HomeDemandAI Revenue: ARR and Products
ARR Path and Product Launches
ARR Path ($B, log) vs Public Launch Events#
On a log scale both companies' ARR climbs in steps, and each step lines up with a product launch.Draft, pending review
How to read this chart
Figures are annualized revenue as disclosed in the press, on a log axis. Shaded bands mark the main growth phases; vertical dashed lines mark product launches.
Source: FinSight compilation and estimates · Updated 2026-09-04
Standalone page ↗d1-arr-path-mon-01
New ARR per Month ($B) = What Data Centers Have to Absorb#
In three and a half years, monthly new ARR at either company has never turned negative; the worst case was only slower growth.Draft, pending review
How to read this chart
Each step is the average monthly increase between two consecutive disclosed figures. The dashed line is a breakdown of OpenAI's July, inferred from management's 20% month-on-month comment, for direction only.
Source: FinSight compilation and estimates · Updated 2026-09-04
Standalone page ↗d1-arr-path-mon-02
Event Scorecard (Breadth and Depth Together)#
How ARR and adoption responded after each product launch, one event per row.Draft, pending review
| Date | Event | Type | ARR response | Breadth response | Read |
|---|---|---|---|---|---|
| 22/11/30 | ChatGPT(O) | 表面〇 | +31%/月(23H1) | 2.1→17.6%(23/1:3) | ✔✔ 史上最大:2C 聊天 |
| 23/8/28 | ChatGPT Enterpri e(O) | 表面 | 無變化 | 常態 | ✘ 2C 包企業殼,未開新客群 |
| 24/11/25 | MCP 開源(A→產業) | 基建 | ◎ agent 表面時代的地基 | ||
| 25/2/24+5/22 | Claude Code 預覽→正式版 GA(A) | 表面 | +20→30%/月(25Q2) | +0.3→+1.0~1.3pp/月 | ✔ 新客群:開發者 |
| 25/4/16+5/16 | Codex CLI/雲端 agent(O) | 表面 | +8~12%/月不變 | +0.5~0.7pp/月常態 | △ 跟進型=撿殘餘:後有小凸起(+12→16%/月)後回落,遠小於首佔者脈衝;品質至 26/2 才轉正 |
| 25/10/6 | DevDay 開發者平台(O) | 表面 | +8~12%/月不變 | +0.5pp/月常態 | ✘ 補課型+品質未到位 |
| 26/2/5 | GPT-5.3 Codex(O) | 模型(品質轉折) | 增速未變 | 滲透仍緩降 | △ 社群反饋轉正、Codex 下載加速;OpenAI Signal 8/12 版以此月為基期:至 6 月 Codex 週活法務 108×/業務 41×/工程 5×(深度),同期 Ramp 家數滲透持平(廣度)→ 見企業採用度分頁「OpenAI Signal 職能滲透」 |
| 26/3/24 | 棄 Sora(O) | 策略 | ◎ 資源集中企業線 | ||
| 26/1/12 | Cowork 桌面版(A) | 表面 | +37→63%/月 | +0.5→+6.6pp/月峰值 | ✔✔ 新客群:全辦公室 |
| 26/4/9 | Cowork 正式版 GA(A) | 表面 | +42%/月延續 | +4.5pp/月 | ✔ 延續 |
| 26/6/9 | Fable 5(A) | 模型 | +42→+14%/月 | ✘ 定價2倍→token 6%/支出 11.4%(Sol:25%/23%);「前1%對聰明模型敏感」假說反向讀值,9月對帳 | |
| 26/7/7 | Cowork 網頁/行動(A) | 表面 | 落在 8:9 月讀值 | 7月 +1.1pp(脈衝已衰減) | ⏳ 9月讀值 >$75B=通過 |
| 26/7/9 | ChatGPT Work+5.6 Sol(O) | 表面+模型 | +20%/月(7月,說法) | 7月 +0.23pp 止跌 | ⏳ 8/25 Ramp 全量定判:廣度脈衝 or 存量深化 |
| 26/9/1 | Fable 5.1(A) | 模型(換代) | ⏳ 同價換代、cache read 降 75%、AA max 65.7 分;Ramp 9 月模型層份額對帳 | ||
| 26/9/3 | GPT-6 A tra(O) | 模型(新價位帶) | ⏳ 官方說取代 Sol、$20 與 Fable 同帶(板內比照 Fable 5 畫成獨立 SKU,Sol 仍在售)+Sol 同時限時降價 29%;ChatGPT/Codex 同步升級但沒有新表面,依假說不造台階;看 9:10 月 ARR 與 Ramp OpenAI 滲透 |
Source: FinSight compilation and estimates · Updated 2026-09-04
Standalone page ↗d1-arr-path-mon-03
