Open research
Follow the physical buildout behind the AI economy
Company filings, official monthly revenue, price indices, and trade statistics—kept in their native definitions and published with the caveats needed to interpret them.
Open research datasets
Company disclosures
AI infrastructure capex
Quarterly spending, annual outlooks, data-center commitments, and disclosure coverage across the largest AI infrastructure buyers.
Evidence through July 19, 2026
Open research →Monthly operating data
Taiwan AI supply chain
A breadth-first view of official monthly revenue across reviewed foundry, packaging, server, power, cooling, and component relationships.
Latest month: June 2026
Open research →Prices and trade
DRAM price index and NAND flash trends
Bank of Korea export-price indices paired with KITA DRAM and NAND export values, with price-versus-mix decomposition.
Latest month: June 2026
Open research →Company capex
Issuer-level histories
Reviewed quarterly spend and annual outlooks in each company’s native definition.
Why these datasets belong together
Demand, suppliers, and pricing are different layers
Hyperscaler disclosures describe spending intentions and recognized capital expenditure. Taiwan company revenue shows activity across reviewed parts of the supply chain. Korean price and trade series add memory pricing and product-mix context.
No single layer proves end demand. Reading them together makes divergences visible without forcing incompatible units into one synthetic “AI capex” number.
Reviewed data, explicit limits
Pending evidence is excluded from published values. Missing company disclosures remain missing, source units stay visible, and historical backfills are not treated as point-in-time publication snapshots.