AI tools now generate an estimated 41% of committed code, and that code carries roughly 1.7x more bugs than human-written code. Every bug becomes an incident someone has to catch. PagerDuty (PD) sells the automated response to those incidents, yet the stock fell 23% the day it reported its first profitable year. Two forces are repricing it now: a forced pivot from seat licenses to consumption credits, and a $173.9M GAAP profit headline that is almost entirely a one-time tax benefit.
GitHub logged roughly 1 billion commits in 2025 and is on pace for ~14B in 2026. More code means more bugs, and more bugs means more incidents to detect and remediate.
The same tools generating the code (Claude Code, Cursor, Copilot) now plug into PagerDuty's SRE Agent through MCP. The wave that breaks production is wired directly into the platform that fixes it.
Of the $173.9M GAAP net income, $169.2M is a one-time tax benefit. Strip it out and underlying GAAP net income is about $4.7M.
The business model is being rebuilt mid-flight.PagerDuty is migrating from per-seat licensing to consumption credits because its own automation removes the human seats it used to bill. Revenue recognition gets far less predictable, which is why DBNR slid from 106% to 98%.
Seventy-nine customers now carry the company.The $1M+ ARR cohort grew 10% to 79 accounts while total paid customers shrank. That is durable enterprise lock-in and concentration risk at the same time.
PagerDuty runs its agents on OpenAI and Anthropic models it does not own, a dependency the 10-K now flags directly. Defensibility comes from 15+ years of operational data and 86B ingested events.
On a non-GAAP basis the company looks deeply undervalued. On an SBC-adjusted basis, where nearly 100% of free cash flow funds buybacks to offset dilution, the base case implies roughly fair value. The blended target lands near $15 (+58%).
38 pages. Two valuation frameworks, three scenarios. The AI code explosion connected to a half-billion-dollar revenue model.
AI tools now generate an estimated 41% of committed code, and that code carries roughly 1.7x more bugs than human-written code. Every bug becomes an incident someone has to catch. PagerDuty (PD) sells the automated response to those incidents, yet the stock fell 23% the day it reported its first profitable year. Two forces are repricing it now: a forced pivot from seat licenses to consumption credits, and a $173.9M GAAP profit headline that is almost entirely a one-time tax benefit.
GitHub logged roughly 1 billion commits in 2025 and is on pace for ~14B in 2026. More code means more bugs, and more bugs means more incidents to detect and remediate.
The same tools generating the code (Claude Code, Cursor, Copilot) now plug into PagerDuty's SRE Agent through MCP. The wave that breaks production is wired directly into the platform that fixes it.
Of the $173.9M GAAP net income, $169.2M is a one-time tax benefit. Strip it out and underlying GAAP net income is about $4.7M.
The business model is being rebuilt mid-flight.PagerDuty is migrating from per-seat licensing to consumption credits because its own automation removes the human seats it used to bill. Revenue recognition gets far less predictable, which is why DBNR slid from 106% to 98%.
Seventy-nine customers now carry the company.The $1M+ ARR cohort grew 10% to 79 accounts while total paid customers shrank. That is durable enterprise lock-in and concentration risk at the same time.
PagerDuty runs its agents on OpenAI and Anthropic models it does not own, a dependency the 10-K now flags directly. Defensibility comes from 15+ years of operational data and 86B ingested events.
On a non-GAAP basis the company looks deeply undervalued. On an SBC-adjusted basis, where nearly 100% of free cash flow funds buybacks to offset dilution, the base case implies roughly fair value. The blended target lands near $15 (+58%).
38 pages. Two valuation frameworks, three scenarios. The AI code explosion connected to a half-billion-dollar revenue model.
AI tools now generate an estimated 41% of committed code, and that code carries roughly 1.7x more bugs than human-written code. Every bug becomes an incident someone has to catch. PagerDuty (PD) sells the automated response to those incidents, yet the stock fell 23% the day it reported its first profitable year. Two forces are repricing it now: a forced pivot from seat licenses to consumption credits, and a $173.9M GAAP profit headline that is almost entirely a one-time tax benefit.
GitHub logged roughly 1 billion commits in 2025 and is on pace for ~14B in 2026. More code means more bugs, and more bugs means more incidents to detect and remediate.
The same tools generating the code (Claude Code, Cursor, Copilot) now plug into PagerDuty's SRE Agent through MCP. The wave that breaks production is wired directly into the platform that fixes it.
Of the $173.9M GAAP net income, $169.2M is a one-time tax benefit. Strip it out and underlying GAAP net income is about $4.7M.
The business model is being rebuilt mid-flight.PagerDuty is migrating from per-seat licensing to consumption credits because its own automation removes the human seats it used to bill. Revenue recognition gets far less predictable, which is why DBNR slid from 106% to 98%.
Seventy-nine customers now carry the company.The $1M+ ARR cohort grew 10% to 79 accounts while total paid customers shrank. That is durable enterprise lock-in and concentration risk at the same time.
PagerDuty runs its agents on OpenAI and Anthropic models it does not own, a dependency the 10-K now flags directly. Defensibility comes from 15+ years of operational data and 86B ingested events.
On a non-GAAP basis the company looks deeply undervalued. On an SBC-adjusted basis, where nearly 100% of free cash flow funds buybacks to offset dilution, the base case implies roughly fair value. The blended target lands near $15 (+58%).
38 pages. Two valuation frameworks, three scenarios. The AI code explosion connected to a half-billion-dollar revenue model.
.svg){kind=logo}
