Releasing Hopsworks 5.0 – Introducing the Coding Data and AI Stack

Stockholm, Ju e 30, 2026 (GLOBE NEWSWIRE) — The team at Hopswo ks is excited to a ou ce Hopswo ks 5.0, the u ified a d Sove eig Data a d AI platfo m built a ou d the Codi g AI a d Data Stack; a ew pa adigm whe e codi g age ts eplace f agme ted tooli g to accele ate the developme t a d mai te a ce of ML pipeli es, data wo kflows, a d AI applicatio s a d much mo e.

It ma ks a ew chapte i how AI systems get built a d mai tai ed. Fo the fi st time, teams have a full codi g age t a d te mi al embedded i side the platfo m itself, p e-loaded with Claude Code a d Codex. The e is o co text switchi g as develope s desc ibe what they wa t to build, a d the platfo m ha dles the est f om i gesti g aw data to deployi g p oductio – eady dashboa ds a d i fe e ce pipeli es.

The elease, ge e ally available, is desig ed to help small a d la ge teams alike build, ite ate, a d ope ate AI pipeli es faste tha eve befo e f om batch a d eal-time systems to LLM-powe ed applicatio s.

Key capabilities of Hopswo ks 5.0:

Developme t Co tai e s, Codi g Age t a d Te mi al A te mi al a d developme t co tai e s  a e built di ectly i side Hopswo ks to allow age t to i st ume ts the platfo m, givi g develope s a u ified e vi o me t to build a d ope ate the full ML stack; featu e pipeli es, t ai i g pipeli es, i fe e ce pipeli es, St eamlit applicatio s, a d a alytics dashboa ds without leavi g the platfo m. Claude Code a d Codex come p e-i stalled, a d multiple age t sessio s ca u i pa allel, so develope s ca wo k ac oss diffe e t pa ts of a p oject simulta eously. Develope s desc ibe what they wa t to build i plai la guage, a d the age t w ites, u s, a d ite ates o the code. Sessio s a e fully pe siste t ac oss esta ts, so wo k is eve lost betwee codi g sessio s.

New Use I te face A completely edesig ed use i te face that delive s 50% lowe late cy ac oss the most commo platfo m tasks a d mea t to simplify use expe ie ce ac oss the whole platfo m, with the help of the Wiza d i te face.

• The Wiza d allows use s to select a y case, such as time se ies fo ecasti g, auto esea ch, a d othe s , automatically d ivi g th ough steps a d a guided setup that delive s a p oductio – eady pipeli e with a codi g age t. 
• Platfo m I tellige ce A i tellige t age t is embedded ac oss key platfo m wo kflows. Whe co ecti g exte al data sou ces f om Datab icks, S owflake, BigQue y, Mo goDB, SAP HANA, a d othe s, Platfo m I tellige ce uses a LLM to ge e ate huma – eadable colum ames a d desc iptio s, i fe p ima y keys a d eve t time colum s, a d co figu e i gestio schedules automatically. 

A alytics with T i o a d Apache Supe set Hopswo ks ow suppo ts ative SQL que y capabilities powe ed by T i o a d a full dashboa di g laye via Apache Supe set. Teams ca u explo ato y SQL que ies, build a d sha e dashboa ds, a d expo t esults as PDFs o PNGs without leavi g Hopswo ks. Batch i fe e ce pipeli es ca w ite p edictio s di ectly to featu e g oups, maki g esults immediately que yable a d visualisable.

Expa ded Data I teg atio  Hopswo ks 5.0 i t oduces a sig ifica tly expa ded set of data sou ces alo gside two ew ways to wo k with exte al data: mou ti g exte al tables without copyi g data, a d i gesti g data to Hopswo ks usi g DLTHub o a co figu able schedule. 

Colum -Level Access Co t ol Colum -level access co t ol e ables the sha i g of selected colum s f om tables (featu e g oups) to teams i diffe e t p ojects. I ML, it is commo fo data ow e s to ma age wide tables with some colum s (featu es) used i ma y models, but othe colum s co tai i g se sitive i fo matio . Colum -level access co t ol gives data ow e s, p ecise co t ol ove what colum s f om the (wide) tables a e sha ed, with whom.

Model Regist y E ha ceme ts Models ca ow be impo ted di ectly f om Huggi g Face, i cludi g the model ca d docume tatio , maki g it st aightfo wa d to b i g models i to o -p emise o ai -gapped deployme ts. Deployed models be efit f om Platfo m I tellige ce fo co figu atio , a d full p ove a ce li ki g t ai i g data, featu e views, a d model ve sio s is t acked automatically.

Read mo e about out the Hopswo ks 5.0 elease otes he e

“Hopswo ks 5.0 is ou most ambitious elease yet. It is a big step i movi g Hopswo ks f om a MLOps platfo m to a AI Lakehouse, with suppo t ow fo dashboa ds at scale, a ew que y e gi e, T i o, a d ative Apache Icebe g suppo t. But, Hopswo ks 5.0 is ot a me-too Lakehouse. It is a ew class of AI Lakehouse powe ed by codi g age ts. The low-code a d o-code tooli g that domi ates othe Lakehouses is obsolete. I stead, we have added fi st class suppo t fo buildi g a d ope ati g pipeli es, age ts, apps, a d dashboa ds f om a si gle Te mi al UI (a dev co tai e ). All this as we co ti ue to expa d ou lead i eal-time AI, with mo e suppo t fo eal-time t a sfo matio s i Ro DB.” — Jim Dowli g, CEO, Hopswo ks

Availability

Hopswo ks 5.0 is ge e ally available. Hopswo ks SaaS, which i cludes the full Codi g Data Stack expe ie ce a d the Wiza d, is ope fo   egist atio . Both o -p emise a d cloud deployme ts a e suppo ted fo e te p ise custome s, with GPU a d dist ibuted t ai i g via Ray a d Spa k fo la ge-scale wo kloads.

About Hopswo ks

Hopswo ks is the leadi g u ified platfo m fo buildi g batch, eal-time, a d LLM-powe ed AI systems, e abli g o ga isatio s to go f om aw data to p oductio – eady AI a d Data p oducts at scale. As a Eu opea -fou ded, sove eig AI platfo m, Hopswo ks is pu pose-built fo o ga isatio s that equi e data sove eig ty a d egulato y complia ce. At its co e is a powe ful AI Lakehouse, combi i g a best-i -class featu e sto e, model egist y, a d vecto database  cove i g the full pipeli e lifecycle f om data i gestio to batch, eal-time, a d age tic i fe e ce.