A permanent map of the physical economy behind artificial intelligence, from lithography and accelerators to grid connections, cooling, debt and inference capacity.
AI semiconductor demand is spreading from data-centre accelerators into premium edge devices, making advanced-node manufacturing and packaging increasingly important to a wider set of chip companies.
The AI investment cycle is transmitting through semiconductors, data centres, electricity systems and export supply chains, creating very different opportunities and constraints across Asian economies.
Calls for a slower pace of frontier AI development have exposed how much of the equity rally now rests on an extraordinary infrastructure buildout. Investors need to watch orders, not rhetoric.
The Federal Reserve begins its September meeting with oil prices elevated and Treasury yields around levels that tighten financial conditions before policymakers make any move.
A planned benchmark covering roughly $330 billion of local-currency government debt could make frontier markets easier for global investors to compare and allocate to.
If frontier-model training slows while deployed AI keeps growing, spending can shift from giant training clusters toward inference chips, networking, power and software that serves live users.
Big Tech infrastructure spending has moved into a new scale. This tracker separates disclosed company plans from the much larger ecosystem estimates often quoted around the AI boom.
Hyperscalers are issuing debt at unusual scale to finance data-centre expansion, pushing AI investment beyond equities and into credit spreads, duration and refinancing risk.
AI compute demand is colliding with generation, transmission and grid-connection limits, making electricity availability a gating factor for data-centre growth rather than a background utility input.
Independent GPU-cloud providers can grow faster than traditional cloud platforms, but heavy hardware spending, supplier concentration and rapid depreciation make utilisation the decisive metric.
The Federal Reserve meets on 15-16 September with markets heavily pricing a rate increase as stronger inflation and oil above $100 complicate the policy outlook.
AI-linked shares fell after leading lab executives backed slowing frontier development, while oil-driven rate expectations simultaneously raised the discount rate applied to growth stocks.
Many AI equities derive a large share of their valuation from profits expected years into the future, making changes in discount rates unusually important even when current revenue is still growing.
AI capital spending moves through a chain from cloud-company budgets to accelerator orders, memory, networking, data-centre construction and power equipment. Each layer turns at a different time.
Permanent reference coverage, company and location guides, and comparison-led research. News remains separately curated above.
Dividend and reinvestment calculators designed to sit alongside GMR's permanent investing guides.
Company pages with sourced market fields and freshness information.
Country profiles connected to companies, sectors and market coverage.
Dated company and sector comparisons with methodology and sources.