来自 Maksym Misichenko · CNBC ·
AI智能体对这条新闻的看法
The panelists debate the sustainability of Salesforce's growth and margin expansion as it transitions to a consumption-based model, with AI-driven headcount cuts potentially offsetting tokenized pricing benefits. The key concern is whether the 205% growth in Agentforce ARR can be sustained and if it will lead to margin expansion.
风险: Potential erosion of per-seat revenue as AI tools enable in-house builds and the shift to consumption-based pricing that could mute near-term billings.
机会: Successful navigation of the transition to a consumption-based architecture, allowing Salesforce to capture value from AI-driven efficiency gains.
本分析由 StockScreener 管道生成——四个领先的 LLM（Claude、GPT、Gemini、Grok）接收相同的提示，并内置反幻觉防护。 阅读方法论 →
Salesforce 的季度业绩未能说服持怀疑态度者，人工智能与其业务是朋友而不是敌人。但它增强了 Jim Cramer 对该股票的信心。“继续持有 Salesforce 是值得的，我们将这样做，”Jim 在星期三的 Morning Meeting 上说。此前，他在 CNBC 上表示，如果未受限制，他将购买更多股票。如果 Jim 在 CNBC 上谈论过某只股票，他必须在执行交易前等待 72 小时。Salesforce 的股价在企业软件制造商周三报告强劲季度后上涨了大约 1%。根据 LSEG 的数据，收入增长了 13.3% 至 111.3 亿美元，超过了市场预期的 110.5 亿美元。调整后的每股收益为 3.87 美元，比共识预期高出 76 美分。Salesforce 一直试图反击市场普遍存在的观点，即人工智能对其业务构成生存威胁。空头认为，由于人工智能效率导致人员缩减，该公司的人均席位业务面临危险，并且人工智能代码编写工具将导致客户用内部应用程序替换一些 Salesforce 工具。包括 Wells Fargo、UBS、Bernstein 和 D.A. Davidson 在内的几位分析师要么降低了对该股票的目标价，要么保持不变。但 Jim 确信，首席执行官 Marc Benioff 应该得到应有的认可，因为 Salesforce 正在悄然转向基于消费模式。该策略包括标记定价，客户可以在各项服务中兑换标记，作为一种更灵活的支付使用方式。他将 Salesforce 的人工智能驱动模型 Agentforce 指出为这种转变的证明。Agentforce 帮助客户构建可以自主执行任务的人工智能代理。Salesforce 表示，本季度完成了创纪录的 98 笔交易，年度经常性收入 (ARR) 现在达到 12 亿美元。与第四季度 8 亿美元相比，这增长了 205%。 “Agentforce 的收入现在超过了 10 亿美元，Jim。这太不可思议了。而且 Agentforce 现在已包含我们所有产品，从销售到服务，”Salesforce 首席执行官 Marc Benioff 在星期三晚上“Mad Money”节目中说。“没有一家企业软件公司比 Salesforce 做的更多。”Jim 将 Salesforce 的转型与另一家软件公司 Snowflake 的转型进行比较，Snowflake 的股价在周四的财报发布后飙升了近 39%。不仅 Snowflake 在周三晚间通过拥抱人工智能实现了超预期和提高预期，还承诺从亚马逊花费 60 亿美元用于计算。“[Benioff] 也在改变他的模式。他只是没有大声宣扬。”（Jim Cramer 的慈善信托持有 CRM。请在此处查看股票的完整列表。）作为 CNBC Investing Club 的订阅者，您可以获得交易警报，在 Jim 在其慈善信托的投资组合中进行交易之前，您将收到交易警报。Jim 在 CNBC 电视上谈论过某只股票后，会在发布交易警报后的 72 小时内执行交易。上述投资俱乐部信息受我们的条款和条件以及隐私政策，以及我们的免责声明的约束。由于您收到与投资俱乐部相关的信息，不构成或产生任何信托义务或职责。不保证任何特定的结果或利润。
四大领先AI模型讨论这篇文章
"Analyst price target cuts after the beat indicate the AI threat to Salesforce's core model is understated relative to Agentforce's still-minor revenue contribution."
Cramer's endorsement of Salesforce's consumption-based pivot via Agentforce overlooks that the $1.2B ARR remains under 3% of annualized revenue despite 205% growth. Multiple firms including Wells Fargo and UBS left targets unchanged or cut them after the 13.3% revenue beat, signaling skepticism that tokenized pricing offsets per-seat erosion from AI-driven headcount cuts. The 72-hour wait rule and charitable trust disclosure also limit immediate signal value. Broader enterprise software faces similar AI substitution risks that Snowflake's compute commitment does not fully address.
Agentforce could scale faster than modeled if enterprises adopt autonomous agents broadly, validating Benioff's quiet transition before peers and driving multiple expansion beyond current 11-12x forward sales.
"Agentforce is real but remains too small (11% of revenue) and too new to validate that it solves the existential headcount-reduction threat to Salesforce's legacy per-seat model."
