August 2026 Summaries

Factory has announced the Factory Partner Network and a $100 million investment aimed at helping consulting firms, systems integrators, cloud providers, ISVs, and AI labs deploy AI-enabled software engineering capabilities in enterprises. The company argues that while AI models can generate code, greater value comes from managing the full software development lifecycle, including planning, testing, security, deployment, monitoring, and governance. The network will initially focus on consulting and SI partners through Factory Partner Strategists, who support enterprise planning, business cases, pilots, and adoption, and Factory Partner Engineers, who are trained to prepare, deploy, enhance, modernize, and scale Factory platforms. Factory says qualifying customers may receive support for early engagement activities, while partners will gain access to training, certifications, sales, marketing, engineering, and customer-success resources. A small, invitation-only founding cohort across the United States, EMEA, and APJ will be selected based on enterprise reach, technical capability, leadership commitment, and growth potential.

Factory has opened applications for the Factory Guild, a selective community for students, independent builders, researchers, creators, and technical community leaders interested in building with its Droids platform and influencing the future of agent-native software development. Applicants are evaluated by a Guild Council for curiosity, initiative, originality, and community commitment, with contributions including projects, technical content, events, educational materials, research, and member support. Members begin as Apprentices and can advance through Craftsman to Artisan based on contribution quality, consistency, impact, and value to other builders, earning benefits such as Droid credits, merchandise, mentorship, early feature access, event invitations, and career support. Factory describes the Guild as a feedback loop in which members test Droids in real projects, share successes and challenges, and help guide product development, while exceptional contributors may be considered for potential Factory roles.

Factory has introduced Agent Effectiveness, a private-preview feature in Factory Analytics that links Droid and Factory session activity to engineering delivery outcomes, helping leaders assess AI return on investment beyond usage metrics and surveys. Drawing data from connected Jira, Linear, GitHub, and GitLab integrations, it provides Throughput, Output, and Attribution views to show changes in cycle times, categorize work by intent such as feature development or maintenance, and connect spending and sessions to specific projects, issues, pull requests, and users. Administrators can enable the capability by configuring organization-wide integrations and activating the Advanced Analytics enterprise control, which supplies local attribution signals and backfills historical effectiveness data. Factory positions the feature as a way for organizations to monitor whether AI-assisted capacity is improving delivery and serving strategic priorities while there is still time to adjust resources.

Factory has introduced local deployment of its autonomous software engineering platform on NVIDIA DGX Spark using NVIDIA Nemotron 3.5 Lightning, targeting organizations that must keep source code, operational data, and execution within controlled environments. The platform can investigate issues, modify code, run tests, and prepare changes for human review, supporting uses such as legacy modernization and maintenance of critical, regulated, financial, and government systems. Its local deployment model supports Nemotron and other approved open-weight models, allowing enterprises to govern and customize AI workloads within their own networks, including air-gapped environments. Factory describes Nemotron 3.5 Lightning as a 30B mixture-of-experts model with 3B active parameters designed for high-throughput, always-on agents, while emphasizing that agentic workflows can route different tasks to specialized or frontier models. Example use cases include banks responding to production issues without externalizing sensitive data and government organizations modernizing applications in approved secure environments.

Factory announced that it has signed the Open Weights and American AI Leadership letter alongside Microsoft, NVIDIA, and more than 230 other organizations, arguing that U.S. AI leadership depends on an open ecosystem rather than concentration among a small number of closed models. The company describes itself as model-agnostic, supporting both open and closed providers, and highlights its role in developing AGENTS.md, an open format intended for AI agents. It also cites security-focused open-source work, including releasing components of its Droid Shield 2.0 secret-scanning models, auditing open-source projects through its Automated Security Review program, disclosing CVE-2026-42876 in External Secrets Operator, and joining the NVIDIA-launched Open Secure AI Alliance. Factory emphasizes that open models should be accompanied by safeguards such as threat modeling, red teaming, sandboxing, and human approval for high-risk actions under its Safe Autonomy Readiness Policy.

Factory’s Organization Model is designed to help enterprises deploy AI development tools across complex, multinational, and regulated business structures while retaining centralized governance. It separates a centrally managed enterprise account, which handles billing, identity, and administration, from nested organizations that define operational boundaries for users, integrations, credentials, policies, model access, spending limits, and data residency. Organizations can mirror departments, regions, subsidiaries, projects, client delivery groups, regulated environments, or deployment stages, allowing groups with distinct security, compliance, residency, or access needs to operate independently under one commercial relationship. Enterprise administrators use the Enterprise Admin Console to create and oversee the hierarchy, manage SSO, SCIM directory synchronization, domain verification, deployments, usage, and organization-level controls without managing each group locally. The model supports configurations such as customer-hosted inference or bring-your-own-key access for sensitive workloads while allowing other teams to use Factory-managed models, aiming to scale AI-enabled software development without requiring companies to compromise existing governance structures.