Cloud backlogs swell as hyperscalers post soaring commitments
Hyperscale commitments expanded past half a trillion dollars as model developers posted rapid usage gains alongside widening operational pressure.
Big cloud providers reported mounting multi-year enterprise commitments alongside rapid user adoption across foundation model developers [1].
Google Cloud stated that its backlog reached $561 billion, and Oracle registered $664 billion in remaining performance obligations [1]. Amazon Web Services brought in $46 billion in revenue during the third quarter of 2026, with Amazon representatives stating that "almost as fast as we bring capacity online, customers are consuming it" [1]. Meanwhile, Microsoft forecast Azure revenue growth between 49% and 50% for the quarter ending in December [1].
Demand for computing capacity matches expanding scale at key foundation model companies [1]. ChatGPT reached 1.2 billion weekly active users in September 2026. OpenAI recently saw its valuation revised from $1.4 trillion to $1.25 trillion [1]. Anthropic disclosed that more than 6,000 customers now commit at least $100,000 annually, while its engineering team cut central processor utilization by 40% to 50% after rewriting its Messages API from Python into Rust [1].
Infrastructure rollouts continue to face practical hurdles. A federal court denied xAI's motion to dismiss an environmental lawsuit over natural gas turbines deployed at its Southaven facility [1]. Model sizes also continue to expand, with Reflection AI's Beam 1.0 reaching 500 billion parameters and Mistral AI releasing its Le Chonk model with 1 trillion parameters [1].
Questions and answers
What were the reported backlog figures for Google Cloud and Oracle?
Google Cloud's backlog reached $561 billion, while Oracle reported $664 billion in remaining performance obligations [1].
How many users did ChatGPT reach in September 2026?
ChatGPT reached 1.2 billion weekly active users in September 2026 [1].