Image: Factory logo mark, from Menlo Ventures' October 5, 2026 post announcing its investment in Factory.

AI software development is moving past the "write me a function" stage. On October 5, Menlo Ventures said it has invested in Factory. The startup wants to automate the entire software lifecycle, not just the coding part.

Factory is already valued at $5 billion. Its tools are used by hundreds of thousands of developers at companies like Nvidia and Blackstone, according to Menlo.

I coordinate software projects for clients in India and the USA, so I read this news with one question in mind. What does it change for agencies and the teams that build software for others? Below I cover what was announced, the drama around it, and what I think delivery teams should do now.

What Menlo Ventures Announced About Factory

Menlo partner Matt Murphy and his team shared the news in a blog post on Menlo's investment in Factory. A Business Wire release from Factory followed the same day.

Here are the main facts:

Menlo joined Factory's latest round. The firm did not share how much it put in.
The round came last month. On September 15, Factory said it raised $200 million at a $5 billion valuation.
The valuation jumped fast. Factory said $5 billion is more than triple the $1.5 billion it was valued at in April.
Total funding is now over $400 million. Backers include Sequoia Capital, Blackstone, NEA, Khosla Ventures, Insight Partners, Nvidia and J.P. Morgan.
Big customers are on board. Factory names Nvidia, Blackstone, Palo Alto Networks, Adyen and EY among its users.

Factory was founded in 2023 by CEO Matan Grinberg and CTO Eno Reyes. Grinberg left a theoretical physics Ph.D. at UC Berkeley to start it. Reyes was a machine learning engineer at Hugging Face. The company now has offices in San Francisco, New York, London and Japan.

Why the Timing Made Headlines

On its own, a VC joining a round is not big news. This one landed after a messy week. TechCrunch's Julie Bort reported that Grinberg said he had fired investor Chris Degnan of RPT Partners as a board advisor. Grinberg said he feared Degnan may have shared confidential information with Cognition, Factory's biggest competitor. Degnan had just joined Cognition as chief revenue officer.

Degnan disputed that account. He said he resigned. Then Vinod Khosla, whose firm backs both Factory and Cognition, called Factory a "struggling second tier competitor" and accused Grinberg of lying.

Murphy is best known for leading Menlo's pre-product bet on Anthropic. In Factory's release he said: "When we invested in Anthropic, there was plenty of skepticism that they could overcome larger competitors." He added that he sees "that same combination of principles, talent, and execution at Factory."

Menlo's post praised the founders, the product and the customer relationships. TechCrunch noted that Menlo is not an investor in Cognition. A source told the outlet the investment was significant for Menlo and not a token move.

What Factory Actually Builds

Most AI coding tools you know focus on one step: writing code. Menlo argues that was never most of an engineer's job.

In its post, Menlo says most development time goes to the work around coding:

Planning and breaking work into tasks
Code review
Testing
Security checks
Documentation
Coordination between people and teams

Factory's pitch is to run all of that as one system. Menlo calls it the software factory, or "the machine that builds the machine." Factory's own release describes a system that "runs 24/7 to build software under their control, with humans governing measurable outcomes."

Three design choices stand out for anyone who builds software for clients:

Shared context and guardrails. Every agent and user works from the same knowledge base, inside the same rules.
Model choice. Factory is model-agnostic. Customers, or Factory's router, pick which AI model runs, trading off capability against cost.
Where it runs. Factory says it can run in its managed cloud, on the customer's own servers, or in fully air-gapped setups.

How AI Software Development Is Changing

Menlo's post lays out three stages. I find it a useful way to explain this shift to clients who are not technical.

Stage	How it works	Who drives it
Autocomplete	The model finishes lines of code as you type	The developer, hands on keyboard
Coding agents	You hand off a scoped task or feature to an agent	The developer starts each session
Outcome-driven development	You set the goal, agents do the work, you review before shipping	The engineer as reviewer and owner

The money behind this is real. Menlo says the AI coding market grew from $550 million to more than $30 billion in two years. It now makes up almost half of all AI application spending, by Menlo's count. Menlo also says commits on GitHub are up 6x.

Grinberg summed up the goal in Factory's release: "We are moving from a world where engineers build software to a world where engineers build the factories that build the software."

