← sources

Alphabet shows a P/E of about 15. But, take out the gains on its stakes and I get about 27.

source: reddit (reddit:ValueInvesting)published 2026-11-02 19:33ingested 2026-11-03 02:43

retained text

I think the cheap P/E on Alphabet is misleading right now. At $284 today, Yahoo shows about 15x on $19.15 of EPS.

Most of that EPS came from marking its stakes in other companies. Gains on equity securities were $37B in Q1 and $99B in Q2. Q3 flipped to an $8.7B loss, and net income dropped from $112B to $26B while the business kept growing. When I strip those swings out of the last four quarters I get about $10.60 a share, so more like 27x.

27x isn't crazy here. Q3 revenue grew about 25%, and Cloud doubled at a 36% margin.

The cash is what bothers me. FCF was -$5.9B in Q2 and -$9.1B in Q3. Capex over four quarters was about $160B, 5.7x depreciation, and the company says it rises again in 2027. There has been no buyback for three quarters, and the share count is up 1.7% in nine months.

I ran the DCF backwards when I wrote my report on 29 October, with the stock at $325. That price needed $120-125B of cash profit every year. After tax, operations brought in around $127B over the last four quarters, so you reach that number if capex shrinks to roughly the level of depreciation. With capex settling at 19% of revenue, which is my assumption, I get about $60B.

My Monte Carlo median in that report is $151, around 14x that $10.60. It looks low for a company growing like this, and I'd attack it first if I were you. Most of the gap comes from that 19%. If this capex earns a proper return, I'm too low. The stock has fallen to $284 since then and I haven't rerun the model.

The full report with sources:[Complete report](https://waitingisthework.com/reports/googl/?ref=reddit) 

Research, not advice.

What share of revenue do you think capex settles at once the build slows down?

document fetched 2026-11-03 00:45 · canonical https://reddit.com/r/ValueInvesting/comments/1xm71fj/alphabet_shows_a_pe_of_about_15_but_take_out_the