Stonks: Prefer low starting multiple + margin expansion + robust topline

Joined April 2020
BetterIRR retweeted
Now I understand the fund performance
unavailable
3
3
2
100
27,165
Anthropic's S1 looks fine on the surface. Bear's argument was never that they couldn't show an OK quarter, rather the core arguments are: 1) existing ARR is inflated via a VC funding bubble 2) open source is an existential threat 3) if they misjudge demand their mandatory 1/2
8
2
77
19,602
spending commitments will bankrupt them.
18
2,277
Even if you pay with a convertable note, it would still be recorded as an expense no? Unless there is some funky SPV stuff.
Yikes. Anthropic just pulled the exact accounting trick that companies did in 2000 bubble for IPO. They hid all their expenses via convertible notes redeemable at IPO. $108 billion worth. Dotcom firms often paid suppliers, ad agencies, lawyers and even landlords with stock or equity to save cash. AOL swapped shares for services and ads (e.g., Homestore, DrKoop). Telecom suppliers like Lucent, Cisco and Nortel used vendor financing, sometimes with equity or warrants, to drive equipment sales.
6
20
10,246
The AI boom is narrower than almost anyone admits. The top 10% of firms account for 99.5% of all model-serving spend, 99% of neocloud, and 95%+ of OpenAI and Anthropic billings. And 80.4% of the heaviest AI spenders run on venture capital, not their own revenue. This is a concentrated, circular buildout. A handful of companies are the entire trade.
34
39
13
255
31,027
Is the AI boom a bubble outside Tech? 📉 Rev per worker: S&P 493 flat 📉 Profit margins: S&P 493 flat 📉 Productivity boost: None Non-tech sees AI token spend as a cost center, not growth. The macro AI story will be decided in the S&P 493, not the Mag 7. T at Apollo
12
25
15
134
15,550
This is a worse data set than ramp but tells the same story. All the AI growth (declerating as it is) is being driven by the top 1% VC funded companies. Ex AI lab VC funding is already 0.5T. This bubble only goes on if ever larger VC cheques are written.
👀Yipit coverage of Anthropic below
3
20
4,879
(And 500B is a lot already.) Either $NVDA (or someone?) starts parabolically funding VC firms or this is all over.
684
Modeling $REAX is an incredibly frustrating experience. That said I feel pretty good that we are close to 14x synergized pre-tax earnings (48M). Not the worst price to pay for a secular winner in REAL at the trough of the cycle.
3
8
2,832
If you think they can get there in say 5 years you're looking at a tripling of pre-tax income in 5 years. On top of that, you likely have a housing recovery sometime within those 5 years further juicing things. The base biz will also be growing, REAL had impressive share gains.
1
1
498
Anyone have a strong view on the name? Think it's a good setup.
1
496
Your hyperscalers are dead money for 2 to 3 years: -> Over earning across the board (implemented price hikes on non AI workloads via compute + memory constraints) -> Neocloud BKs will lead to distressed GPU pricing -> Equity investors will start paying attention to massive 1/x
3
2
1
69
14,206
Off balance sheet lease + obligations -> hypers will either have to keep funding the labs funding w/ no hope of return or do it themselves in house to not loose edge (perma a worse business than pre-ai) -> Over capacity comes with structurally higher cost base need absorb
2
11
2,469
Was it this easy to call the top in the dot com bubble?
22
1
3
82
20,611
When aggregate AI spend goes negative the real fun starts!
1,288
It's all fake. Meanwhile, token prices are down 25% m/m.
Great chart from Slok. Everybody knows that a very small % off companies are consuming the lion's share of token consumption What's interesting is that of those top 1% of spenders, they're overwhelmingly VC-backed companies. May have implications if VC-funding starts to slow.
2
7
2,467
What's the best way to play this? Want trough valuation @ trough margins.
Pending home sales fell to a record low in September: Lower than during a global financial crisis Lower than when the entire country was on lockdown for a deadly pandemic
16
1
19
9,205