 is filled as fiscal year minus its three reported quarters and cites all four filings; quarters where any of the five segments is still missing are dropped; segment operating income, not gross profit; only two layers so far.

Formula

WITH q AS (
  SELECT subject, series_key, as_of_date, period_start, value_numeric AS v, id
  FROM observations WHERE series_key LIKE 'sec_seg.%.operating_income.q'),
fy AS (
  SELECT subject, replace(series_key, '.fy', '.q') AS series_key, as_of_date, period_start, value_numeric AS v, id
  FROM observations WHERE series_key LIKE 'sec_seg.%.operating_income.fy'),
fill AS (  -- FY minus the three quarters inside it, when exactly three are reported
  SELECT fy.subject, fy.series_key, fy.as_of_date, fy.v - sum(q.v) AS v, list_prepend(fy.id, list(q.id ORDER BY q.id)) AS ids
  FROM fy JOIN q ON q.subject = fy.subject AND q.series_key = fy.series_key
                 AND q.as_of_date > fy.period_start AND q.as_of_date < fy.as_of_date
  GROUP BY fy.subject, fy.series_key, fy.as_of_date, fy.v, fy.id
  HAVING count(*) = 3 AND NOT EXISTS (SELECT 1 FROM q q2 WHERE q2.series_key = fy.series_key AND q2.as_of_date = fy.as_of_date)),
allq AS (
  SELECT subject, as_of_date, v, [id] AS ids FROM q
  UNION ALL SELECT subject, as_of_date, v, ids FROM fill),
seg AS (
  SELECT date_trunc('quarter', as_of_date) AS cq, subject, v, ids,
         CASE WHEN subject IN ('nvda', 'amd') THEN 'compute_semis' ELSE 'compute_cloud' END AS layer_id
  FROM allq),
complete AS (SELECT cq FROM seg GROUP BY cq HAVING count(DISTINCT subject) = 5),
by_layer AS (SELECT s.cq, s.layer_id, sum(v) AS oi, flatten(list(ids)) AS ids FROM seg s JOIN complete c USING (cq) GROUP BY 1, 2)
SELECT (cq + INTERVAL 3 MONTH - INTERVAL 1 DAY)::DATE AS as_of_date, layer_id, oi / sum(oi) OVER (PARTITION BY cq) AS value,
       list_distinct(flatten(list(ids) OVER (PARTITION BY cq))) AS obs_ids
FROM by_layer

Concentration by layer

Is any layer fragmenting while its neighbours stay concentrated?

0.35 indexas of 2026-03-31(3 obs)compute_cloud

obs:01a2687f obs:5b994889 obs:a26cee82

Only covered filers; Broadcom, TSMC, Oracle and the neoclouds are absent, so levels overstate concentration. Fiscal quarters mapped to the calendar quarter of their period end. Ramp lab shares allow multi-homing and are normalised to sum to one.

Formula

WITH rev AS (
  SELECT date_trunc('quarter', as_of_date) AS cq, subject, value_numeric AS v, id,
         CASE WHEN subject IN ('nvda', 'amd') THEN 'compute_semis' ELSE 'compute_cloud' END AS layer_id,
         CASE WHEN subject IN ('nvda', 'amd') THEN 2 ELSE 3 END AS need
  FROM observations
  WHERE series_key IN ('sec_seg.nvda.data_center.revenue.q', 'sec_seg.amd.data_center.revenue.q',
                       'sec_seg.amzn.aws.revenue.q', 'sec_seg.googl.cloud.revenue.q', 'sec_seg.msft.intelligent_cloud.revenue.q')),
complete AS (SELECT cq, layer_id FROM rev GROUP BY cq, layer_id, need HAVING count(DISTINCT subject) = need),
sh AS (SELECT (r.cq + INTERVAL 3 MONTH - INTERVAL 1 DAY)::DATE AS as_of_date, r.layer_id,
              r.v / sum(r.v) OVER (PARTITION BY r.cq, r.layer_id) AS s, r.id
       FROM rev r JOIN complete c USING (cq, layer_id)),
lab AS (SELECT as_of_date, 'model' AS layer_id, value_numeric / sum(value_numeric) OVER (PARTITION BY as_of_date) AS s, id
        FROM observations WHERE series_key LIKE 'ramp.%.business_paid_share.m')
SELECT as_of_date, layer_id, sum(s * s) AS value, list(id ORDER BY id) AS obs_ids
FROM (SELECT * FROM sh UNION ALL SELECT * FROM lab) GROUP BY 1, 2

Lab recoupment ratios

How much run-rate has each lab earned per dollar of capital raised?

2.87×as of 2026-08-31(2 obs)

obs:89540992 obs:b7dae308

Tier-5 inputs; excludes debt, compute credits and cloud commitments; run-rates are press-reported annualised months.

