matthew

The carried interest loophole allows the 1% to pocket billions of dollars per year in avoided taxes, and I'd bet you've never posted about that. But sure, be mad at free lunches for schoolchildren.

For an Austin example: From 2001-onwards, about 90% of permits turned into starts for 5+ units (mostly within a year) Starts in 2023-2025 were only 69% of the units permitted in 2022-2024

There’s a lot of discussion about data center yields as an industry. The customer, contract and full project cost all matter. Digital Realty provides a useful public benchmark. Its Q3 2026 financial supplement reports an estimated stabilized cash yield of 11.6% across its development pipeline, with investment including allocated acquisition, shell and infrastructure costs, plus data center fit-out. That reflects a broader business mix, from smaller colocation requirements to large hyperscale deployments. Smaller deployments generally command higher pricing per unit of capacity and can support higher yields. A hyperscaler leasing an entire building or campus is purchasing wholesale capacity. The size of that commitment and the customer’s purchasing power typically translate into lower pricing per megawatt. The hyperscale businesses of private operators such as QTS, Switch, Vantage and Aligned are relevant comparisons too. I would generally expect comparable projects to have similar economics after accounting for customer credit, contract terms, location, design and scope. That is an expectation, not a claim about their undisclosed yields. At the end of the day, hyperscale data center development is a spread business. Selected 2026 approximately 20-year U.S. dollar bond offerings carried coupons of 5.55% for NVIDIA, 6.00% for Amazon, 6.15% for Microsoft, 6.20% for Meta, 6.25% for Alphabet and 8.65% for Oracle—a range of approximately 5.6%–8.7%. Those offerings were priced at different times. They represent borrowing rates, not current market quotes or total costs of capital. But they help frame the customer’s decision: finance and own the infrastructure, or lease it from a developer. The developer earns a premium for securing power, committing capital, managing construction risk and delivering capacity. When power is scarce and speed matters, that premium can be substantial. Getting capacity years earlier has real economic value. But unusually wide spreads create an incentive for customers to self-build, fund more of the development or bring in competing providers. My view is that low-double-digit yields on full project cost are a reasonable long-term expectation for the industry’s large campuses leased to investment-grade hyperscalers—broadly consistent with Digital Realty’s disclosed estimated stabilized yield of 11.6%, although its pipeline reflects a broader business mix. Short-term yields can differ as supply and demand shift. Scarce power, limited available capacity and urgent delivery needs can support higher returns. I wouldn’t assume those conditions, or the resulting spreads, persist indefinitely. The calculation matters just as much as the percentage. A building-only construction budget and a full project budget that includes allocated campus infrastructure produce different yields. First-year cash yields, average yields over a lease with escalators and levered equity returns also measure different things. A developer’s yield is not directly equivalent to the tenant’s borrowing rate. It must compensate for development risk and leave room for overhead and ongoing capital needs. Compare consistent cash flows against the full investment, then evaluate the customer and contractual commitments supporting those returns. Sources: 2026 company bond announcements and SEC offering documents; Digital Realty Q3 2026 Financial Supplement. Long-term yield expectations reflect my view.

Apartment construction starts are already down 50% from peak, and I'm increasingly of the view that they'll decline further here in Q4 and into 2027. Just a brutal environment for apartment developers these days with high interest rates, frozen equity capital and some construction cost increases -- outweighing improved occupancy/rent trends in most cases. No one is immune to math. If you are relying on Census building permits for multifamily construction trends, you are blissfully unaware of realities on the ground.

Claude / Chat will be $2,000+ per month to get the full capabilities. In college Ubering home from the bar literally cost $2-3 and now that ride is $50-60. The VC money ultimately demands a return.

For those who care. This is the same unit layout, same complex in the Village Apartments, the Hills. Updated. Price $2388. Same in 2026 dollars as when I lived there. Split 6 ways. $400. Maybe there are extra charges. Maybe they won’t let 6 guys in there ( we weren’t supposed to be there either ). I don’t know. But the math is the math. Tbh, I didn’t mind living with 5 roommates. We all were in the same boat. I moved in when I was 24. Lived there or across the way for 2 years before I only needed one roommate If it’s not the lifestyle for you. So be it. If you don’t think just getting any job to get started is the right career choice for you. Do what you think is right. I don’t care and I don’t judge. If you ask me a question , and the Steelers or Cowboys are playing , there is lots of down time to post. So here I am.

I am a lot like AI in that if I read a great deal about a subject I trust my own judgement more than someone who has decades of actual experience in the subject. (The difference being I have judgement to second guess that instinct)

$FBRT DOUBLE DIP REO 👀