AI coding tools drive more software output, and downstream SDLC automation becomes the new bottleneck

Overview

This report revisits Morgan Stanley’s previous framework of “More Software and More Developers,” with the core conclusion that AI coding tools are increasing software development speed and output, but are not causing broad shrinkage in developer demand. As development costs fall, companies can launch more projects that were previously not economically viable, and developer roles shift from basic coding toward architecture design, results validation, systems integration, and governance. The report also notes that increased AI-generated code is pushing bottlenecks further downstream in the SDLC, increasing demand for more complete agentic AI and DevOps/DevSecOps platforms.

Core views

The report’s main views include: first, AI coding tools improve automation, but software output increases simultaneously, so the impact on developer seats is more structural than absolute; second, entry-level development roles face stronger pressure, while mid- to senior-level developers are more valuable because they can handle architecture, validation, and production complexity; third, after code generation efficiency rises, review, testing, integration, release, and security validation become the new productivity constraints; fourth, future competitive focus will shift from tools that help developers write code to platforms that can govern and automate the entire software life cycle.

Analysis framework

The report combines developer hiring data, software creation activity, AI tool adoption surveys, CI/CD workflow data, security testing outcomes, and public company business-model comparisons to assess AI’s impact on software development labor demand, SDLC bottleneck migration, and the effects on Atlassian, GitLab, JFrog, and similar companies.

Methodology notes

• Industry frameworkImpact of AI on the software development life cycle

  From code generation to end-to-end life-cycle automation

  The report breaks AI impact into upstream code generation efficiency gains and downstream delivery bottleneck migration, and argues that the true opportunity lies in automation of review, testing, security, release, and operations.

• Labor analysisChanges in developer demand structure

  Aggregate resilience with role-level divergence

  AI absorbs more entry-level and repetitive tasks, while at the same time expanding the feasible scope of development projects, increasing demand for mid- to senior-level engineers who can design systems, validate outputs, and manage complexity.

• Business-model comparisonSeat-based pricing versus usage-based pricing

  Software-platform monetization differences in the AI era

  The report argues that seat-based models remain under market pressure in the near term, while usage-based models tied to software output volume, artifact count, and governance needs benefit more directly.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• Atlassian Corporation PLC (TEAM.O)

  Core covered company, maintains Overweight

  Strengths

  Jira and Confluence are records systems for software planning and collaboration, with strong workflow context and enterprise collaboration entry points, and remain durable in AI-driven software development expansion.

  Weaknesses

  Re-rating now relies more clearly on cloud business re-acceleration, improved FY27 visibility, and AI monetization evidence; target price has been significantly lowered due to compressed valuation multiples.

  Comparison

  Compared with JFrog, Atlassian’s benefits depend more on product integration and monetization evolution beyond seat-based models, making near-term proof-of-execution less direct than usage-based models.

  Risks

  Slower cloud growth, insufficient AI monetization, compression of seat-based models, and continued valuation multiple decline.

• GitLab (GTLB)

  Public coverage DevSecOps automation platform

  Strengths

  Positioned in automation and the DevSecOps layer of the software life cycle, it is theoretically poised to benefit from the shift from code generation to full life-cycle governance.

  Weaknesses

  CY26/FY27 appears to be a transition period; investors need clearer execution and evidence that the Duo Agent Platform is restarting growth.

  Comparison

  Similar to Atlassian, it still needs to prove AI platformization and blended monetization; relative to JFrog, near-term direct benefits are less clear.

  Risks

  Execution shortfall, insufficient AI product adoption, delayed reacceleration of growth, and intensifying competition.

• JFrog (FROG.O)

  Viewed as the clearest direct beneficiary in the near term

  Strengths

  It charges based on growth in binary artifacts, packages, and software distribution rather than depending on developer seats; as a binary artifact repository, it has a moat and is expanding into security and governance.

  Weaknesses

  The market still has concerns about its security business, and it needs to continue demonstrating commercial traction in security and governance expansion.

  Comparison

  Compared with Atlassian and GitLab, JFrog has a more direct link to AI-driven software output growth, and its monetization benefit path is clearer.

  Risks

  Security business underperforming, underwhelming usage growth, and intensifying platform competition.

Key data

• Atlassian target price$120.00Reduced from $290.00 to $120.00, while Overweight was maintained.
• AI tool weekly usage rate75%-95%Multiple surveys indicate a high proportion of developers use AI tools at least weekly.
• Professional developer monthly usage rate93%-98%The report says AI adoption in engineering teams is near universal.
• Software developer job change+11%From around 49K in July 2025 to around 55K in February 2026, while total U.S. jobs declined 7% over the same period.
• Average workflow throughput+59% YoYCircleCI data shows code activity has risen significantly, but gains are concentrated in high-performing teams.
• Median team mainline throughput-7% YoYIncreased feature-branch activity did not broadly translate into higher mainline delivery.
• Mainline success rate71%Down to the lowest level in more than five years, below CircleCI’s benchmark of 90%.
• AI-generated code security test failure rate45%The Veracode 2025 GenAI Code Security Report shows that a substantial amount of AI-generated code still introduces major security vulnerabilities.
• AI-generated code security findings>10K/monthApiiro says that by June 2025, AI-generated code introduced more than 10,000 new security findings per month, a 10x increase over six months.

Impact & implications

The investment implication is that the market may be overestimating concern that AI will lead to a large-scale reduction in developer seats, but traditional seat-based platforms need to prove AI monetization and product-integration capability. The value of software platforms will increasingly come from control over workflow context, enterprise trust, governance, automation, and audit capabilities. JFrog is clearer as a near-term beneficiary due to direct linkage to binary artifacts and software distribution volume; Atlassian and GitLab remain in platform positions but need stronger cloud growth, execution evidence, and AI monetization proof to drive re-rating.

Risks

• If AI tools ultimately significantly reduce developer seats, growth for seat-based software platforms could be weakened.
• AI-generated code may bring higher security vulnerabilities, code review burden, and debugging costs that could offset productivity gains.
• Adoption of agentic AI remains constrained by trust, accuracy, governance, audit, and enterprise control requirements.
• Atlassian’s target price has already been lowered due to valuation multiple compression, and if cloud growth and AI monetization evidence are insufficient, re-rating may continue to be delayed.
• GitLab, Atlassian, and JFrog all face competitive pressure from large platforms and emerging agentic tools.

What to watch

• Whether developer hiring trends continue to remain more resilient than the broader labor market.
• Whether demand divergence between entry-level and mid- to senior-level developer roles is widening.
• Whether AI-generated code can translate into higher mainline throughput and actual production release improvements.
• Whether mainline success rate, recovery time, security vulnerabilities, and code review burden improve.
• Acceleration of Atlassian’s cloud business, FY27 visibility, and AI monetization progress.
• Adoption and blended monetization effect of GitLab’s Duo Agent Platform.
• JFrog usage growth, binary artifact governance, and security business progress.