Key takeaways
Engineering production across six major technology companies increased 116% year over year, according to Navigara's analysis of commit-level activity from 676 engineers at Cloudflare, Vercel, OpenAI, Google, Meta, and Microsoft.
The increase was not evenly distributed. OpenAI's engineering production rose 373%, while Google's increased 56%, highlighting significant variation in how organizations are experiencing changes in developer volume.
More production does not necessarily mean higher productivity. While the data shows a shift toward growth-oriented engineering work, maintenance and fixes still account for most day-to-day activity, underscoring the limits of using production alone as a measure of business value.
For nearly two years, the technology industry has been debating whether AI is actually making software engineers more productive. Navigara’s study analyzed commit-level engineering activity across five quarters at six major technology companies: Cloudflare, Vercel, OpenAI, Google, Meta, and Microsoft.
The result is a large, public commit-level analysis of engineering production, covering 676 engineers and their software commits. Although no causality was established, the study found that production across the companies studied increased by 116% year over year.
The data reveals a significant shift in engineering activity that appears across multiple organizations, follows a distinct timeline, and changes the composition of the work engineers are doing. In an industry hungry for objective evidence, that alone makes the findings worth examining.
A five-quarter analysis of commit data reveals how engineering production changed across six major tech companies
Navigara analyzed five quarters of commit-level activity from 676 engineers across six hand-picked companies: Cloudflare, Vercel, OpenAI, Google, Meta, and Microsoft. The study examined observable development activity to identify shifts in engineering production, workload composition, and organizational trends over time.
A commit-level look across big tech
Rather than relying on surveys or self-reported estimates, Navigara’s study measured production using commit-level data collected over five consecutive quarters from public code repositories. Using commits as the unit of analysis offers a practical way to observe engineering behavior.
Every commit represents work being recorded within a software development workflow, creating a trail of measurable activity that can be analyzed consistently across teams and organizations.
What commit data can and cannot tell
Commit data provides an observable record of engineering activity across software workflows. It allows researchers to compare patterns over time, but it should not be treated as a complete measure of productivity, quality, or business value.
The data can show outcomes, such as how much engineering activity occurred, how that activity changed over time, and how work was distributed across different categories. It can also reveal whether developers are spending more time building new functionality, maintaining existing systems, or addressing fixes and operational issues.
What it cannot directly measure is quality, customer impact, revenue generation, or long-term business value. It also does not evaluate the internal products within the six companies, as it exclusively looks at publicly available information of peripheral work, like open-source tools, developer tools, and software development kits (SDKs). Finally, the tool can’t tell whether a commit was written by a human or with AI assistance.
Commit-level analysis shows engineering production increased 116% across six major technology companies in one year
The 116% figure is the study’s clearest signal, but it is not the whole story. Navigara’s quarter-by-quarter analysis shows that the increase was not a smooth climb throughout the year, but included a sharp inflection in engineering activity in Q1 2026 that warrants a closer look.
A sudden inflection, not a gradual trend
The top-line result from Navigara’s analysis is striking: Across the six companies studied, engineering production rose 116% year over year. However, the quarter-by-quarter pattern matters as much as the final percentage.
According to the analysis, the increase was not a smooth, steady climb. It was concentrated heavily in a single quarter, creating what looks more like a mostly flat curve with a sudden spike at the end. The five-quarter progression was +0%, +9%, +18%, +25%, and +116%, showing that the majority of the jump occurred in the study’s final quarter. Even when holding the engineer population constant at a fixed panel of 418 engineers active every quarter, production still rose by 98% without any new engineers joining the study.
A gradual increase might suggest slow-moving process improvements, growing teams, or normal organizational scaling. A concentrated jump suggests a shift in tooling, priorities, operating cadence, or some combination of all three.
The shape of the curve
A 116% annual increase showcases the magnitude of the change. The quarterly distribution shows how that change unfolded. If most of the gain occurred in a single period, leaders should be cautious about treating the trend as a simple, repeatable growth rate.
This is not evidence that engineering production will keep doubling every year. It is evidence that production, as measured through commits, experienced a major inflection during the period studied.
