8090 Software Factory
by 8090 Solutions
The AI-native SDLC control plane: from business intent to production code, with a full audit trail.
8090 Software Factory is an AI-native SDLC control plane that links business requirements, architecture decisions, work orders, tests and feedback in one knowledge graph, so AI coding agents build from governed, auditable context. It is aimed at engineering and product leaders in regulated enterprises modernising legacy systems or building new software under compliance scrutiny.
8090 Software Factory is an AI-native software development lifecycle (SDLC) control plane from 8090 Solutions, the Menlo Park company Chamath Palihapitiya started in January 2024 as a self-funded incubator pitching enterprise software with 80% of the features at 90% less cost. Rather than competing as another code-completion tool, the Factory manages the context that coding agents consume. Work moves through five modules: Requirements, an agent-assisted workspace where product and business owners define intent and testable behaviour; Blueprints, which record architecture and technical decisions and flag drift between the plan and the code; Work Orders, which turn intent and architecture into implementation tasks grounded in a connected GitHub or GitLab repository; Tests, which tie verification back to requirements; and Feedback, which converts incoming feedback into structured development work. A shared knowledge graph links every artifact, so each change carries its rationale and an audit trail. Developers keep their own tools: an MCP connection with project-scoped API keys lets agents such as Cursor or Claude Code pull work orders and update their status from the IDE. Palihapitiya promoted a 1 September 2025 launch; in March 2026 EY US made it the engine of its EY.ai PDLC offering, to be rolled out to tens of thousands of consultants, and in June 2026 8090 announced a $135M Series A led by Salesforce. Self-serve pricing is $200 per user per month plus tokens; the 8090 Enterprise tier, starting at $1M a year, has 8090 design, build, host and maintain custom systems, including extracting business rules from 18 million lines of Medicare COBOL and Assembly for CMS. It targets regulated industries: healthcare, insurance, financial services, manufacturing and government.
A CTO or VP of Engineering in a regulated enterprise running legacy modernisation or AI-assisted development, who must be able to show auditors and the board why each change was made.
One traceable record linking business intent, architecture decisions and AI-written code, so agents build from agreed context and every change carries an audit trail.
At a Glance
- Category
- Developer Tools
- Pricing
- Subscription, Usage-based, Contact for pricing
- Target Market
- CTOs, CIOs, VPs of Engineering, Product Managers, Enterprise Developers
- Deployment
- Cloud-only
- Founded
- 2024
- Headquarters
- Menlo Park, United States
Key Features
- ✓Requirements module
Agent-assisted workspace where product, business and engineering owners co-author intent and testable behaviour, so coding agents start from agreed scope rather than a prompt.
- ✓Blueprints module
Records architecture and technical decisions as structured specifications; its agent guides design choices and flags drift between the documented design and the code.
- ✓Work Orders module
Converts requirements and blueprints into implementation tasks grounded in the connected codebase, packaging the context a coding agent needs to produce aligned code.
- ✓Tests module
Connects verification back to requirements before release, so teams can show which requirement each test covers and catch misalignment early.
- ✓Feedback module
Captures feedback from any source and automatically turns it into structured development work instead of an untriaged backlog of tickets.
- ✓Knowledge graph
Living documentation that links every requirement, decision and change with its rationale, giving auditors and executives a traceable record of what was built and why.
- ✓MCP connection for coding agents
Project-scoped API keys let Cursor, Claude Code or other MCP-capable agents pull work order details and update status directly from the IDE, avoiding lock-in to one tool.
- ✓GitHub and GitLab repository connections
Connecting a repository lets agents index the existing codebase, which grounds work orders in real code for both greenfield builds and legacy modernisation.
- ✓Real-time coauthoring
Multiplayer editing of specs and documentation lets product, design, engineering and QA work in the same artifacts instead of scattered documents.
Capabilities
Use Cases
- •Legacy system modernisation
Extract and document business rules from decades-old COBOL or Assembly systems before rewriting or decommissioning them; 8090 reports analysing 18M+ lines of Medicare claims code for CMS.
- •Governed AI-assisted development in regulated industries
Give healthcare, insurance and financial-services teams an auditable trail from business requirement to shipped code when AI agents write much of that code.
- •Consulting-led AI-native delivery
Systems integrators standardise delivery on one platform, as EY US is doing with EY.ai PDLC, planned for rollout to tens of thousands of consultants.
- •Reviving undocumented internal applications
Rebuild understanding of an abandoned internal system; customer Red Oak Development reports going from weeks to minutes to iterate on a change to a CRM.
- •Rapid prototyping of new products
Move a new product from written requirements to a usable prototype quickly; Utopia Studio reports three weeks to a working surgical mentoring platform prototype.
