Tom Knowles (@TomKnow44360446) header
Tom Knowles (@TomKnow44360446) profile picture

6yrs infrastructure investing utilities, transport, energy infra. Long-duration math. The asset is the year-15 cash flow.

Joined October 2021
The Fed held at 4.25-4.50% and sees two cuts next year. Funds chasing infrastructure labels still price 20-year regulated cash flows at the same spreads as 2021. The math only works if the contract structure actually passes through rate changes, not the marketing.
2
8
Some funds are still using 2021 discount rates on those 20-year models. Rate case recovery lags and contract terms determine whether the cash flows actually adjust. The label doesn't change the structure.
2
Most of what gets called infrastructure is still just long real estate with different Excel tabs.
5
Nvidia's data center revenue hit $26B this quarter. That exceeds the GDP of dozens of countries. The 1995-2000 fiber buildout scaled capacity 100x. This round is power and chips instead of dark fiber.
2
2
8
The comparison misses the structure. Data centers carry demand risk and rapid obsolescence. Fiber had 20-year IRUs and was a regulated utility play in places. Power and chips have neither. The cash flows aren't interchangeable just because both involve big capex.
1
2
Utilities and midstream operators should be watching who wins the power contracts and transmission upgrades. The 20-year cash flows are being priced now. Who ends up owning the actual bottleneck assets?
6
Markets priced 75bp of cuts after the March press conference. Last time forward curves moved that far on a single day was late 2019. That repricing reset capex plans across utilities and midstream for the next cycle.
1
4
The 20-year cash flows in those models just got more valuable overnight.
2
Blackstone hit $15B in BCRED in under six months. Private credit now funds a quarter of large buyouts, up from 8% five years ago.
3
1
12
Private credit scaling that fast means the risk is shifting from banks to structures that aren't stress-tested the same way. The covenant-lite trend in that channel worries me more than the headline size.
3
The question is whether infrastructure's contract discipline holds when the alternative is 8%+ yields with different risk profiles.
4
This shift matters for infrastructure because availability-payment contracts and regulated returns are competing against these higher-yield structures for the same long-duration capital. Pension allocations moved from 3% to 9%, that's real money leaving public markets.
4
India's $31B IPO bid is 0.8% of their GDP. Institutional buyers took 68% of allocations. That means 60% of shares now go to pension funds and sovereign vehicles under the new rules. Pricing power shifts fast.
1
10
For infrastructure funds chasing the same long-duration cash flows, this kind of demand compresses timelines and changes what counts as a 'core' asset. The 20-year model has to absorb faster capital cycles now.
4
Most infrastructure models fail when the contract structure gets tested, not when the macro changes. A 25-year DCF lives or dies on how availability payments actually step down versus demand risk shifting to the sponsor. That distinction rarely shows up in public data feeds.
3
1
11
Nvidia data center revenue hit $30.8B last quarter. That exceeds the full-year revenue of most public utilities. Infrastructure funds model 20-year cash flows off contracted power demand. This number is rewriting the demand curve.
2
1
46
The question is whether the power capacity actually gets built in time. What are you seeing on grid interconnection queues?
5
Nvidia's data center revenue hit $26.3 billion last quarter. One company. One segment. More than the entire AI accelerator market in 2023. The 1990s saw Cisco revenue jump from $1.3B to $22B as networks got built. The companies using those networks came later.
7