AGI is here, context is the missing piece (with Caroline Hyde, Ed Ludlow)

Ali Ghodsi, CEO of Databricks, argues AGI has already arrived—models are smarter than most coworkers. The real bottleneck is lack of context/data integration, not model capability. Databricks focuses on infusing data context into AI agents. He predicts more software will be written in the next 1-2 years than in all human history, validating Jensen Huang's thesis.

Yields

implicit

RUT

Oil

Metals

USD

Databricks 3.0
Information Technology
Ali Ghodsi 8.5

AGI is here, context is the missing piece (with Caroline Hyde, Ed Ludlow)

ndx

Ghodsi validates Jensen Huang's thesis that more compute drives more revenue and that software demand will explode. He predicts more software written in 1-2 years than all human history, implying sustained demand for AI infrastructure and tech stocks.

Caroline Hyde

AI bubble, compute needs, SaaS apocalypse are key themes. Databricks is one of the most valuable private companies yet to file confidentially.
Ali Ghodsi

We already have AGI. Models are smarter than most coworkers. The problem is lack of context—if we could capture all context and feed it to AI, huge breakthroughs would happen. Databricks focuses on infusing data context into AI agents via Genie.
Ed Ludlow

Jensen Huang's equation: more compute ? more tokens ? more revenues. He says agents become the ultimate users of software, making the 'SaaS apocalypse' narrative nonsense.
Ali Ghodsi

Jensen is spot-on. We calculated that in the next 1-2 years, more software will be written than in all human history. This creates huge demand for the ecosystem around software, including databases like Databricks' Lake Base.
Caroline Hyde

As Databricks scales, do you need more private capital or will you go public?
Ali Ghodsi

We will be a public company. At our scale ($20B raised, 10,000 employees), public markets provide a transaction mechanism for employees. But this is a terrible year to go public—too much happening.

SignalTube

markets at a glance