Ideal For
Best For
- ✓Regulated enterprises (healthcare, insurance, financial services, government) that need an audit trail for AI-assisted code changes
- ✓Legacy modernisation programmes that must extract and document business rules from old COBOL or Assembly systems before a rewrite
- ✓Engineering organisations already using Cursor or Claude Code that lack shared, durable context across product, architecture and implementation
- ✓Product and engineering leaders who want requirements, architecture decisions, work orders and tests tracked in one record
- ✓Systems integrators and consultancies standardising AI-native delivery for clients, as EY US has done
Not Ideal For
- ✗Small teams or solo developers whose repository documentation and prompts already give agents enough context; Ry Walker Research calls this a poor fit, and another $200-per-seat platform may not pay back
- ✗Buyers who need a fixed, predictable all-in cost up front, because model tokens are billed separately on top of the seat price
- ✗Teams looking for an autonomous coding agent out of the box: the Factory supplies context, work orders and traceability, and relies on an IDE agent such as Cursor or Claude Code to write the code
- ✗Organisations that require a self-hosted or air-gapped deployment, which is not documented on the public product, pricing or docs pages
Deployment
Market Analysis
Pros
- ✓Published self-serve list price ($200/user/month), unusual for an enterprise SDLC platform
- ✓Agent-agnostic: the MCP connection works with Cursor, Claude Code and other agents rather than locking teams into one IDE
- ✓Traceability from requirement to blueprint to work order to test suits audit-heavy, regulated environments
- ✓Serious enterprise validation: EY US built EY.ai PDLC on it, and 8090 runs a CMS Medicare modernisation engagement
- ✓Well capitalised after a $135M Series A led by Salesforce in June 2026
Cons
- ✗Tokens are billed separately from the $200/user/month seat price, so total cost cannot be estimated from the list price (Ry Walker Research)
- ✗Headline results are vendor- or partner-reported and not independently verified: TNW calls them the company's own figures, and EY's 70% productivity and 80x speed claims come from a single EY US use case
- ✗Almost no independent user-review footprint yet: no G2, Capterra or Product Hunt listing was found, and Hacker News has only a 2-point post for its agent harness
- ✗Does not guarantee correct generated code or prove that documentation cannot drift from the codebase (Ry Walker Research)
- ✗Compliance attestations (SOC 2, HIPAA, FedRAMP) are not stated on the public product, pricing or docs pages despite a regulated-industry pitch
- ✗Commentators question whether a 'factory for agents' is a durable product category or a slogan on top of a services business (TNW, HackerNoon)
Pricing
8090 Software Factory (self-serve)
From $200/user/mo
- ✓Full access to the Software Factory application
- ✓Guided onboarding
- ✓Standard support (documentation, guides, responsive assistance)
- ✓Tokens billed separately
- ✓More than 50 seats: contact sales
8090 Enterprise
From $1M/yr
- ✓Custom engagement for system modernisation or custom software development
- ✓8090 designs, builds, hosts and maintains the application
- ✓Sold via sales conversation
Self-serve Software Factory is list-priced at $200 per user per month, with model tokens billed separately on top, so the all-in cost depends on usage and cannot be read off the list price; deployments above 50 seats go through sales. 8090 Enterprise is a custom delivery engagement in which 8090 designs, builds, hosts and maintains the system, starting at $1M per year. No free tier, free trial or published token rates were found.
Security & Compliance
Connect
Sources
This page was written from 14 sources, 9 on domains other than 8090.ai.
- 1.8090.ai — 8090.aivendor
- 2.8090.ai — pricingvendor
- 3.8090.ai — software factoryvendor
- 4.8090.ai — customer storiesvendor
- 5.8090.ai — series avendor
- 6.docs.8090.ai — introduction
- 7.docs.8090.ai — quickstart
- 8.thenextweb.com — chamath palihapitiya 8090 135m series a ai coding
- 9.ey.com — ernst young llp and 8090 launch ey ai pdlc
- 10.rywalker.com — 8090 software factory
- 11.hackernoon.com — $135m for a dev shop im bullish on 8090s bet on the simplest
- 12.en.wikipedia.org — Chamath Palihapitiya
- 13.github.com — software factory harness
- 14.hn.algolia.com — search
Stay Ahead of the Curve
Weekly enterprise AI insights for technology leaders. No spam, no vendor pitches—unsubscribe anytime.
SubscribeRelated Products
HackerRank Chakra
AI interviewer that runs hands-on technical interviews and scores how candidates work with AI
Autoheal
Self-improving AI agents for incident response, vulnerability fixes and release work after code is written
Unified.to
Real-time unified API and MCP server giving AI products and agents access to 1,000+ SaaS integrations
Momentic
Agentic QA platform that writes, runs and self-heals end-to-end tests for web and mobile apps