Keep pulling the thread on Brandon McBee.
Uber reportedly exhausted its entire 2026 AI budget within the first four months of the year.
According to an Axios report, an AI consultant's client spent $500 million in a single month on the AI model Claude after failing to implement usage limits.
According to Ian Dunning of Hudson River Trading, the primary scarcity in AI infrastructure is now real estate and finding suitable locations to install GPUs, which is as or more challenging than acquiring the GPUs themselves.
CoreWeave has more than 10 clients with contracts valued at over $1 billion each.
CoreWeave's backlog of contracts from financial services clients is valued in the tens of billions of dollars.
CoreWeave provides infrastructure for 9 of the top 10 AI labs in the world, excluding those in China.
In Q4 of the most recent year, CoreWeave added twice as many new client logos as in any previous quarter, driven primarily by growth in its enterprise customer base.
Brandon McBee asserts that NVIDIA H100 GPUs could have a useful life of 6 to 8 years, and A100s even longer, because different AI workloads will continue to require older, more cost-effective hardware.
CoreWeave's direct backlog from financial services clients is approaching $10 billion.
AI labs are increasing the duration of their committed infrastructure contracts with CoreWeave, moving from 3-year terms a couple of years ago to 5-year terms currently.
CoreWeave's clients are exclusively requesting NVIDIA infrastructure, with no material demand for custom silicon from other manufacturers.
Inference workloads represent "well in excess of 50%" of the total infrastructure utilization on CoreWeave's platform.