FAQ
What is Slicast?
Slicast is a real-time intelligence service covering the global AI infrastructure buildout — GPU clouds, data centers, AI chips, power, and the capital behind them. It surfaces primary sources (SEC filings, company announcements, press wires) first, with a daily brief and breaking alerts.
Which companies are spending the most on AI infrastructure capex?
By annual capital expenditure, the hyperscalers (Microsoft, Amazon, Google, Meta) lead in absolute dollars, while neoclouds such as CoreWeave and Core Scientific show the highest capex intensity (capex as a share of revenue, often above 100%). Slicast's Capex Tracker ranks them directly from SEC XBRL filings.
What is a neocloud?
A neocloud is a specialized GPU cloud provider — such as CoreWeave, Nebius, Crusoe or Lambda — that rents out AI compute (GPU clusters) at scale, distinct from the general-purpose hyperscalers. They are central to the AI buildout because they expand GPU capacity fastest.
Where can I track AI data center capex and SEC filings?
Slicast's Capex Tracker pulls annual capital expenditure for the public companies building AI infrastructure straight from SEC XBRL filings, and the Primary Disclosures stream surfaces their 8-K filings, newsroom posts and press wires in real time.
How often is Slicast updated?
The daily brief publishes every morning (Sydney time), and breaking alerts are pushed in real time around the clock as high-importance developments are detected.
Does Slicast cover China and other regions?
Yes. Slicast covers the AI buildout across five regions — the US, China and APAC, Australia, Europe, and the Middle East — so global compute, chip, power and capital developments are in one place.
How much are hyperscalers spending on AI data centers?
Microsoft, Amazon, Google and Meta together commit hundreds of billions of dollars a year in capital expenditure, much of it for AI data centers and compute. Slicast's Capex Tracker shows each company's latest annual figure straight from SEC filings.
What is HBM and why does it matter for AI?
HBM (high-bandwidth memory) is the stacked memory that feeds AI accelerators like Nvidia GPUs. It is a key bottleneck in the AI buildout — SK Hynix, Samsung and Micron dominate supply, and HBM demand drives much of the memory-chip cycle.
What is the difference between a neocloud and a hyperscaler?
Hyperscalers (AWS, Azure, Google Cloud) are general-purpose cloud giants; neoclouds (CoreWeave, Nebius, Crusoe, Lambda) are specialized GPU clouds focused purely on AI compute, typically expanding capacity faster and at higher capex intensity.
See also Capex Tracker · Primary Disclosures · Methodology.