Oracle and Salesforce have both sold off sharply, but for opposite reasons—one is spending heavily on AI infrastructure, the other faces fears that AI could disrupt its core business. At current valuations, Salesforce's strong profitability and triple-digit AI product growth may offer better value than Oracle's expensive expansion.

What happened: Oracle reported fiscal fourth-quarter revenue of $19.2 billion（約3.1兆円） (up 21% year over year) and cloud infrastructure revenue surging 93%, yet the stock fell after the company revealed it is spending heavily on AI data centers. Free cash flow was negative $23.7 billion（約3.8兆円） in fiscal 2026 as capital spending ramped up, and management expects to raise about $40 billion（約6.4兆円） more in fiscal 2027. Meanwhile, Salesforce shares have fallen about 37% year to date and touched a 52-week low, despite fiscal first-quarter revenue of $11.1 billion（約1.8兆円） (up 13% year over year) and its AI agent product Agentforce reaching $1.2 billion（約1900億円） in annual recurring revenue, up 205% year over year.

Why it matters: Investors are pricing the two companies very differently based on their AI strategies. Oracle trades at a price-to-earnings ratio of about 32 and a forward price-to-earnings ratio of about 23, reflecting confidence in its $638 billion（約100兆円） backlog of contracted revenue yet concerns about the debt burden required to deliver it. Salesforce, by contrast, trades at a price-to-earnings ratio of about 19 while generating substantial cash and returning $27.5 billion（約4.4兆円） to shareholders in its first quarter, yet the market appears to be pricing it as if AI will undermine its subscription software business—despite no sign of that erosion yet.

What to watch: Oracle's ability to convert its $638 billion（約100兆円） backlog into profit without additional equity dilution and Salesforce's execution on organic revenue growth in the second half of fiscal 2027, which management says will be driven by Sales, Service, Slack, Agentforce, and Data 360. The risk calculus cuts both ways: if AI demand compounds for years, Oracle could deliver far greater returns; if AI agents erode demand for traditional subscription software, Salesforce's discount could prove justified.