US marketing AI startup Alembic Technologies is procuring a GPU-based supercomputer.

The company this week announced it is deploying an Nvidia DGX H100-based AI supercomputer to power its new composite and causality AI systems.

Nvidia DGX H100
Nvidia's DGX H100 GPU – Nvidia

The system will help Alembic “infuse its applications with differentiated AI capabilities," the company said.

“The new Alembic release is the first to use causality mathematics and composite AI to mathematically demonstrate causality in large datasets, initially focusing on marketing ROI. Imagine being able to trace and understand the direct impact of your brand's spending, similar to reversing the butterfly effect of big brand moments,” said Tomás Puig, Alembic co-founder and CEO.

Based in San Francisco, California, Alembic says it can trace and prove the results of marketing programs and “apply composite AI, causal AI, a graph neural network, and advanced contact-tracing mathematics” developed during the pandemic to marketing analytics.

Details on where the DGX system is hosted were not shared. In a recent LinkedIn post, Puig said the system offers “256 petaflops of Nvidia horsepower,” but didn’t specify at which precision.

Founded in 2018, the startup has raised $26.8 million in funding to date, including $14 million in a Series A round that closed in February.

Nvidia is a named customer of Alembic. Nvidia CEO Jensen Huang has previously said: “Alembic is using AI developed for scientific research applications to predict ROI from marketing. Nvidia marketing is using Alembic with great success.”