MLQ.ai
About Sign in Subscribe

Data accuracy

Measured against the filings themselves — not another vendor's feed · methodology

Our extraction pipeline reads SEC XBRL. To verify it, an independent reader — an LLM with zero shared code — reads the rendered filing and we diff the two. One bug cannot corrupt both paths, so agreement is evidence, not circularity. Sampling is seeded and stratified across company sizes: re-runnable by anyone.

VERIFIED VS THE RENDERED SEC FILING — 250 STRATIFIED COMPANIES, SEEDED SAMPLE, JULY 2026 (3rd audit cycle)
99.4% adjudicated · 99.0% raw agreement across 1,330 data points (±1% tolerance)
EPS (diluted)99.5%
Total assets99.6%
Total liabilities99.5%
Stockholders' equity99.6%
Net income97.5%
Revenue98.1%

The adjudicated figure counts the disagreements we have examined against the filing by hand: nearly all sit in two documented classes — bank revenue composition (we report the standard net-interest basis) and net income before vs after preferred dividends (we report the filed attributable-to-parent line). Both are definitional, both are published, none are smoothed over.

13F HOLDINGS — RECONCILED TO EACH FILER'S OWN COVER PAGE
99.4% of sampled filings reconcile to the filer's self-reported portfolio total (300+ filing sample, July 2026). CUSIP→company mapping covers ~99% of operating-company dollars. Spot adjudication against EDGAR: our BlackRock NVDA position matches the filing to the share (1,925,533,174) — a figure a leading competitor's published material understates by ~118M shares.
CROSS-SOURCE — FULL-UNIVERSE DIFF VS A COMMERCIAL PROVIDER
206,935 field comparisons across 4,906 companies: 92% agreement on core statements (revenue, net income, assets, liabilities, equity, operating cash flow, ±1%). Every divergent field is annotated: most are deliberate — we report as-filed numbers where providers recompute (operating income, EBITDA, debt incl. operating leases). The differences are documented in the data dictionary, not smoothed over.

Every value the API returns carries its provenance — the exact accession number and filing URL — so any number can be checked against the source in one click. Found a discrepancy? [email protected] — adjudications feed the next audit cycle.

Fuel-Cell Deployments at U.S. Data CentersReport
Research

Fuel-Cell Deployments at U.S. Data Centers

Public disclosures confirm at least 154 MW of operating fuel cells at U.S. data centers. Seven proposed projects account for another 6.31 GW, led by large developments in New Mexico, Texas, and Wyoming.

No spam. Unsubscribe anytime.