--- title: "CRM: Is the 'AI replacement' thesis busted? Can legacy software stage a comeback?" type: "Topics" locale: "en" url: "https://longbridge.com/en/dolphin/post/43616657.md" description: "Traditional SaaS bellwether $Salesforce(CRM.US) reported FY27 Q2 results for the quarter ended Jul on Aug 27 (pre-mkt), with headline numbers still muted. Both growth and core profitability were broadly in line with the Street, and trends remained steady.However, with the recent narrative reversal, steady prints at least suggest the 'AI replacement' thesis is overblown. Legacy SaaS names have seen a notable recovery.Specifically: 1) growth is still slowing. Subscription revenue, the core business, rose about 10.8% YoY on a nominal basis this quarter, with a clear QoQ deceleration..." datetime: "2026-08-27T02:33:43.000Z" locales: - [en](https://longbridge.com/en/dolphin/post/43616657.md) - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/43616657.md) - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/43616657.md) author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)" generator: "portal-rs" --- # CRM: Is the 'AI replacement' thesis busted? Can legacy software stage a comeback? The legacy SaaS bellwether $Salesforce(CRM.US) released Q2 FY27 results (through Jul) early on Aug 27. On the face of it, the print was muted, with both growth and core profitability roughly in line with Street expectations and trends broadly steady. Recently, as the narrative has turned, steady execution at least suggests the AI-replacement thesis was likely overblown, and legacy SaaS stocks have seen a notable rebound.Details below: **1) Growth is still slowing:** Subscription revenue, the core line, grew ~10.8% nominally, decelerating QoQ. On a constant-currency basis, growth was 11%, down roughly 1ppt QoQ. The revenue contribution from consolidating Informatica was similar to last quarter, indicating growth is indeed continuing to ease. By segment, the legacy applications in the **Agentforce Apps** bucket grew 8% cc, a modest 1ppt acceleration QoQ. However, the **data and platform bucket, which should benefit more from AI and had been the growth engine,** saw cc growth **decelerate by 3ppt.**In short, the AI-exposed legacy apps did not perform as badly and even showed slight repair, while the AI-favored areas where the company is focused did not deliver standout growth. **2) AI progress is steady, not explosive:** Consistent with the above, the company’s AI businesses (AI Agent under Agentforce and Data Cloud) reached $3.9bn ARR, steadily scaling to about 9% of total revenue. But growth slowed, with ARR up just 15% QoQ vs. 17% last quarter.Thus, AI is advancing steadily but has not hit an inflection, nor has it re-accelerated company-wide revenue. Within AI, the more application-driven Agentforce contributed about 38% of AI ARR, with a rising mix and stronger growth. Management attributed this mainly to recent momentum in **Slackbot** (the AI Agent within Slack) and **Headless 360** (front-end removed, callable via APIs by agents). **3) Leading indicators offered no upside surprise:** cRPO, the key metric, was $33.5bn, up 14% cc, a small 1ppt acceleration QoQ. However, total RPO and over-12-month RPO growth slowed.This implies part of the short-term cRPO acceleration came from shifting longer-duration RPO into the near term. Similarly, new bookings were about $9.9bn, up nearly 10% YoY, rebounding from last quarter’s low, but still growing more slowly than cRPO, again suggesting tenure mix was a driver of cRPO. **4) GPM under pressure, biz. model deteriorating:** With limited improvement on the top line, **gross margin continued to compress.** Subscription GPM was 81.3%, down QoQ, with YoY contraction widening to a recent high. This suggests AI, before reigniting growth, is already hitting margins due to higher costs, pointing to a deteriorating model. As a result, **total GP was about $8.7bn, up less than 9% YoY, missing Street expectations** and ranking among the lowest GP growth rates historically. **5) AI not only drags GPM but also lifts opex?** To make matters worse, margin pressure was not offset by tighter cost control. **Total opex rose about 12.4% YoY,** slightly below Street, but outpacing revenue growth, lifting the opex ratio. **R&D and S&M each grew by about 14% and 12%,** respectively. This likely reflects investment to build AI features and the need to promote and drive user adoption of these features. With GPM under pressure and a higher opex ratio, **GAAP OPM narrowed by over 2ppt, with OP at $2.3bn, flat YoY.** Net-net, AI has not accelerated growth, while profit gains have been consumed. **6) Shareholder returns fell back:** After a massive ~$25bn return last quarter, **buybacks dropped sharply to about $360mn,** even below this quarter’s stock-based comp. That said, last quarter’s repurchases drove a ~10% YoY reduction in total shares outstanding, effectively doing a year’s worth in one quarter. Also, this quarter is historically a seasonal trough for cash flow (about $1.1bn), limiting funds available for capital returns. **7) New reporting buckets:** Subscription revenue disclosure shifted from five lines (Sales, Service, Marketing & Commerce, Platform, Analytics) to two: **Agentforce Apps** and **Data 360 & Platform & Others.