As I have mentioned earlier, I've been putting data (Shiller's CAPE; S&P500 daily prices 1927-2024) on the rack, so to speak. Caveat, the problem with torture is the victim will give the answer his tormentor requires of him.
Anyway, you probably forgotten some years back I pointed out that a Robert Shiller before 1929, using only the data available, would have come to the exact opposite conclusion than he did 60 years later! A high CAPE was better (R^ was weak). Shiller now shows an Excess CAPE yield, he turns CAPE into a yield and subtracts the 10 yr note yield. It's R^ is better than the plain CAPE-Future10YR Real CAGR; .3381 to .258 for all the data. There is that anomaly that I pointed out and I have a hypothesis. There is a pivot around 1914, and what was special about 1914? The US finally has a Central Bank and it's mission isto combat deflation. Stocks usually did well before then when there was inflation and badly when there was deflation. After 1914 inflation will drive down CAPE (and the SP500) as the market believes the Fed will increase rates. A lower CAPE will increase it's yield and offset the rise of the 10yr yield. The excess CAPE yield will also rise.
As I said, using plain CAPE-Future 10yr real returns before 1914 showed a higher CAPE would have a high return, but the R^ was.0973. Using excess CAPE Yield (ECY) it has a .4904 R^. After 1914 the ECY has a R^ of .4842 vs plain CAPE R^ of .3561. Current ECY is about 1.20%; to get into the lowest 10% it must get to .82%. Lower 10% average Real CAGR is .33% and 1Std are +3.55% and -2.89%.
Time to run.
Bumping this for CG 3:
Excess Cape Yield is currently 1.19-1.20%. Average real total returns from 1.20% and below for the next 10 years averaged 1.18% since 1915! 1 STD are 5.14% and -3.52%!
🦔Microsoft canceled its internal Claude Code licenses this week after token-based billing made the cost untenable, even for a company with effectively infinite cloud resources. Uber's CTO sent an internal memo warning the company burned through its entire 2026 AI budget in just… pic.twitter.com/zBGccaP3e6
The AI trade, Nvidia, the whole thing was supposed to collapse by now. That was the forecast about a year ago, and certainly all throughout lass fall.
Not happening. Atleast not yet. Not by a long shot...
And Nvidia mostly flat post earning announcement.
There are plenty of examples of things that are provably false, yet they have their believers. Not sure about AI, however it is provably true that the AI build out is not funded by demand, not close. Crypto mania was phony, so was lab created meat, wet market Covid origin, and trans-women in sports.
I like the comment by one of the commentators of that x post, by Pravnov: "a bomb releases its energy in a microsecond. a data center spreads the same heat over a full day, low-grade, cooled with water and air. comparing the two is built to scare, not inform."
Only way this can happen is if i) there are rolling blackouts across the US (with electricity prices 3-4x higher) or ii) there is a nuke around every corner.
You sure know how to pick them. He's either an idiot or hopes you are. From NVDA's 10K Note 12;
Commitments "Manufacturing, supply, and capacity commitments reflect datacenter-scale production and longer future ordering horizons across current and future product architectures. We enter into agreements with our supply vendors that allow them to procure inventory based upon our defined criteria, and in certain instances, these agreements are cancellable, able to be rescheduled, or adjustable for our business needs prior to placing firm orders. Changes to these agreements may result in additional costs. As of January 25, 2026, these commitments were $95.2 billion, of which substantially all will be paid through fiscal year 2027. Multi-year cloud service agreement commitments as of January 25, 2026, were $27 billion, for which $7 billion, $6 billion, $5 billion, $5 billion, $2 billion, and $2 billion will be paid in fiscal years 2027, 2028, 2029, 2030, 2031, and 2032 and thereafter, respectively. Some cloud service capacity may be reduced, terminated or sold to others by the CSPs, in which case our commitments will be reduced. We expect cloud service agreements to be used to support our research and development efforts. Investment commitments are $11.4 billion as of January 25, 2026, subject to certain contingencies, of which we expect substantially all will be made through fiscal year 2027. Other commitments were $3.4 billion as of January 25, 2026, of which the majority will be paid through fiscal year 2027"
These off balance sheet items will enter the balance sheet as cost of goods sold when delivered and paid for. Current projections for the current fiscal year (ends 1/31/2027) are in the $360+ billion range. Last fiscal year COGS was almost 30% of sales. Sales for Fiscal year 2026 were $215.938 billion. NVDA, and any other company that believes sales will increase Y-o-Y, will increase their commitments from their suppliers accordingly (over 60% for NVDA).
Valor, led by a sitting SpaceX board member, holds >$15B in failed sale-leaseback obligations, including a newly identified $6.58B April 2026 lease. That is extreme related-party leverage concentration against SpaceX. /2
SpaceX S1 is out. I have to get my kid to soccer, will post stuff about this later or tomorrow. https://www.sec.gov/Archives/edgar/data/1181412/000162828026036936/spaceexplorationtechnologi.htm
Anthropic spends $3 to make $1 and that’s before you include any and all other costs like staff or electricity.
Microsoft dumped $300B in capex, made ~$18B in AI revenue. OpenAI and Anthropic alone make up 43-54% of Microsoft, Google, Amazon… pic.twitter.com/QNnV0pP0r0
You sure know how to pick them. He's either an idiot or hopes you are. From NVDA's 10K Note 12;
Commitments "Manufacturing, supply, and capacity commitments reflect datacenter-scale production and longer future ordering horizons across current and future product architectures. We enter into agreements with our supply vendors that allow them to procure inventory based upon our defined criteria, and in certain instances, these agreements are cancellable, able to be rescheduled, or adjustable for our business needs prior to placing firm orders. Changes to these agreements may result in additional costs. As of January 25, 2026, these commitments were $95.2 billion, of which substantially all will be paid through fiscal year 2027. Multi-year cloud service agreement commitments as of January 25, 2026, were $27 billion, for which $7 billion, $6 billion, $5 billion, $5 billion, $2 billion, and $2 billion will be paid in fiscal years 2027, 2028, 2029, 2030, 2031, and 2032 and thereafter, respectively. Some cloud service capacity may be reduced, terminated or sold to others by the CSPs, in which case our commitments will be reduced. We expect cloud service agreements to be used to support our research and development efforts. Investment commitments are $11.4 billion as of January 25, 2026, subject to certain contingencies, of which we expect substantially all will be made through fiscal year 2027. Other commitments were $3.4 billion as of January 25, 2026, of which the majority will be paid through fiscal year 2027"
These off balance sheet items will enter the balance sheet as cost of goods sold when delivered and paid for. Current projections for the current fiscal year (ends 1/31/2027) are in the $360+ billion range. Last fiscal year COGS was almost 30% of sales. Sales for Fiscal year 2026 were $215.938 billion. NVDA, and any other company that believes sales will increase Y-o-Y, will increase their commitments from their suppliers accordingly (over 60% for NVDA).
Big Tech’s AI Accounting Loophole🤯
Microsoft, Google & Amazon “invest” in AI startups like OpenAI & Anthropic… but not with cash. They give them cloud credits.
The startups then spend those credits right back on the big tech clouds.
🚨 THE ENTIRE AI BOOM MIGHT BE BUILT ON FAKE REVENUE.
Latest corporate filings show that OpenAI and Anthropic alone make up over half of the entire $2 trillion future cloud backlog held by Microsoft, Oracle, Google, and Amazon.
