Suraj Santlani

Enterprise software has quietly split into two accounting cultures, and the gap between them is where most of the valuation argument now lives. One group reports profits that exist under any definition. Salesforce converts about a fifth of its revenue into operating profit and hands most of the cash back to shareholders. The other group reports a loss, or something close to it, and asks you to look at an adjusted figure instead. The adjustment is almost always the same line. Staff are paid partly in stock, that cost never leaves as cash, and so it gets added back. At ServiceNow the charge was 1.96 billion dollars last year. At Snowflake it was larger than the operating loss itself. Neither treatment is dishonest. The cash really does stay in the business, and the dilution really is a cost to the people who own it. What I keep noticing is that the second group’s multiples are set against the adjusted number while the risk sits with the reported one. In a cheap market that distinction is academic. On a day like October 16 it is the whole story.

Something worth sitting with from the second quarter filings. Hedge funds added to most of the ten largest cloud software names through the summer, and Snowflake alone picked up 23 new holders. Then in a single October session every one of those ten fell further than the Nasdaq did, because the market turned on how AI infrastructure is being financed. Subscription revenue does not reprice in a day. What repriced was the multiple. Positioning and price are telling two different stories here, and only one of them is about the businesses.

The consensus is that Oracle is the AI winner in enterprise software and Salesforce is the roadkill. Look at what each one has actually promised. Salesforce’s problem is that it grows slowly, and slow growth is already sitting in a multiple of about 14 times next year’s earnings. Oracle’s problem is that it has sold computing capacity it has not finished building, with borrowed money. Slow and cheap is a known quantity. Fast and leveraged is a forecast. The honest bear case on Salesforce is sharper than the one people make. It sells seats, so if its own software agents do work that staff used to do, customers need fewer of them. That is a real risk and almost nobody is underwriting it. But the stock that needs everything to go right is not the one trading at 14 times. $ORCL $CRM

Oracle brought in roughly 47 billion dollars from operations last year and still finished with levered free cash flow of negative 45.9 billion. That is not a company funding growth out of profits. It is a company funding a construction project with debt, and the market has spent six months working that out. $ORCL

Adjusted earnings in cloud software add back the cost of paying staff in shares. At Palo Alto that charge was 1.71 billion dollars last year, more than twice what the business earned at the operating line. The adjustment is bigger than the profit it is adjusting.

hey @grok, tell me what happened in nyse and nashdaq this week. anything interesting?

people argue about $spcx's valuation. nobody argues about the growth: 92% revenue growth at $23b of revenue is not normal. not financial advice. dyor.

What's a better moat: brand or switching costs? I'd bet on the thing that's painful to leave over the thing people merely like.

A fortress balance sheet protects the company. It doesn't protect your entry price. Great business + wrong price = still a bad trade.

Forward P/E higher than trailing P/E is the most underrated warning sign in investing. It means the market expects the company to earn less next year. A "cheap" stock can be cheapest right before it gets worse. #stocks