Mango Contango

Usually NG

[John Coogan @johncoogan]

Very few people pricing in a scenario where a rogue AI agent hacks into the FDA and unbans the original Four Loko

[Mango Contango @MangoContango_]

@OCM_C2 @USAEnergyNatGas Tighttttt

[Josh Billinson @jbillinson]

Miami is that bad, huh

[Mango Contango @MangoContango_]

Glad to see Neel is getting attention. One of the only guys on the internet who teaches commod trading the right way- from a fundamental perspective- and does it well

[Neel Somani @neelsomani]

I'm a former Citadel quant who covered power & gas.

There's constant talk about chips & memory, but power is the central bottleneck for AI.

Very few people understand it, so I'm posting a canonical primer on power pricing & data centers: https://t.co/LO5ovj2imA

[Neel Somani @neelsomani]

I'm a former Citadel quant who covered power & gas.

There's constant talk about chips & memory, but power is the central bottleneck for AI.

Very few people understand it, so I'm posting a canonical primer on power pricing & data centers: https://t.co/LO5ovj2imA

[Yellowbrick Investing @joinyellowbrick]

Some very valid pushback on building models like this:
1) you are compounding 100s of inaccurate inputs (it's impossible to properly forecast all of these numbers)
2) the majority of the inputs are completely irrelevant to the outcome
3) the model becomes your identity. Nobody has ever spent 18-months modeling something and then said "yeah, there is nothing here"
4) it's performative. "I've modeled every well, pipeline, etc" is unfalsifiable and (likely) useless. The only reason to add data point number 1,412 is to tell people that you added it

[Mango Contango @MangoContango_]

I’m confused at the attention this has been getting today. The 20 pager reads as if it’s a first draft to an undergrad research paper on an interesting topic. Not saying he’s not posing interesting questions that should be discussed, but why this podcast/paper in particular?

[Patrick OShaughnessy @patrick_oshag]

Matthew Smith has spent the last 18 months modeling every well, pipeline, storage facility, and power plant in the American natural gas system.

His conclusion is that the US is heading toward a natural gas shortage with no precedent, beginning in 2028.

By 2030, he believes we could exhaust our working natural gas storage entirely.

The fuel everyone in AI is counting on, and that everyone assumes is abundant, is not there.

And because gas sets the price of electricity in most of the country, he argues Americans will pay for the shortage in their power bills.

Matthew has worked in energy markets for over 20 years and is the CIO of Chronometer Partners.

This is his second time on the show, and he's one of my favorite people to talk to about energy.

We discuss:
- Why the bottleneck is moving from power to fuel
- Why we can't just shut off exports
- 2028 as the inflection point
- Large-scale nuclear v. SMRs
- Who wins, who loses, and what can still be done

Enjoy!

TIMESTAMPS
0:00 Intro
1:30 What Drives the Deficit
11:00 Why Supply Can’t Catch Up
20:35 The 2030 Gas Crisis
25:05 Winners and Losers
29:00 Nuclear and Solar
33:30 Consumers Pay the Bill
37:20 AI’s Next Shortage
45:25 Solutions and Global Stakes
51:15 The Coming Gas Knife Fight

[Michael Spyker @ShaleTier7]

The price at which North America can produce 200 Bcf/d is like $5 USD/MMBtu. It's not $10 USD/MMBtu, and there will never be a shortage.

Matthew Smith, DM me and we can talk, your conclusion is very, very wrong.

[NextWave EFT @NextWaveEFT]

I’ll give my brief take on this interview Smith gave of nat gas long term. In short, I respectfully disagree.

Smith’s thesis relies heavily on a single monster assumption which would take the story down if not true: If primary basins today can’t fill demand, nat gas prices 🚀

If we were reliant on primary basins to fill demand, that would be true. But that’s not the case.

The last 15 years our industry has been lulled into thinking the only basins with meaningful resource is Appalachia, the Haynesville, and the Permian. This is not true. Those are just the basins that work at $3/mcf.

When you make small increases in nat gas prices (ie, $4-5/mcf) there are vast reserves across the L48 that would be activated. None of them on their own are large enough to move the needle, but when you’ve got 50 small programs each capable of growing to 0.5 bcfd, thats more meaningful than finding a new monster gas basin.