** **Agentforce Apps** are more SaaS-like, user-facing apps including Sales, Service, Slack, etc. **Data 360 & Platform** skews PaaS, providing the middleware and data services layer, including Data 360, the newly launched Headless Platform, and acquired businesses like Informatica and MuleSoft. **8) Core financial highlights below; earnings call recap linked** **Dolphin Research view:** **1) A lukewarm quarter** As discussed, Salesforce’s results were uninspiring and even somewhat negative. In brief: a) overall revenue growth is still decelerating; while AI ARR is scaling, from a small base its own growth has slowed and has not re-accelerated company revenue. b) AI’s higher compute needs make GPM declines immediate and notable, effectively shifting the model from light-asset SaaS with near-zero marginal costs to a heavier model where compute costs scale roughly with usage. Thus, with limited top-line uplift from AI, the impact of lower GPM and higher opex is clearly dilutive to profit. **2) Outlook improves, but not strongly** **Near term,** at the midpoint, **next quarter revenue growth is ~11.7%,** a modest acceleration vs. this quarter and slightly ahead of Street, with consolidation still contributing a bit over 4ppt. Revenue re-acceleration, long awaited by the market, could finally arrive next quarter, albeit modestly. **cRPO growth guidance at cc is 14%,** flat vs. this quarter. On profits, management guided **GAAP diluted EPS at $1.82 (mid),** down nearly 17% YoY and below the Street’s $1.92. Profit deterioration could be more pronounced next quarter.Overall, growth improves slightly while profit pressure deepens. **Medium term,** full-year FY27 revenue guidance was nudged up, but by **less than 0.5%.** Implied Q4 revenue growth would drop below 11%. **FY27 profit guidance was cut,** with **GAAP OPM down another 0.8ppt.** FCF growth is guided at 4%–5% YoY, again pointing to heavier AI investment weighing on margins.Net, a touch better growth but a more noticeable margin deterioration. **3) Recent move — acquiring Fin** In mid-Jun, the company announced the acquisition of AI Agent unicorn Fin for about $3.6bn. Disclosures indicate Fin (formerly Intercom) started as a small but full-suite SaaS across sales, marketing, and support, then **pivoted in mid-2023 to AI-powered customer support** with a proprietary base model, **Alex.** Currently, **ARR is about $400mn, with AI about $100mn,** implying a purchase at roughly **9x P/ARR,** not expensive. Dolphin sees this as broadly neutral: it strengthens Agentforce’s AI support capabilities at a fair price, but given Salesforce’s >$10bn quarterly revenue, Fin’s scale is too small to move the needle soon. The deal is expected to close by FY27 year-end, with no near-term P&L impact. **4) Valuation and investment view** We see no clear signs of fundamental improvement, and from a profit lens trends are getting worse. This is true for both the quarter and the outlook. Hence, the after-hours surge was not driven by fundamentals, but rather by a narrative shift and a re-rating of software multiples. More broadly, the **IGV** index has rebounded over 30% from the low, now within ~10% of its all-time high, retracing most of the AI-replacement drawdown. In recent quarters, legacy software firms like Salesforce showed middling results and slow AI transition, but crucially no evidence of severe AI-driven deterioration. In other words, the notion that AI would replace legacy software looks overstated. With fundamentals broadly stable, prior multiple compression tied to AI fears should reverse. **Salesforce’s edge is a lower multiple,** leaving more room for re-rating: on FY27 non-GAAP EPS guidance (~$16.7), the stock trades at ~14x P/E vs. ~26x avg. and ~20x median for the sector.We see meaningful near- to mid-term upside from multiple repair alone. That said, while AI has not directly hit results, the shift toward AI does impair the business model and long-term value. As noted, it shifts from near-zero marginal cost economics toward compute-intensive operations. Also, while AI will not fully replace enterprise software, it likely intensifies competition. Valuation can repair, but a return to the prior SaaS premium looks unlikely. \ **Prior Dolphin Research coverage of Salesforce:** **Earnings reviews:** Sep 4, 2025 review **‘**[**Salesforce: Is AI a Savior or a Death Knell?**](https://longportapp.cn/en/topics/33617374)**’** Sep 4, 2025 recap ‘[Salesforce (Trans): AI is an assistant, not a killer; SMBs are the new growth vector](https://longportapp.cn/en/topics/33623964)’ May 29, 2025 review **‘**[**Salesforce: AI agents still a distant dream?**](https://longportapp.cn/en/topics/30140966)**’** May 29, 2025 recap ‘[**Salesforce: SMB demand solid; pushing Agentforce**](https://longportapp.cn/en/topics/30145828)’ Feb 27, 2025 review **‘**[**Salesforce: Spend first, profit later with agents?**](https://longportapp.cn/en/topics/27617946)**’** Feb 27, 2025 recap **‘**[**Salesforce (Trans): The holy trinity of apps, data, and agents**](https://longportapp.cn/en/topics/27624545)**’** **Deep dives:** Jan 15, 2025 initiation II ‘[**Salesforce: Can a legacy SaaS sprout anew?**](https://longportapp.cn/en/topics/26625009)’ Jan 7, 2025 initiation I ‘[**Will AI replace ‘artificial’? How much can Salesforce gain?**](https://longportapp.cn/en/topics/26491224)’ **Risk disclosures and statements:** [**Dolphin Research disclaimer and general disclosures**](https://support.longbridge.global/topics/misc/dolphin-disclaimer) ### Related Stocks - [CRM.US](https://longbridge.com/en/quote/CRM.US.md) - [INFA.US](https://longbridge.com/en/quote/INFA.US.md) --- > **Disclaimer: This article is for reference only and does not constitute any investment advice.**