What is Databricks?

Why a cheaper-per-token model can cost more per task: lower-capability models require more iterations and reasoning tokens to reach the same result, making the nominally expensive model the economical choice at the task level.

“However, Sonnet cost around $2 per task or 2.09 per task versus 1.94 for Opus. So, because Sonnet needed more iterations and more reasoning had to spend way more tokens to get to the same results, overall Opus, which is significantly on paper more expensive model, it was cheaper to operate.”

Databricks processes 50 to 60 million virtual machines per day across three clouds.

“We maybe 50 or 60 million virtual machines a day across our three clouds.”

Over 90% of databases at Databricks Neon are created by AI agents, not humans.

“At data bricks neon, which is their superbase competitor, over 90% of the databases are built by agents, not by humans.”

Databricks Neon: a Supabase competitor where >90% of database creation is already agent-driven, making it a leading indicator of agentic infrastructure adoption.

“At data bricks neon, which is their superbase competitor, over 90% of the databases are built by agents, not by humans.”

Neon, acquired by Databricks, is scaling serverless Postgres to 13 million database launches per day, making it a key infrastructure bet on agent-era ephemeral compute needs.

“I think 13 million databases a day now.”

Databricks argues that unifying storage (via L-TAP) delivers 99% of the benefits of HTAP without requiring a single database to handle both OLTP and analytics workloads simultaneously.

“We think you can get 99% of what you need by unifying the storage.”

Opus 4.8 cost $1.94 per completed coding task vs $2.09 for Sonnet 5, despite Sonnet being cheaper per token (1.7x), because Sonnet required more iterations and reasoning tokens (Databricks benchmark).

“However, Sonnet cost around $2 per task or 2.09 per task versus 1.94 for Opus. So, because Sonnet needed more iterations and more reasoning had to spend way more tokens to get to the same results, overall Opus, which is significantly on paper more expensive model, it was cheaper to operate.”

Neon (acquired by Databricks) launches approximately 13 million databases per day.

“I think 13 million databases a day now.”

Arvind Jain says AI enterprise automation failures are a human organization problem, not an AI problem.

“Initial attempts at automate a lot of the software engineering at Databricks kind of failed even there's nothing wrong with the AI. The problem is the humans and how we were organized.”

Open-source model self-training pipelines, where the same model generates its own training environments, can outperform frontier proprietary models like Opus and GPT-5.5 on specific tasks, challenging the assumption that frontier closed models hold a durable edge.

“We have pipelines just using open-source models. Like the same model generates training environments and trains itself and beats like Opus and GPT 5.5 and stuff at a task.”