Formula

WITH rr AS (SELECT subject, as_of_date, value_numeric AS rev, id FROM observations WHERE series_key LIKE 'epoch.%.revenue_run_rate_usd.pt'),
     eq AS (SELECT subject, as_of_date, value_numeric AS eq, id FROM observations WHERE series_key LIKE 'epoch.%.round_equity_usd.pt')
SELECT rr.as_of_date, rr.subject AS entity, rr.rev / sum(eq.eq) AS value,
       list_concat([rr.id], list(eq.id ORDER BY eq.id)) AS obs_ids
FROM rr JOIN eq ON eq.subject = rr.subject AND eq.as_of_date <= rr.as_of_date
GROUP BY rr.as_of_date, rr.subject, rr.rev, rr.id

Labour-tracker concordance

Do the four monthly labour trackers agree?

0as of 2026-07-31(4 obs)

obs:2ec713fc obs:83c12bef obs:a9276b9d obs:cbb77dc6

Different geographies (CA claims vs US employment), different exposure measures; a count, not a test.

Formula

WITH latest AS (
  SELECT series_key, value_numeric, as_of_date, id,
         row_number() OVER (PARTITION BY series_key ORDER BY as_of_date DESC) AS rn
  FROM observations
  WHERE series_key IN ('adp_research.us_high_exposure.employment_yoy.m', 'revelio.us_exposed.growth_gap.m',
                       'cait.ca_high_exposure.claims_3mma_mom.m', 'stanford_del.us_entry_level_exposed.employment_gap.pt'))
SELECT max(as_of_date) AS as_of_date,
       sum(CASE WHEN series_key = 'adp_research.us_high_exposure.employment_yoy.m' AND value_numeric <= -0.03 THEN 1
                WHEN series_key = 'revelio.us_exposed.growth_gap.m' AND value_numeric <= -0.10 THEN 1
                WHEN series_key = 'cait.ca_high_exposure.claims_3mma_mom.m' AND value_numeric >= 0.05 THEN 1
                WHEN series_key = 'stanford_del.us_entry_level_exposed.employment_gap.pt' AND value_numeric >= 0.25 THEN 1
                ELSE 0 END)::DOUBLE AS value,
       list(id ORDER BY id) AS obs_ids
FROM latest WHERE rn = 1

Capex to AI revenue

How far ahead of AI revenue is hyperscaler capex running?

2.51×as of 2026-06-30(16 obs)

obs:027237d8 obs:143dcf3e obs:196a6174 obs:29221e09 obs:416fff66 obs:4910df35+10 more

Hyperscaler capex is not all AI; the revenue side is run-rates (annualised, press-reported) plus one annual survey and double-counts where enterprises buy from labs; grade C by construction.

Formula

WITH capex AS (
  SELECT date_trunc('quarter', as_of_date) AS cq, subject, value_numeric AS v, id FROM observations
  WHERE series_key LIKE 'sec.%.capex.q' AND subject IN ('msft', 'googl', 'amzn', 'meta', 'orcl')),
qs AS (SELECT DISTINCT cq FROM capex WHERE cq >= DATE '2024-01-01'),
capex4 AS (
  SELECT q.cq, sum(c.v) AS v, list(c.id ORDER BY c.id) AS ids FROM qs q JOIN capex c ON c.cq > q.cq - INTERVAL 12 MONTH AND c.cq <= q.cq
  GROUP BY q.cq HAVING count(DISTINCT c.subject) >= 4),
rr AS (SELECT entity_id, as_of_date, value_numeric AS v, id FROM observations WHERE series_key LIKE 'epoch.%.revenue_run_rate_usd.pt' AND entity_id IS NOT NULL),
rr_latest AS (
  SELECT q.cq, r.entity_id, arg_max(r.v, r.as_of_date) AS v, arg_max(r.id, r.as_of_date) AS id FROM qs q JOIN rr r ON r.as_of_date <= q.cq + INTERVAL 3 MONTH AND r.as_of_date > q.cq - INTERVAL 9 MONTH GROUP BY 1, 2),
ent AS (SELECT as_of_date, value_numeric AS v, id FROM observations WHERE series_key = 'menlo.us_enterprise.genai_spend_usd.fy'),
ent_latest AS (SELECT q.cq, arg_max(e.v, e.as_of_date) AS v, arg_max(e.id, e.as_of_date) AS id FROM qs q JOIN ent e ON e.as_of_date <= q.cq + INTERVAL 3 MONTH GROUP BY 1)
SELECT (c.cq + INTERVAL 3 MONTH - INTERVAL 1 DAY)::DATE AS as_of_date,
       c.v / ((SELECT sum(v) FROM rr_latest r WHERE r.cq = c.cq) + e.v) AS value,
       list_distinct(flatten([c.ids, (SELECT list(id) FROM rr_latest r WHERE r.cq = c.cq), [e.id]])) AS obs_ids
FROM capex4 c JOIN ent_latest e USING (cq)

Circular financing total

How much circular financing has been signed?

