The Twenty Minute VCClickHouse CEO: AI Margins Need to Improve | Revenue Concentration Should be a Concern
At a glance
WHAT IT’S REALLY ABOUT
ClickHouse CEO on AI durability, margins, agents, and infrastructure shifts
- Katz argues AI is still early and accelerating faster than prior platform shifts, with unprecedented revenue growth and new technical demands from agentic applications.
- He believes the biggest AI investing risk is revenue durability because switching costs are low for many agentic apps, even as infrastructure software tends to have high switching costs.
- He expects AI margins can be tolerated while growth is exceptional, but companies must show credible paths to margin expansion as token costs fall and efficiency improves.
- He predicts agents will become active infrastructure selectors and will require identity, authorization, and budgets—creating a governance layer that doesn’t exist yet.
- ClickHouse’s growth thesis centers on low-latency, efficient analytics for exploding query volumes, with strong retention/expansion dynamics and deliberate moves across cloud and on-prem deployments.
IDEAS WORTH REMEMBERING
5 ideasAI’s current wave is more compressed and operationally demanding than prior cycles.
Katz compares AI to internet/mobile/social but says agentic products are maturing faster and stressing systems with highly concurrent, exploratory workloads rather than predictable human dashboard queries.
Margin concerns matter less than credible margin expansion while hypergrowth persists.
He’s relatively relaxed about low gross margins in AI-native businesses if they maintain strong balance sheets and can show a path to improving margins as token prices trend downward.
Durability of revenue is the critical risk for many AI applications.
Switching costs can be very low for agentic apps and model-driven products because providers are leapfrogging quickly; investors should scrutinize how defensible and “sticky” usage really is.
Token spend should be governed by outcomes (roadmap velocity and revenue), not fear of the bill.
ClickHouse prioritized accelerating engineering output (“AI Awakening”) and treats token cost as controllable later—especially since per-token cost is falling and product velocity is pulling roadmaps forward by years.
Agents will drive new requirements: ultra-low latency plus efficiency at massive scale.
Katz frames agent workloads as dozens of simultaneous SQL queries across systems, making the slowest link define the experience; infrastructure winners must ingest, store, and query at lower cost than incumbents.
WORDS WORTH SAVING
5 quotesWe're just getting started. This seems to be accelerating at a unprecedented pace.
— Aaron Katz
Would be durability of revenue, because the switching costs that you and I talked about are very high for infrastructure software. The switching costs can be very low for agentic applications.
— Aaron Katz
That you can overestimate what you can achieve in one year and underestimate what you can achieve in five.
— Aaron Katz
We can talk about the fact that agents will need to have an identity... they'll need to have a budget, and how do you authorize an agent to consume a service?
— Aaron Katz
If I've got one customer that, or one category or one industry that accounts for more than 10% of revenue, I spend a lot of time thinking about it and that seems to be kind of the industry standard, that threshold that if some dimension of your revenue base accounts for more than 10% of your revenue, you've got exposure and I wanna limit exposure, right?
— Aaron Katz
High quality AI-generated summary created from speaker-labeled transcript.
