Krishna Rao, the newly appointed Chief Financial Officer of Anthropic, is known for orchestrating a $1 billion lifeline for Airbnb during the pandemic. Now he is applying that strategic rigor to the world of AI compute.
"The compute that we procure is the lifeblood of our business," Rao said. He emphasized the fine line between excess and scarcity, warning, "If you buy too much compute you go out of business; if you buy too little compute you can't serve your customers."
Flexibility is also key. Rao revealed Anthropic uses three different chip platforms-Amazon's Trainium, Google's CPUs, and Nvidia's GPUs-a capability that took significant time to develop, but it allows a dollar of compute to go further than in most organizations.
Rao is also breaking away from traditional financial mindsets, describing the need to shift from linear to exponential thinking to navigate the rapidly evolving landscape. Dynamic compute resource allocation, he says, improves model development efficiency and cost-effectiveness.
On the performance front, Rao stated, "For us, the scaling laws are alive and well" and progress is accelerating. He noted Anthropic started the year with roughly $9 billion in annualized run-rate revenue, underscoring that major leaps in model intelligence directly correlate with significant revenue growth.