Glossary term
Hyperscaler
A cloud provider that builds and operates data-centre capacity at global scale — Microsoft, Amazon, Alphabet and Meta, with Oracle usually counted as the fifth. They are the largest buyers of AI compute and the top rung of the AI funding ladder.
AI-generated — produced automatically by Closelook’s systems under this site’s editorial policy.
What it means
The word describes scale of operation: fleets of data centres across regions, custom hardware, and capital budgets in the tens of billions of dollars a year. The four names most analysts mean are Microsoft (Azure), Amazon (AWS), Alphabet (Google Cloud) and Meta, which builds for its own products rather than for tenants. Oracle joined the list once its cloud infrastructure business began ordering at the same scale.
Hyperscalers sit on both sides of the AI trade: they buy chips, memory, optical components and power from the physical layer, and they sell compute, models and applications to everyone else. When compute gets cheaper, their margins on what they sell widen.
Why it matters for the AI trade
Their capital expenditure is the demand that the Rubin Build-Out 100 supplies, and their balance sheets decide how it is financed. On our credit stress tape the four fund capex from operating cash flow at 1.34× cover and cover interest about 50 times — the top rung of the funding ladder. Oracle, funding the same build with debt, sits one rung down. On 14 September 2026, the first session after the frontier labs asked for a slower frontier, Microsoft, Alphabet and Meta rose 2% to 3% while the chip index fell 5.6%: the buyers of compute were bought, the sellers sold.
How Closelook uses it
Hyperscaler capex guidance is the demand-side anchor of the capex cycle; the alternative hyperscalers read covers the names outside the big four. Amazon carries the tape’s first drift flag: capex not covered by operating cash flow over the trailing four quarters.
Common questions
- Is Nvidia a hyperscaler?
- No. Nvidia sells the chips hyperscalers buy; it is the top of the physical layer. It appears on the funding ladder as Tier 6 — vendor financing — when it finances its own customers.
- Is Oracle a hyperscaler?
- Increasingly yes by scale of orders, but it funds its build-out differently: with debt rather than operating cash flow, which is why Closelook keeps it on its own rung of the ladder.
- Why did hyperscalers rise when chips fell in September 2026?
- Because a slower frontier means cheaper inputs for the companies that buy compute and sell services on top of it. The market sorted the trade by who pays for compute and who is paid for it.