Factory Commits $100M to Partner Network, Pushes to Scale Software Factories

The starting point for this announcement is a statistic the company has been citing. Even with the latest models, actually writing code accounts for only about 20% of the full software development lifecycle. The remaining 80% involves deciding what to build, designing it, testing it, checking security, deploying it, and monitoring it afterward. Citing research firm Gartner, Factory said teams that apply AI tools across the entire lifecycle—rather than just for code generation—saw productivity gains of 25 to 30%.

Why Partners Are Needed

"Value is created when intelligence is deployed." That's the line Factory used to frame this announcement. The company pointed out that Anthropic, OpenAI, Google, and Microsoft are pouring massive investment not just into the models themselves but into service capabilities, arguing that the model is only the starting point—actual deployment is what creates enterprise value. At the same time, Factory noted that these AI labs spend hundreds of times more on AI infrastructure than on service-oriented companies, and it sees that gap as the opening for consulting and SI partners to fill.

Factory laid out four things consulting and SI partners bring that models cannot: trust built over years, people with empathy, creativity, and drive, engagement structures that leave intellectual property with the client, and accountability for outcomes rather than token consumption. At the same time, the company said these partners also need to change, since projects that once took large teams quarters or years can now be completed by small, software-driven teams in weeks to months—meaning billing is shifting from hourly labor rates to outcome-based pricing.

Two Tracks: Strategists and Engineers

FPN launches around two pillars. The first is Factory Partner Strategists (FPS), which helps partners plan and run enterprise adoption of software factories. This includes executive briefings and events, business value consulting, use-case assessment and roadmapping, proof-of-value (PoV) and production pilots, and certification training. Factory said it will cover the cost of these activities for qualifying clients.

The second is Factory Partner Engineers (FPE), where partners embed directly with client teams to design, deploy, and operate software factories. Factory broke this journey into six stages.

Stage	Description
Readiness	Assessing adoption readiness, security planning, value baselines, deployment roadmap
Learning	Role-specific hands-on training and certification for teams that will run the factory
Deployment	Production rollout of infrastructure, software, data, security, and governance aligned to cost and value
Enhancement	Strengthening workflows and data with integrations, skills, droids, and policies
Modernization	Addressing technical debt in legacy systems, mainframes, and data platforms
Expansion	Broadening the scope of transformation across functions, applications, teams, and roles

Factory also said that routing tasks across multiple models can cut token costs by more than 50% while maintaining the same quality.

Selection Process and Timing

The first partner cohort will be recruited across the US, EMEA, and APJ. Selection will be by invitation, based on enterprise reach, engineering capability, leadership commitment, and growth ambition, the company said. Factory said it intends to keep the cohort deliberately small so it can give each partner substantive attention, with the first cohort forming within this quarter. The company added that it plans to share what it learns from this process with the broader ecosystem.

Editor's Take

This announcement is best read as a signal that the code-generation race has moved to a new stage. Over the past year or two, AI coding tools competed on how fast and how accurately they could generate code. But if writing code really is only 20% of the development process, as Factory itself states, then any tool that fails to cover the remaining 80% will stall at the pilot stage. In practice, enterprises both in Korea and abroad that have tried adopting AI coding tools tend to hit the same wall—prototypes come together in a day, but pushing them through security review, legacy integration, and operational rollout takes months. Factory's bet is to close that gap not with its own workforce, but through the hands of consulting and SI partners.

For Korean companies, the practical takeaway is clear. Organizations considering adoption of AI coding tools should look less at benchmark scores and more at whether the tool bundles deployment, security, and governance together. This is especially relevant for industries like finance and manufacturing, where moving source code outside the organization is difficult—Factory's August rollout of a local-run environment via DGX Spark and this partner network announcement both point in the same direction: a "software factory that stays entirely within the enterprise." That said, since the partner network is starting small and by invitation only, whether Korean SI firms will actually gain access to the program remains to be seen once the first cohort takes shape this quarter.

What happens over the next few months likely comes down to two competing paths: model developers like Anthropic, OpenAI, and Google building out their own service organizations, versus deployment-focused startups like Factory teaming up with consulting and SI firms. As the gap in model performance narrows, the outcome may ultimately hinge on who can push AI deeper into enterprises faster and more cheaply.