In Episode 11 of the Magento Association Podcast, Artur Fridberg (co-founder of JetRails) and Nicanor Gioffré (Improntus) share a practitioner's perspective on self-hosted AI, open source ecommerce, and why predictability beats frontier models when you're running mission-critical Magento stores.

Artur started what became JetRails back in 2000 as a shared hosting company, eventually narrowing in on enterprise managed hosting for Magento and other open source ecommerce platforms. Nicanor has run Improntus on Magento and Adobe Commerce exclusively since version 1.3 — and today, the two teams are working together on something newer: hosting open source LLMs on private infrastructure for ecommerce clients with real compliance and privacy requirements.

This episode goes deep on the economics and mechanics of bringing AI in-house. Why token-based billing creates uncertainty agencies can't plan around. How freezing a model in place lets you update everything around it without breaking your stack. Why Artur thinks the future of software is a utility layer on top of resources — and how that reframes what Magento really is in an AI-first world. Along the way: open source as a market signal, the "exploding" requirements pipeline agencies are seeing, and a closing reflection on why the Magento Association still matters.

🔍 Episode Breakdown

0:00 – Welcome & JetRails OriginsArtur's journey from shared hosting in 2000 to a Magento-focused managed hosting brand.

2:30 – From Shared Hosting to Mission-Critical EcommerceHow JetRails evolved into "the anesthesiologist" for stress-free ecommerce ops.

5:00 – Partner, Not Vendor: How JetRails Works With AgenciesWhy JetRails acts like a department inside the agency, not an outside provider.

7:00 – Improntus Meets JetRails: AI Hosting Meets ComplianceNicanor on bringing JetRails an AI hosting problem most providers wouldn't touch.

9:30 – Why Predictability Beats Token-Based PricingThe case for cost and roadmap certainty when AI is part of your production stack.

14:00 – Open Source as a Market SignalArtur on why open source contribution patterns reveal where the market is actually going.

16:30 – The Uber Module & Magento's Open Source SpiritNicanor on shipping a real-time delivery module free to the community — and why.

23:30 – The Exploding Requirements PipelineWhat agencies are seeing as client expectations accelerate beyond anything from a few months ago.

26:00 – Plumbers, Coders, and Why Magento Is the PlumbingThe "learn to code" reframe and why Magento is the stable base layer for AI-era ecommerce.

30:30 – Freezing the Model in Place: How Self-Hosted LLMs WorkThe technical pitch for hosting your own open source models — and why stability beats frontier.

33:00 – Queueing AI Work Overnight: A Different Cost ModelRunning batch AI workloads on private infrastructure and what it changes about agency economics.

37:30 – Why the Magento Association MattersClosing thoughts from Artur and Nicanor on community, contribution, and ecosystem strength.

💬 Key Quotes

"We act like a department in your organization." — Artur

"We help your people not be woken up at 3 a.m. when a customer goes down." — Artur

"It's not so much the cost as the lack of control over your future." — Artur

"We've been with Magento since version 1.3. That's basically the only thing we do." — Nicanor

"Client expectations are much higher than even two or three months ago." — Nicanor

"I feel like I'm losing a tremendous amount of value if I'm not actually connected to the association." — Nicanor

#MagentoAssociation #Magento #MagentoCommunity #AdobeCommerce #Ecommerce #ArturFridberg #NicanorGioffre #JetRails #Improntus #SelfHostedAI #OpenSourceAI #LLMHosting #OpenSourceEcommerce