BBP

Pronoia is the belief that the world is quietly conspiring in your favor.

Stocks, AI, energy & credit. I follow the invoice.

[BBP @BBPronoia]

Kosmos Energy $KOS : +179.1% YTD.

i called it on July 8. More barrels, lower costs. Now the cash needs to pay down the debt.

The thesis, in four sketches 👇

[BBP @BBPronoia]

🩸 Comeback Watch: $BTC, Uptober edition

$86K. Still 31% below the all-time high.

That high was set last October. So was the start of the fall.

Improving: best Q3 since 2017, $2.65B of ETF inflows in September.
Still open: 10-year above 5%, and 2026 ETF inflows under $1B net.
Uptober giveth. Uptober taketh away.

Bounce or comeback?

[BBP @BBPronoia]

@usehotbot incredible work man, could I get an invite?
I’m holding!

[BBP @BBPronoia]

And
To who?

[Bull Theory @BullTheoryio]

🚨 Global debt now exceeds $365 trillion.

That's roughly 310% of global GDP, an ALL TIME RECORD.

Now the cost of servicing that debt is exploding everywhere at the same time.

US 5-year Treasury: 5.136%, highest since 2007 (19 year high)

US 10-year Treasury: 5.340%, highest since 2002 (24 year high)

US 20-year Treasury: 5.730%, highest since 2002 (24 year high)

US 30-year Treasury: 5.681%, highest since 2002 (24 year high)

UK 5-year Gilt: 5.062%, highest since 2008 (18 year high)

UK 10-year Gilt: 5.510%, highest since 2007 (19 year high)

UK 30-year Gilt: crossed 6.029%, highest since 1998 (28 year high)

Japan 10-year bond: 3.126%, highest since 1996 (30 year high)

Japan 30-year bond: 4.223%, a record high since Japan started issuing this bond in 1999

Germany 10-year bond: 3.65%, highest since 2009 (17 year high)

France 10-year bond: 4.963%, highest since 2002 (24 year high)

Australia 10-year bond: 5.448%, highest since 2011 (15 year high)

The entire world is drowning in debt.

[BBP @BBPronoia]

Expected free cash flow growth in next 5 years:

Micron $MU: +4,797%
Lumentum $LITE: +2,206%
Amazon $AMZN: +1,817%
SK Hynix $000660: +1,436%
Astera Labs $ALAB: +1,329%
Broadcom $AVGO: +1,056%
Palantir $PLTR: +933%
Arm $ARM: +903%
AMD $AMD: +825%
Marvell $MRVL: +719%
Credo $CRDO: +600%
NVIDIA $NVDA: +554%
Celestica $CLS: +472%
Vistra $VST: +349%
TSMC $TSM: +326%
Sterling Infrastructure $STRL: +319%
Applied Materials $AMAT: +278%
Onto Innovation $ONTO: +254%
GE Vernova $GEV: +251%
Lam Research $LRCX: +215%
Dell $DELL: +212%
Vertiv $VRT: +207%
Microsoft $MSFT: +191%
Arista Networks $ANET: +190%
Quanta Services $PWR: +180%
KLA $KLAC: +153%
ASML $ASML: +138%
Apple $AAPL: +95%
Tesla $TSLA: +94%
Meta $META: +55%
Alphabet $GOOGL: +50%

[BBP @BBPronoia]

Nike $NKE crashes to new 10-year low after missing earnings.

[BBP @BBPronoia]

💰 Insider Bets — October 1

my favourite management presentation is a purchase receipt.

Four insiders. Three companies. Here’s where they put their money:

🎮 $GME — Ryan Cohen: c. $10.6M
GameStop’s CEO bought 450,000 shares on September 29 at c. $23.48, taking his direct holding to 40.95M shares.

A big cheque. Still only a c. 1.1% increase in his existing position.

My test: can the operating business generate sustainable cash? Cohen buying more doesn’t answer that for us.

🧬 $XENE — CEO + CFO: c. $1.68M
On September 30, Ian Mortimer bought 30,000 shares at $37.38. CFO Thomas Kelly bought 15,000 at c. $37.32.

The context matters: Xenon recently paused new enrollment in psychiatry trials following adverse events. Its epilepsy studies continue, and its drug application for focal seizures has been submitted to the FDA.

My test: regulatory progress in epilepsy and clarity on the psychiatry safety issues. Two executives buying cannot replace clinical evidence.