None of those small programs are being modeled today. Nobody has “300 mmcfd coming from a no name private operator in the piceance” or “400 mmcfd from a new operator in Mississippi” in their model. But when you look at the economics of those programs at $4/mcf, they make a hell of a lot of sense to drill. I’ve seen these deals and projects. They’re popping up everywhere at an accelerated pace. With increased confidence in a $4 world, they will get funded, and those volumes will show up.

Over the last 20 years the nat gas business has become hyper concentrated in the hands of a few key players in a few key basins. The next 20 years what I believe you’ll see in a $4-5 world is a much more distributed network of nat gas producers who make a killing at those prices.

We had 100 nat gas operators, we now have 5. What I believe Matt and many others miss is in 10 years, in a $4-5/mcf world, it’s likely we once again have 100 nat gas operators across the L48.

You model that and the thesis of nat gas going to the moon completely crumbles. That flawed thesis is underpinned by a complete disregard of what new nat gas teams are capable of delivering if they were given even a moderate opportunity ($4/mcf).

I’m bullish relative to strip today, but the idea of runaway nat gas prices over the long term is highly improbable. There’s simply too much talent, resource, and $ sitting on the sidelines that would storm the field at $4-5/mcf.

[Ben Smith @firstenercast]

i read the 20pg natgas bull thesis presented by matt smith and @colossusmag

like their angle on deliverability/infrastucture limits (not resource size) as the binding constraint .....refreshingly novel approach

but here are 9 cracks in their logic (in no particular order) 🧵

[Mango Contango @MangoContango_]

This is how I often feel when grinding away at fundies. If I put that same effort to something else, I think my peak could be the same/higher, but like many I struggle to figure out what I should do in life. For that reason, I hope my time in this industry be on the shorter side

[Lauris @lzminsky]

An experience that complely decimated the high school striver prestige maxxing mindset for me was living 2 months at a well-known Palm Beach country club (courtesy of an ex) when I was 20.

Quickly realized that the people who had half a yard in NW literally produced either a single shingle for some production supply chain, owned some estoteric patent, or did some very niche import-export thing (outside of lawyers who made money on the 60s-80s Floida real estate boom).

Whenever someone is stacking their credentials infront of me in our first conversation (which is a really weird faux pas anyhow), the image of Roger from Pittsburgh, who went to community college, and owns 30 boats, pops into my mind to remind me of what actually matters for value capture.

I try to imagine what he would feel about someone if he heard then trying to establish frame through signal stacking in a conversation.

Great guy. Hope he’s still alive and kicking.

[Marco Foster @MarcoFoster_]

Tim Dillon: “We can’t handle people chanting death to America? I take more shit than that in comments. This is crazy. Can you imagine your kid dies in Iran and you’re telling someone about it: yeah, it was a tough loss, but they were chanting death to America. We had to show them you can’t do that. I know my kid’s dead, but you know, no more chanting”

[Mahmoouud @Mahhmooud]

البيزو الارجنتيني vs الجنيه المصري
اللي هيكسب هياخد قرض من صندوق النقد

[Peter Gostev (SF 24-28 August) @petergostev]

We are now in a position where a tiny proportion of the population uses Fable or soon GPT-5.6, while everyone else's experience of AI is 8-30b-model level - Google's AI Overviews, Meta AI, ChatGPT free tier, maybe MS Copilot at best. People outside of tech must be completely baffled how this is supposed to take their job, and annoyed that hundreds of billions are being poured into it.

[Milo Smith @mil000]

im gonna pass on installing your OS bro

[Paarug Sethi @paarugsethi]

all timer

[Sami Gold @souljagoyteller]

One of my favorite current political bits is how Brandon Gill is positioning himself as the bulwark against the Great Replacement and radical Islam only because he’s married to an Indian woman (Dinesh D’Souza’s daughter). Once you know that, every statement of his becomes funny

[Brett Harrison @BrettHarrison]

In light of today’s SpaceX IPO, a story from inside Jane Street on the day of the 2012 Facebook IPO:

[alex bronzini-vender @alexbronzini]

"You've never applied to the University of Austin (UATX). We're admitting you anyway."

Surely a sign of a thriving institution