Salesforce's beat is real—13.3% revenue growth, $0.76 EPS surprise, 98 Agentforce deals closed—but the article conflates Cramer's confidence with a solved problem. Agentforce's $1.2B ARR is impressive until you note it's 10.8% of total revenue; the 205% YoY jump is off a tiny base ($800M in Q4). The consumption-model pivot is strategic, but tokenization doesn't solve the core bear thesis: if AI agents reduce customer headcount needs, per-seat pricing collapses regardless of how flexible the billing is. Cramer's comparison to Snowflake is instructive—Snowflake committed $6B to compute spend, signaling conviction. Salesforce hasn't announced equivalent capex. The stock is up 1% post-earnings; that's muted for a 'strong quarter' and suggests the Street remains unconvinced.
Agentforce could be the inflection point—if $1.2B ARR compounds at even 50% annually while core business stabilizes, CRM rerates on growth narrative alone. The article omits that several large enterprise customers have publicly praised Agentforce adoption, which isn't priced in yet.
"The rapid scaling of Agentforce ARR proves that Salesforce is effectively monetizing AI agents, successfully offsetting the structural risks posed by per-seat licensing declines."
Salesforce (CRM) is successfully navigating the transition from a rigid per-seat model to a consumption-based architecture, which is critical for long-term relevance. The 205% year-over-year growth in Agentforce ARR to $1.2 billion validates that their AI agents are moving beyond pilot projects into production. While bears fear AI cannibalization, the shift to a consumption-based model allows Salesforce to capture value from AI-driven efficiency gains rather than losing it to headcount reductions. At current valuations, the market is severely underpricing the margin expansion potential of this pivot. If they maintain this momentum in enterprise adoption, a re-rating toward a higher multiple is likely as the 'AI-threat' narrative fades.
The transition to consumption-based pricing risks significant revenue volatility and margin compression if enterprise clients aggressively optimize their usage to cut costs during a cyclical downturn.
"CRM's AI pivot can drive meaningful, long-term ARR growth and margin upside, but only if Agentforce monetizes sustainably and the consumption-based model meaningfully outperforms the old per-seat model in a cooling macro."
Jim Cramer flags a real AI pivot at Salesforce: Q2 revenue $11.13B, +13.3% YoY; ARR $1.2B, up 205%; Agentforce cited as over $1B in revenue; tokenized pricing and 98 deals cited as proof points. Yet the AI narrative isn't proven to be durable. Risks include a shift to consumption-based pricing that could mute near-term billings, potential erosion of per-seat revenue as AI tools enable in-house builds, and stiff competition from Snowflake, Microsoft, Oracle. The piece glosses over margin trajectory and cash flow while couching optimism in marketing claims. The decisive question remains: can CRM sustain 2H growth and margin expansion as AI normalization unfolds?
Agentforce revenue exceeding ARR could be front-loaded or non-recurring. Even durable AI adoption may not translate into sustained margin expansion if pricing shifts prove unsustainable or customers push to in-house automation.
"Consumption volatility plus absent capex commitments risks stalling ARR growth despite deal momentum."
Gemini's margin expansion claim overlooks how consumption pricing amplifies revenue swings when enterprises optimize usage, a risk ChatGPT noted but Gemini downplayed in the against stance. Without Salesforce matching Snowflake's $6B compute commitment that Claude highlighted, the $1.2B ARR could stall even if 98 deals close. This explains why Wells Fargo and UBS targets stayed flat despite the beat, tying AI substitution fears to real billings uncertainty rather than pilot momentum alone.
"Salesforce's capex gap versus Snowflake is a category error; the real risk is whether consumption pricing creates unpredictable billings, not whether they're investing enough in compute."
Grok conflates two separate issues: consumption pricing volatility (real) and Salesforce's capex commitment (a false equivalence). Snowflake's $6B pledge funds infrastructure they sell; Salesforce doesn't need matching capex—they license agents. The real test is whether tokenized pricing stabilizes or swings wildly. Wells Fargo/UBS flat targets likely reflect billings timing uncertainty, not capex skepticism. That's a different bear case than the one being constructed here.
"Salesforce's pivot to consumption-based pricing will structurally compress margins due to their dependence on third-party hyperscaler compute costs."
Claude is right to call out the capex false equivalence, but both panelists miss the real issue: Salesforce’s reliance on hyperscaler infrastructure. Unlike Snowflake, CRM remains hostage to Microsoft and AWS margins. If Agentforce scales, their cost of revenue will balloon because they lack the proprietary compute layer to defend margins. The market is ignoring that Salesforce is essentially a high-margin software layer trying to pivot into a low-margin consumption model, which is a recipe for multiple compression.
"Hyperscaler margin drag could erode incremental profitability even as ARR grows, risking multiple expansion."
Gemini rightly flags hyperscaler reliance as a potential drag, but misses how that dependency could channel margin compression beyond a 'consumption' narrative. If Agentforce adoption scales, Salesforce will still pay for compute via partner clouds, and cloud-provider margins could squeeze CRM's incremental margin, not just capex. That risk isn't priced into flat targets; even with ARR rising, GM and COGS momentum could stall multiple expansion.
The panelists debate the sustainability of Salesforce's growth and margin expansion as it transitions to a consumption-based model, with AI-driven headcount cuts potentially offsetting tokenized pricing benefits. The key concern is whether the 205% growth in Agentforce ARR can be sustained and if it will lead to margin expansion.
Successful navigation of the transition to a consumption-based architecture, allowing Salesforce to capture value from AI-driven efficiency gains.
Potential erosion of per-seat revenue as AI tools enable in-house builds and the shift to consumption-based pricing that could mute near-term billings.