🏘️ $HHH — Marc Grandisson: c. $1.60M
Vantage’s executive chairman bought 25,000 Howard Hughes shares on September 23 at c. $64.12, increasing his direct holding by c. 35%.

HHH completed its $2.1B acquisition of insurer Vantage in June. Someone close to that business is increasing his exposure to the parent.

My test: underwriting profitability and investment returns. The holding-company story needs numbers.

The purchases get these names onto my research list. What happens in the business keeps them there.

Which deserves a deeper dive?

Trade dates: September 23–30. Disclosed September 24–October 1, 2026.

[BBP @BBPronoia]

⚡ Who Gets Paid? $MU edition

Micron made $54.2B in one quarter. Gross margin: 87%. Somebody is paying that bill. Here's who.

Data centers and cloud AI → $34.3B.
The AI build-out is the obvious payer.

Phones and PCs → $13.1B, at a 90% margin. The same margin as data centers. That cost lands somewhere. Eventually, probably on you.

26 big customers → $32B committed to lock in supply through 2030. Buyers are paying today to make sure they get memory tomorrow.

Equipment makers → $11.5B of Micron capex next quarter alone. Every memory shortage becomes someone else's order book.

Yesterday I asked if it was too late to buy. Last night the customers answered: they're still paying.

Which one surprised you?

[BBP @BBPronoia]

@RJCcapital 🤣🤣

[BBP @BBPronoia]

Now it's $GOOG, but OpenAI will be back in top in weeks, then Anthropic again... and so on.
On the while we'll see 4 more Chinese disruptions.
"artificiar interigence"

[BBP @BBPronoia]

@KawzInvests

[KawzInvests @KawzInvests]

OpenAI can’t catch a break lol.

Claude Opus 5.5 leads Artificial Analysis. Now Google’s Gemini 4 Argon matches GPT-6 Astra’s score.

Cost per benchmark task:
Opus 5.5: $5.98
Astra: $3.26
Google: $1.99

Google is 39% cheaper than Astra and 67% cheaper than Opus.

[BBP @BBPronoia]

This is literally the best model in the world. Congrats, Google. $GOOG

[BBP @BBPronoia]

🧨 Too Late to Buy? $MU edition

In July I called Micron the place where AI FOMO goes to meet God. The FOMO won. Up 555% in a year. Now a $1.2 trillion company. Report is out.

The bull case: it looks cheap. Around 7x next year's expected profit. Analysts expect c. $150 a share, up from $44 over the last twelve months.

The catch: memory always looks cheapest at the top. 2018 peak profit: $14.1B → $2.7B two years later. 2022 peak: $8.7B → a $5.8B loss the next year. Today: $50.5B, almost 6x the last peak.

The business: never been better. The price: needs this boom to be different.

Too late? Tell me why I'm wrong.

[BBP @BBPronoia]

@KawzInvests Yes
Case similar to meta with muse

[BBP @BBPronoia]

@KawzInvests Great post

[BBP @BBPronoia]

@CKCapitalxx

[BBP @BBPronoia]

The calculation:

$280 × 1.10⁵ ÷ 30 = $15.03 in annual earnings per share.

$15.03 × 323.331M shares ÷ 20% net margin = $24.3B in annual revenue.

Lower margins, a lower final valuation or additional dilution would raise that hurdle.

[BBP @BBPronoia]

🧨 Too Late to Buy? $BE Bloom Energy edition

i love a stock right up until i read what it costs.

• The price: around $280, up roughly 280% over the past year. The power shortage has shareholders’ attention.

• What it needs: roughly $24.3B in annual revenue by 2031 to deliver a 10% annual share-price return under the assumptions below.

• What it has done: $2.02B in revenue in 2025. Its latest 2026 guidance is $3.9–4.2B.

• The hurdle: about 43% annual sales growth from the 2026 guidance midpoint for five years, reaching a 20% net profit margin.

The business has a real bottleneck to solve. The price needs years of profitable expansion.

Assumptions: $280 starting price; 30× earnings in 2031; 20% net margin; Q2 diluted share count held flat at 323.3M; no dividends. An illustrative scenario, not a forecast.

Data as of September 30, 2026. Sources and calculation in the first reply.

Too late? Which assumption would you change?

[BBP @BBPronoia]

@maxwin Who would want to interact with a crypto girl (if that exists)
Touch grass!