Portrait of Menlo Ventures managing partner Matt Murphy smiling outdoors — Image: Menlo Ventures managing partner Matt Murphy, who led the Factory investment. Photo via TechCrunch's October 5, 2026 report.

A Reality Check on Adoption

Big funding does not mean every developer is on board. CNN recently covered a survey suggesting most developers try to avoid using AI. We wrote about that in our post on AI coding tools and the Stack Overflow developer survey.

Cost is another open question. Some large companies have started trimming their AI tool spending. We covered that in our breakdown of AI costs and the Claude cutbacks at Meta and Microsoft. Model choice and routing, which Factory pitches, matter more when budgets get tight.

What AI Software Development Platforms Mean for Agencies

Here is my honest view as someone who plans sprints and talks to clients every week. Platforms like Factory will not replace agencies. They will change what clients pay agencies for.

1. Clients will ask about it before you raise it

Enterprise buyers read the same headlines. Expect questions like "Do you use AI agents?" and "Why does this take six weeks?" in proposals. Have a clear, honest answer ready about where AI helps your process and where it does not.

2. Review becomes the bottleneck

When agents write more code, someone still has to check it. In our experience, senior review time is already the tightest resource on AI-assisted projects. Plan for that in staffing, not just in tooling.

3. Hourly billing gets harder to defend

If features ship in days instead of weeks, billing by the hour looks odd. Menlo says customers told it Factory ships features "in days rather than quarters." Fixed-scope and outcome-based pricing will make more sense for many projects.

4. Governance becomes a selling point

Regulated clients in banking, healthcare and government care where code and data live. Options like on-premises or air-gapped deployment are not extras for them. Agencies that can explain their controls clearly will win those deals.

A Practical AI Software Development Checklist for Teams in India and the USA

You do not need a $5 billion platform to get ready. Here is the checklist I now share with our delivery leads.

Map your lifecycle. List every step from ticket to release. Mark where AI already helps and where it could.
Start with review and tests. AI-written tests and first-pass reviews often save more time than code generation.
Write down your guardrails. Decide which repos, data and secrets an agent may touch. Get the client to sign off.
Keep a human gate before merge. No agent should push to a client's main branch without a named reviewer.
Follow a secure development standard. The NIST Secure Software Development Framework (SP 800-218) is a solid base for AI-assisted work too.
Track model costs per project. Know what each feature costs in AI usage, not just in hours.
Avoid lock-in. Prefer tools that let you switch models, so one price change does not break your margins.
Measure results. Track cycle time, defect rates and rework before and after adding agents.

Questions to Ask Any AI Development Vendor

Which models does it use, and can we choose?
Where does our code and context get stored?
Can it run inside our cloud or on our own servers?
What audit logs do we get for agent actions?
How are secrets and credentials kept away from the model?

What This Means for Businesses Hiring Developers

If you hire an agency or an offshore team, this shift affects you too. The best partners will not just say they "use AI." They will show you how it fits into their process and how they check its work.

Here is a simple example of the difference. One team says a feature will take three weeks of developer time. Another says it will take one week, with AI drafting tests and code, and two days of senior review. The second quote is only better if that review is real and the results are measured.

So ask for proof. Request sample pull requests, test coverage reports and a written policy on AI use. A good partner in India or the USA should be happy to share all three.

How We Use AI Software Development at Debugged Pro

We use AI tools on most projects now, but with clear limits. Agents help with boilerplate, tests and first-draft documentation. Architecture, security decisions and final reviews stay with senior engineers.

One pattern works well for us. We let an agent draft tests for a module first, then a developer writes or fixes the code until those tests pass. It keeps the human in charge of what "done" means.

Thinking about bringing AI into your product's build process? Our software development consulting team can help you plan a setup that fits your stack, budget and compliance needs.

Key Takeaways

Menlo Ventures invested in Factory on October 5, 2026, joining a round that valued it at $5 billion.
Factory aims to automate the full software lifecycle, from planning and review to testing, security and docs.
The deal followed a public dispute involving Factory's CEO, investor Chris Degnan and Vinod Khosla.
AI software development is shifting to outcome-driven work, where engineers set goals and review results.
Agencies should prepare now with guardrails, human review gates, cost tracking and clear client answers.

About the Author

Nitin Kumar is the Lead Project Coordinator at Debugged Pro. He manages software and AI delivery projects for clients in India and the USA, with a focus on practical rollouts that teams can maintain after launch.