$903Bas of 2026-08-17(14 obs)

obs:0020effe obs:15768aa0 obs:15f854e5 obs:47769aec obs:5b741dc1 obs:86bf3cb0+8 more

Mixes instruments (equity, guarantees, commitments) with different risk; "up to" and floor values taken at face; cumulative, so it only rises until a row is superseded.

Formula

WITH deals AS (
  SELECT as_of_date, value_numeric AS v, id FROM observations
  WHERE series_key LIKE 'circular.%_usd.pt'
    AND series_key NOT LIKE '%loi_usd%' AND series_key NOT LIKE '%talks_usd%' AND series_key NOT LIKE '%total_commitments%'
    AND series_key NOT LIKE '%commercial_rpo%')
SELECT last_deal AS as_of_date, sum(v) OVER (ORDER BY q_end) AS value,
       flatten(list(ids) OVER (ORDER BY q_end)) AS obs_ids
FROM (SELECT date_trunc('quarter', as_of_date)::DATE AS q_end, max(as_of_date) AS last_deal, sum(v) AS v, list(id ORDER BY id) AS ids
      FROM deals GROUP BY 1)

Price per capability

What does a METR-hour of capability cost on OpenRouter?

$0.30as of 2026-09-10(32 obs)

obs:03f5354f obs:0862bc67 obs:0d21fe65 obs:261c1f51 obs:2eb7c897 obs:37e97a6f+26 more

List prices on one router, prompt tokens only; the horizon is METR's suite, not general capability; only models present in both datasets count; a price series starts on 10 Sep 2026 and records change points, so the trend needs months.

Formula

WITH h AS (
  SELECT subject, arg_max(value_numeric, as_of_date) AS minutes, arg_max(id, as_of_date) AS hid
  FROM observations WHERE series_key LIKE 'metr.%.horizon_50.pt' AND NOT coalesce(disputed, false) GROUP BY subject),
p AS (SELECT subject, as_of_date, value_numeric AS usd, id FROM observations WHERE series_key LIKE 'openrouter.%.price_prompt_usd_per_mtok.pt')
SELECT p.as_of_date, min(p.usd / (h.minutes / 60)) AS value, list_distinct(flatten(list([p.id, h.hid]))) AS obs_ids
FROM p JOIN h USING (subject) GROUP BY p.as_of_date

Frontier share of venture dollars

What share of tracked venture dollars goes to the frontier labs?

94.8%as of 2026-06-30(32 obs)

obs:021fc48d obs:03a62453 obs:08c5c905 obs:0f018dc2 obs:14dda4d0 obs:15d383db+26 more

Denominator is the seed list, not the market; frontier rounds are mostly Epoch (tier 5) because the labs raise through vehicles that do not file Form D.

Formula

WITH f AS (SELECT entity_id, as_of_date, value_numeric AS v, id FROM observations WHERE series_key LIKE 'formd.%.amount_sold_usd.pt' AND entity_id IS NOT NULL),
e AS (SELECT entity_id, as_of_date, value_numeric AS v, id FROM observations WHERE series_key LIKE 'epoch.%.round_equity_usd.pt' AND entity_id IS NOT NULL),
rounds AS (SELECT * FROM f UNION ALL SELECT * FROM e WHERE NOT EXISTS (SELECT 1 FROM f WHERE f.entity_id = e.entity_id AND date_trunc('quarter', f.as_of_date) = date_trunc('quarter', e.as_of_date))),
m AS (SELECT entity_id, sublayer_id FROM entity_membership WHERE is_primary AND sublayer_id IS NOT NULL),
q AS (SELECT date_trunc('quarter', r.as_of_date) AS cq, m.sublayer_id, r.v, r.id FROM rounds r JOIN m USING (entity_id) WHERE r.as_of_date >= DATE '2023-01-01'),
qs AS (SELECT DISTINCT cq FROM q)
SELECT (qs.cq + INTERVAL 3 MONTH - INTERVAL 1 DAY)::DATE AS as_of_date,
       sum(CASE WHEN q.sublayer_id = 'frontier_labs' THEN q.v ELSE 0 END) / sum(q.v) AS value,
       list(q.id ORDER BY q.id) AS obs_ids
FROM qs JOIN q ON q.cq > qs.cq - INTERVAL 12 MONTH AND q.cq <= qs.cq GROUP BY 1 HAVING sum(q.v) > 0

Ask the data

Query

Saved analyses

Console