The Token Budget Problem Nobody Is Talking About (Matan Grinberg, Co-Founder & CEO of Factory)

2026-06-23 · Source: The Generalist · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Robotics & Autonomous Systems · Depth: Intermediate, extended

Summary

Matan Grinberg, co-founder and CEO of Factory, leads a \$1.5 billion company automating software development for enterprises like Nvidia, Morgan Stanley, and Adobe. Factory's "droids" are AI agents handling the full software development lifecycle, from coding to testing and security. Grinberg, a former Berkeley physics PhD, emphasizes the shift from building software to building "software factories" that optimize resource allocation for AI models. He highlights the critical "token budget problem" where organizations must strategically distribute compute resources across departments, moving beyond blind, uniform allocations. Factory's model-independent approach aims to abstract away the complexity of managing diverse AI models, allowing engineers to focus on higher-leverage tasks. The company has grown from 25 to over 100 people this year, doubling revenue month over month.

Key takeaway

For CTOs and AI/ML Directors grappling with AI integration, recognize that the "token budget problem" demands strategic, data-driven resource allocation, not uniform distribution. Prioritize solutions like Factory's model-independent "software factories" to automate low-leverage engineering tasks and optimize compute spend across diverse models. This approach frees engineers for high-impact work, ensuring better ROI and avoiding vendor lock-in, ultimately enhancing business agility and product quality.

Key insights

Factory automates software development using AI "droids" and "software factories" to optimize resource allocation and engineer leverage.

Principles

• Market timing: being early is equivalent to being wrong.
• Model independence improves agent performance across all models.
• AI resource allocation requires data-backed, non-uniform token budgeting.

Method

Factory's droids automate the full software development lifecycle, including coding, testing, reviewing, and security, by operating in sandboxed environments and routing across various AI models.

In practice

• Automate low-leverage engineering tasks like CI/documentation.
• Implement AI agents for regulatory compliance and security testing.
• Utilize model routing to optimize AI cost, performance, and latency.

Topics

• AI Agents
• Software Development Automation
• Model Routing
• Token Budgeting
• Enterprise AI
• Factory

Best for: Director of AI/ML, CTO, AI Engineer

Related on AIssential

• 'Nobody has budgeted' for tokenmaxxing, Box’s Levie says — Unbudgeted AI token costs, especially for coding agents, are a significant and growing financial challenge for companies.
• The Real Lesson from OpenAI’s Top Customers: Tokens Aren’t Spend. They’re Leverage — AI is becoming embedded infrastructure, automating cognition-heavy work across diverse sectors and roles, leveling the competitive landscape.
• The Annual AI Slowdown Panic is Here — AI's current "slowdown panic" reflects market adaptation to scarce compute and usage-based pricing, not a bubble burst.
• The AI Token Shortage Begins [AI Monthly Recap] — AI's next competitive phase is defined by effective access, affordability, optimization, and deployment of tokens amidst scarcity.
• Your AI Bill Is A Context Problem — AI budget overruns stem from "context debt" in agentic workflows, requiring a new discipline focused on context fidelity.
• TAI #208: Open Models Find Their Role as Agent Token Bills Rise — Rising agent token costs necessitate a tiered model strategy, using cheaper open models for high-volume, verifiable tasks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Generalist.