FEATURE STORE SUMMIT 2026 · community survey

The Production AI Data Blueprint 2026

We are asking the people at the summit how they get data into their ML and AI systems, what keeps breaking, and what they plan to change. We publish the results in aggregate, with architecture examples from the teams who agree to talk to us.

Pick one system you work on yourself (in production or not) and answer for that one. If you do not use a feature store, your answers count just the same.

press Enter

01What does this system primarily do?

Other: what does it do?

02What stage is it at today?

Planning Prototype Pilot with real users In production for less than six months In production for six months or longer

03Which components currently supply its model inputs or agent context?

Select all that apply.

Products you use (optional)

04For its most time-sensitive input, how quickly must a change in source data become available to the system?

FreshnessHow long a change in the source data takes to reach the system. API response time is a separate question.

Under one second One second to under one minute One minute to under one hour One hour to under one day One day or longer No defined requirement

06For the most recent new data input, how long did it take from starting integration to production use?

07Across your team's predictive ML and LLM or agent systems, how much data infrastructure is shared today?

08What is the main infrastructure change your team has committed to delivering within the next six months?

CommittedIt has an owner and a delivery date.

Build or expand a feature platform Improve data freshness Improve serving performance Reduce operating costs Add agent-context infrastructure Consolidate existing systems Improve governance Other No committed change

Other: what change?

09What is one architecture decision you would change if you were rebuilding this system today, and why?

Optional. A sentence or two is plenty.

Only if you have no current system 02What would you build first?

Only if you have no current system 03What is holding you back?

Select up to two.

Last oneHow do you want to take part?

Add my details Stay anonymous

Each choice is separate.

Send me the Blueprint report when it is published I am open to a 30-minute architecture interview Credit me and my company by name in the report Send me Hopsworks news and events

Name

Company

Work email

We only publish aggregates. Anonymous answers are stored without a name, email or company. If you add your details, your answers are linked to them. Credit puts your name and company in the list of contributors, never next to your answers. We ask before we publish any quote or architecture detail.

THANK YOU · answers saved

Thanks, we have your answers.

Here is how they compare with everyone who answered so far.