When production rises this dramatically across companies with different cultures, products, and engineering systems, it suggests that there may need to be a new way to measure productivity. The debate is no longer only about whether developers feel faster but whether their workflows are producing observable changes at scale.
Not all companies experienced the same growth, revealing significant differences in engineering trends across big tech
While the overall increase in engineering production is significant, the gains were not evenly distributed across the companies studied. Navigara's analysis found substantial differences in growth rates, suggesting that organizational structure, priorities, and development practices may influence how engineering teams respond to new tools and workflows.
OpenAI: The largest increase
The aggregate number hides substantial variation between companies. OpenAI recorded the largest increase in the study, with engineering production rising 373%.
That figure makes intuitive sense in one respect. OpenAI has been operating at the center of the AI boom, shipping new products, infrastructure, model capabilities, enterprise features, and developer tools at an unusually fast pace. A sharp rise in commit activity fits the broader picture of a company in expansion mode.
However, it should still be interpreted carefully. OpenAI’s cohort only entered the study in Q2 2025 and grew from six to 31 engineers, affecting the ability to interpret this finding as a stable trend. Additionally, a 373% increase in production does not automatically mean OpenAI engineers became nearly four times more productive. It may reflect rapid organizational scaling, changing product priorities, increased infrastructure work, or a larger number of active engineering initiatives.
Google: A different story
Google’s production rose 56%, a much smaller increase than OpenAI’s but still meaningful at the scale of one of the world’s largest engineering organizations. The contrast shows that even among companies with deep AI investment, elite engineering teams, and mature software systems, the pattern is not uniform.
Different organizations appear to be moving through this transition at different speeds. That variation may be the real takeaway. The impact of new tools and workflows is unlikely to appear evenly across the industry. It will depend on the company's structure, product maturity, technical debt, internal tooling, and how engineering teams choose to integrate automation into their daily work.
What engineering leaders, executives, and boardrooms can learn from commit-level data about modern software development trends
For executives trying to understand whether AI is reshaping software development, the most important takeaway from Navigara's research is not that engineering production increased by 116%. It's that the increase is large enough to demand attention.
The data suggests that something meaningful changed across multiple technology companies during the study period. Whether the catalyst was AI-assisted development, workflow improvements, organizational changes, or some combination of factors, the shift appears too substantial to dismiss as statistical noise.
Just as notable is the changing composition of engineering work. The analysis found a growing share of production being directed toward growth-oriented initiatives. Using a scoring system that measures the complexity of each contribution, Navigara explored both the quantity and quality of new commits. While commits per engineer rose 35%, the Engineering Throughput Value (ETV) rose 51%, indicating that developers were doing more work and each piece of work was more complex.
Data potentially suggested that engineers could be spending less time on maintenance and routine upkeep and more time on growth work. Overall, the data paints a picture of organizations that may be creating more capacity while also increasing the value of the additional work produced.
The future of engineering productivity debates will depend less on opinions and more on measurable operational data
The technology industry has spent years debating AI's impact on software development. Until recently, most of that debate has relied on anecdotes, surveys, and competing narratives.
Navigara’s commit-level analysis offers something different: observable data from 676 engineers across six major technology companies over five quarters. The findings reveal a 116% increase in engineering production, significant differences between organizations, and a noticeable shift toward growth-focused work.
It’s important to note that the data shows a sharp increase in commit-level engineering activity, but it does not isolate AI as the cause. The rise could reflect AI-assisted development, changes in hiring, shifting priorities, process improvements, or a mix of other factors.
Ultimately, the study suggests that engineering work is becoming more measurable. Leaders can use commit-level analysis to understand how outcomes, workload composition, and development priorities are changing, while recognizing that production is only one part of the productivity picture.
For an industry searching for evidence instead of opinions, that may be the most important contribution of all. The next chapter of AI productivity will be shaped by measurement and by the organizations willing to follow the data wherever it leads.
To learn more, you can read about Navigara’s findings on WebWire.
VentureBeat newsroom and editorial staff were not involved in the creation of this content.