Factory 2.0 reframes coding agents as end-to-end software factories
Factory.ai has announced Factory 2.0, shifting its product vision from individual coding agents to interconnected, self-improving "software factories" that span the entire software development lifecycle.
Score breakdown
Factory 2.0 reframes the enterprise AI coding market from point-in-time agent assistance to a self-improving, organization-wide system — a shift the post argues makes individual productivity tooling insufficient on its own.
- 01Factory.ai announced Factory 2.0 on June 15, 2026, authored by Matan Grinberg and Eno Reyes.
- 02The 'software factory' is described as a continuous feedback loop spanning signal ingestion, planning, building, testing, reviewing, securing, shipping, and monitoring.
- 03Three required pillars: Model Independence, Sovereign Intelligence, and Continual Learning and Self-Improvement.
Factory.ai, in a post authored by Matan Grinberg and Eno Reyes, has announced Factory 2.0, a significant expansion of its product vision. Where the company originally launched in 2023 to deploy autonomous Droids across the enterprise software development lifecycle, it now describes a broader ambition: the "software factory," an interconnected, agent-native, end-to-end system that observes itself and improves over time. The system is framed as a continuous feedback loop — ingesting signals from bug reports, internal conversations, customer feedback, and business requirements, converting them into planned changes, then building, testing, reviewing, securing, shipping, and monitoring the resulting software, with monitoring generating new signals to restart the cycle.
The post outlines three pillars a robust software factory must have: Model Independence (the ability to deliberately choose different models or use a Router to automatically select the best model per task, with the expectation that model costs will decrease as they commoditize); Sovereign Intelligence (full organizational ownership of the system, whether cloud-hosted, bring-your-own-key, self-hosted, EU-specific, or air-gapped, with every agent session, code review, and resolved incident feeding back into the organization's own loop); and Continual Learning and Self-Improvement (all lifecycle stages — code review, security analysis, documentation, QA, and incident response — running on a shared agent core, model router, and organizational context so that findings in one area inform others). Factory describes a spectrum of autonomy mechanisms including simple Droid agents, Automations for recurring workflows, Droid Computers for remote and persistent execution, and multi-agent "Missions" for complex tasks decomposed into parallel tracks running over hours or days. The company states that software factories are already in production at large organizations including NVIDIA, EY, Adobe, Palo Alto Networks, Adyen, Blackstone, Wipro, and Comarch.
Key facts
- 01Factory.ai announced Factory 2.0 on June 15, 2026, authored by Matan Grinberg and Eno Reyes.
- 02The 'software factory' is described as a continuous feedback loop spanning signal ingestion, planning, building, testing, reviewing, securing, shipping, and monitoring.
- 03Three required pillars: Model Independence, Sovereign Intelligence, and Continual Learning and Self-Improvement.
- 04A model Router automatically or rule-based selects the best model for each task across the factory.
- 05Sovereignty options include fully cloud-hosted, bring-your-own-key, self-hosted data plane, EU-specific, and fully air-gapped deployments.
- 06Autonomy mechanisms include simple Droid agents, Automations, Droid Computers for persistent execution, and multi-agent Missions for complex parallel tasks.
- 07Software factories are already in production at NVIDIA, EY, Adobe, Palo Alto Networks, Adyen, Blackstone, Wipro, and Comarch.
Topics
Summary and scoring are generated automatically from the original article. We always link back to the publisher and never republish images or paywalled content. Last processed Jun 16, 2026 · 23:11 UTC. How this works →