Oracle (ORCL)

Has good momentum.

Hyperscaler. Ran into issues with a series of announcements that amounted to running up a bill of $80B. So then they had to get financing and raise $20B in stock issuance. Reports next week. Price target of $258, and you won't find a much longer runway. Yield is 1.35%.

The market expects Oracle to put up as many data centres as it can, so any sign that they won't put up as many triggers a sell-off in the stock. Buy Nvidia instead.

Generated $32B of operating cashflow last year. FCF is negative because they spent roughly $56B building AI data centres. Investors are willing to tolerate that spending because of extraordinary demand. Wouldn't expect a FCF reversal while the buildout continues, but those new data centres will produce revenue. 

Rebound potential. Wait and see, earnings next week. She has only a 1/2 position now, and will see what the earnings report says.

He added shares 2 weeks ago. They enjoy a big backlog while no new debt is expected in 2026. If AI succeeds, so will Oracle. Their EPS is growing over 20%. It's a great time to buy.

Their debt is a massive red flag. They are spending billions to build AI data centres to become the most indebted hyperscaler. Avoid.

He's excited by the neocloud story. Trades at 18x forward PE. The software headwinds have weakened. If AI succeeds, Oracle succeeds.

They changed their business model the past year, from software to a hyperscaler by investing a lot in that area, but taking on debt. Prefers other stocks among the hyperscalers.

Bears are decrying the cost of building AI and asking if AI will generate that much business. After a sharp rise in April and May, the stock has fallen 51% since June 1. Meanwhile, its debt has risen to fresh highs. Oracle needs to make some money anywhere.

It is in its longest losing streak since 2021. It spent too much money and raised a lot of money through bond offerings and stock options. You could wait and buy at $145 and it may have found its support level. He has a 12 month price target of $250.

The stock has been killed lately, but is expected to have a $15 EPS on 2029 and over $20 in 2030 while trading at just over 7x PE. Cheap.

Prime reason for underperformance is debt. Investment in data centres is all about cost of capital. If you can't control costs, your business model doesn't really make sense.