Early Impacts of Generative AI

GENERATIVE AI: IT'S FUNDAMENTALLY DIFFERENT

DigitalBridge explores Generative AI and the new layers of creativity that go beyond traditional AI. To unlock those capabilities every enterprise software platform is being re-architected to incorporate generative AI and that new layer of creativity translates into a lot more compute.

GENERATIVE AI ADOPTION IS THE FASTEST ON RECORD AND WE BELIEVE A CLOUD-SCALE OPPORTUNITY 

With the speed of adoption of Generative AI being the fastest on record.  We believe the Generative AI opportunity is going to be at least as big as the cloud market today over the next 5-10 years. We describe Generative AI as a cloud-scale opportunity and one that's frankly just in the first inning.

CLOUD-TRAINED GENERATIVE AI IS POWER HUNGRY

Generative AI model training is compute-bound; more compute resources are tied to a better product. Access to digital infra in size, at the lowest total cost, is a key success factor. Chip power is exploding.  The new specialized AI chips and GPUs consume 2-3x the power of prior generations. There is going to be higher power density on a per rack basis with each rack filled with power-hungry GPUs drawing 40 or more Kw compared to traditional data center racks drawing 10kw or less.

AI training is not as location or latency sensitive as applications. Its not about uptime and latency, it is about the ability to access and invest in large concentrated MW capacity.  Access to digital infrastructure in size, at the lowest cost, is a key success factor.