Wall Street Wrote Business Software’s Obituary This Year. Somebody Forgot to Check With the Customers
An AI scare knocked Salesforce down about 30% early this year. Then it beat forecasts, raised its outlook and jumped more than 22% in a day. Software isn’t safe — but a headline is a poor reason to sell a whole sector.
By Sean Anees Saifi · Capital Wealth · Published Thursday, September 10, 2026 · Source: The Wall Street Journal, September 9, 2026 edition (Heard on the Street)
Key Points
An AI-fueled selloff knocked Salesforce (CRM) down about 30% early this year after Anthropic and OpenAI released strong coding tools.
Salesforce beat forecasts and raised its outlook last month on quarterly net profit of $3.53 billion, and its stock jumped more than 22% the next day.
It trades at roughly 15 times forward earnings, against its own 10-year average of 43 times and 24 times for an S&P software index.
The selling flared again Tuesday: ServiceNow (NOW) fell 5.0%, Workday (WDAY) 4.9% and Salesforce 3.9%.
The real risks, per the column: AI labs chasing growth ahead of IPOs could move into enterprise software, and AI-native startups may push shorter contracts and outcome-based pricing.
~30%
Salesforce’s AI-scare drop early this year
22%+
Salesforce’s one-day jump after earnings
15x
Salesforce forward P/E; 10-year avg. 43x
$3.53B
Salesforce net profit, quarter to July 31
Much of the software big companies rely on is customized, interconnected and costly to replace, which tends to slow how fast any new technology can displace it.
In one line: The evidence so far suggests AI makes business software a harder business, not a dead one — and a sector-wide scare is a poor reason to sell everything wearing the label.
Somewhere in a big company right now, someone is afraid to touch a piece of software nobody fully understands. That fear may be the most underrated asset in business software. Wall Street spent the early months of this year writing the industry’s obituary: after Anthropic and OpenAI released strong AI coding tools, a selloff knocked Salesforce (CRM) down about 30%. The selling flared again Tuesday, when ServiceNow (NOW) fell 5.0%, Workday (WDAY) 4.9% and Salesforce 3.9%. Heard on the Street columnist Asa Fitch thinks the funeral was premature.
Why the old code stays put
His case starts with what software work actually is. Most of it isn’t writing new code; it’s maintaining and adapting systems that already run the business, and AI agents still do that poorly. Big companies run tangled, interlocking software that’s costly to rip out — most corporate tech buyers, Gartner’s Arun Chandrasekaran notes, operate software rather than build it. Inertia buys time, too: even OpenAI’s Sam Altman has said disruption is arriving more slowly than he expected. So far, the numbers agree. Salesforce beat forecasts and raised its outlook last month, and its stock jumped more than 22% the next day; it also deepened a deal to build Anthropic’s Claude into its software. ServiceNow, Snowflake (SNOW) and Workday reported well, too.
Harder isn’t the same as dead
None of that makes software safe. AI labs hunting for growth stories to tell IPO investors could push into enterprise software themselves, and AI-native startups may force shorter contracts and pricing tied to outcomes. Fitch’s verdict is that AI makes the business harder without killing it, and he calls Salesforce a bargain at roughly 15 times forward earnings, against its own 10-year average of 43 times and 24 times for an S&P software index. That’s his stock call. The lesson underneath it travels better.
Narratives move whole sectors in a day; fundamentals answer one company at a time. When a scary story hits an industry, the reflex is to sell everything wearing the label, and that’s often how a headline hardens into a permanent loss. Software isn’t one bet. It’s a long list of businesses with different customers and very different odds of being replaced, from single-product vendors to giants like Microsoft (MSFT), which the Capital Wealth book holds. The better question isn’t whether AI reshapes the industry. It’s which customers would find leaving too costly to bother.
Obituaries are quick to write; earnings take a quarter or two to answer them. If a headline has you itching to dump a whole corner of your portfolio, spend fifteen minutes with what you actually own first — and with what, specifically, has changed in those businesses. A story is a reason to look, not a reason to sell.
What It Means For Your Portfolio
Hold — a scary narrative isn’t a reason to sell a sector
A narrative can knock down an entire sector in a day, while results arrive company by company; when a headline threatens an industry you own, the sound move is checking results and concentration, not selling everything with the label.
General planning principles, not advice for anyone in particular: selloffs driven by a single narrative tend to lump strong and weak businesses together. Before acting on one, it’s worth separating the story from the results — revenue, profit, whether customers are staying — and checking whether your exposure is concentrated in a few names or spread across the market through broad funds.
In the book, Microsoft (MSFT) is held and stays held, and the book doesn’t own Salesforce. A columnist calling a stock cheap doesn’t override the written rule — no new positions into Friday’s inflation report — and an AI scare isn’t a reason to sell what’s already owned. Nothing is bought or sold today.
