Weave Raises $13.5M Series A to Measure AI Engineering ROI

Weave has raised a $13.5M Series A led by Standard Capital, with Y Combinator, Moonfire, Burst Capital, IrregEx, and Agent Fund participating. The round brings Weave's disclosed funding to $17.7M across two announced rounds.

The San Francisco company is building an engineering intelligence platform that measures work from prompt to production. Weave connects AI tool usage and cost with engineering output, code quality, and delivery data, then uses that information to help companies determine where AI is creating value and where it is simply creating a larger bill.

This funding matters because AI coding has moved from experimentation to a real operating expense. Engineering leaders no longer need another dashboard celebrating token consumption. They need evidence that human engineers and software agents are producing better work together.

What Happened

Weave announced its Series A on July 28, 2026, with Standard Capital leading and Y Combinator, Moonfire, Burst Capital, IrregEx, and Agent Fund participating. The 16-person company plans to use the capital for product development and go-to-market expansion.

The financing follows Weave's $4.2M seed round, announced in July 2025 and led by Moonfire and Burst Capital, with Y Combinator participating. Weave is also part of Y Combinator's Winter 2025 batch, reflecting how quickly the company has progressed from demonstrating that engineering work can be measured to building a broader AI accountability platform.

Weave was founded in 2024 by Adam Cohen, CEO, and Andrew Churchill, CTO. Churchill was previously employee number one at Causal and studied computer science and mathematics at MIT. Cohen's operating thesis begins with a simple imbalance: sales organizations are managed against measurable outcomes, while engineering organizations are often evaluated through proxies that reveal little about value or quality.

Why This Matters

AI coding tools made activity inexpensive before they made outcomes easy to measure. A team can generate more code, open more pull requests, and consume more tokens without knowing whether delivery improved, defects increased, or experienced engineers ultimately had to rewrite low-quality output.

That creates a new management challenge. Token volume appears precise, but it can reward the same behavior as measuring developers by lines of code. Quantity becomes the score simply because quantity is easy to count. Weave is betting engineering organizations will pay for a measurement layer that connects AI cost and usage with the work that survives review and reaches production.

The customer is not limited to the CTO. Finance leaders increasingly need visibility into AI subscriptions and usage-based costs, while engineering leaders need to compare those expenses against output, quality, and cycle time. A shared operating view gives both sides a way to evaluate returns without turning the discussion into a contest between model enthusiasm and spreadsheet anxiety.

The Product Bet

Weave's core platform analyzes engineering work from prompt to production. Its AI ROI tooling tracks subscription costs, per-user spending where available, adoption patterns, and the relationship between AI usage and engineering output.

The platform also measures how teams use interactive coding assistants and autonomous agents. Weave's agent observability product distinguishes foreground tools used directly by engineers from background agents that execute tasks, open pull requests, or resolve bugs with less direct supervision.

That data supports a model-routing strategy. Weave says its router classifies prompts and directs work toward the most cost-efficient model without sacrificing speed or quality, but the more important sequence is measurement before optimization. A lower-cost model is not actually cheaper if its output generates enough rework to consume senior engineering time.

The company also offers Wooly, an AI engineering agent that allows leaders to query organizational data and receive recommendations about where teams or workflows can improve. The broader strategy is straightforward: measure the work, explain the economics, then help organizations act on the results.

Market Context

Weave is grouped alongside engineering analytics and AI infrastructure companies such as getDX, Jellyfish, LinearB, and OpenRouter. Those companies do not solve identical problems, but together they illustrate a market shifting beyond AI adoption metrics toward cost control, quality, governance, and measurable return on investment.

Weave says more than 500 engineering organizations use its platform, including teams at Robinhood, PostHog, Telnyx, and Reducto. The company's website also reports activity across millions of pull requests and tens of millions of prompts, providing an expanding dataset for comparing how AI-assisted engineering performs across organizations.

That dataset could become a meaningful competitive advantage if Weave can keep its metrics transparent and trustworthy. Measuring engineers is both politically sensitive and technically difficult, particularly when those measurements can influence hiring, compensation, or technology budgets. Methodology will matter as much as the sophistication of the underlying models.

What This Signals

The Series A reflects a belief that the next phase of AI coding will be judged by economics and outcomes rather than novelty. Companies have already learned how to buy coding assistants and autonomous agents. The harder question is which tools deserve to remain, which workflows require improvement, and whether lower inference costs actually reduce the total cost of engineering.

For operators, the lesson is to establish a measurement system before optimizing the model stack. Teams need visibility into useful output, code quality, delivery, rework, and cost before they can intelligently route work or compare one model against another.

For Weave, the opportunity is to become the accountability layer between AI vendors, engineering organizations, and finance teams. The $13.5M Series A gives Adam Cohen, Andrew Churchill, and their team more room to prove that the smartest AI budget is not the one that purchases the most tokens, but the one that can explain what every token changed.