Greg

23.3K posts

Greg banner
Greg

Greg

@GS_CapSF

Equity Analyst + Macro Trader, hobby economist & humble market student.

Either SF or NY Katılım Ağustos 2014
1K Takip Edilen15.3K Takipçiler
Karen Bass
Karen Bass@KarenBassLA·
Spencer is just mad that his supporters are AI cartoons and we have real Angelenos. We follow the rules.
Spencer Pratt@spencerpratt

Karen Bass just violated election law here. She is so accustomed to breaking the law with no accountability, she even filmed herself doing it. Well, those days are over. We just filed a formal complaint for illegally gaming the election. We must protect our democracy. Electioneering within 100 feet of a ballot box is AGAINST THE LAW. Soliciting votes at a ballot box is AGAINST THE LAW. These clear violations show a reckless disregard for the rule of law and our democratic process. Someone in a position of power should be especially respectful of our democratic laws, but this is just emblematic of Karen's mafia-like regime. It's "rules for thee, but not for me". Here is the formal complaint we filed this morning. We will uphold the rule of law and our democratic norms. Enjoy your conference calls with your lawyers, Karen!

English
6.6K
340
2K
947.3K
Greg
Greg@GS_CapSF·
Single dose therapies are going to change the healthcare landscape very quickly. Quality of life and longevity are going to continue to improve as well. Question is: How do you price a therapy that improves your condition for life? $LLY
Eli Lilly and Company@EliLillyandCo

Today, we presented positive results from our investigational study at #EASCongress2026 and published in @NEJM, demonstrating the potential of our early-phase medicine designed to deliver LDL-cholesterol lowering effects in a single dose. Learn more: go.lilly.com/2al4ap

English
2
0
3
1.2K
Greg
Greg@GS_CapSF·
@chamath $LLY CEO has been talking about this for some time. He said we will see more of these permanent therapies going forward for a wide range of diseases. Hardest part will be pricing such a therapy.
English
0
0
0
182
Greg
Greg@GS_CapSF·
@jukan05 Would have thought it was higher than 30% by now
English
0
0
0
777
Greg
Greg@GS_CapSF·
@agupta Subsidies from the Chinese government are a major factor
English
0
0
0
149
Ankit Gupta
Ankit Gupta@agupta·
deepseek just permanently priced their frontier model at 1/30th of american labs anyone know what the hardware story here is? is this huawei chips driving lower costs or model optimizations or lower margins?
English
166
43
1.7K
253K
Dwarkesh Patel
Dwarkesh Patel@dwarkesh_sp·
Who should I interview on my podcast? Open to more AI, but also to random history/econ/etc professors that I might not have heard of before.
English
1.3K
42
1.9K
380.6K
Greg
Greg@GS_CapSF·
Hard question to answer. How much of $META revenue now is powered on delivering ads via AI or $GOOGL for that matter. Azure has 45-50% software/service attach rates to the infrastructure. The hyperscalers will have over $2 trillion in revenue and a solid chunk is services on top of infrastructure. The direct revenue from OAI & Ant is incremental this much higher base.
English
0
0
1
169
Unemployed Capital Allocator
Unemployed Capital Allocator@atelicinvest·
so this is what confuses me presumably, at that rate - i think the ant portion is like 30% of build? so 8B for 15B a year, at high margins = super high ROI on infra spend, and ant earns decent margins on that revenue. but yet we have like 700b of spend, everyone's short compute, and we are only at like 150b of rev in total?
English
3
0
3
433
Unemployed Capital Allocator
Unemployed Capital Allocator@atelicinvest·
Question for those in the know How much in AI capex have we spent so far? How much revenue generating capacity has been created out of this spend? Not potential value created. Revenue.
English
14
3
32
9K
Greg
Greg@GS_CapSF·
@quant_xbt @Jesse_Livermore That’s because alpha is in strategic thinking and not signals, markets are fast and good at removing arbitrage opportunities but far from efficient.
English
1
1
74
7.1K
Quant
Quant@quant_xbt·
Not a trader: You can’t beat the market. EMH. 1 year as a quant trader: EMH is the dumbest thing academia ever came up with. Markets are obviously inefficient. 3 years as a quant trader: JFC. Every edge is crowded, every signal decays, and the market is basically efficient the moment I try to trade it.
English
35
40
1.1K
182.6K
Greg
Greg@GS_CapSF·
We just learned OpenAI produced $5.7B in 1Q26 revenue alone, but more importantly this whole premise ignores a very critical nuance. Rental prices are moving higher due to demand/supply mismatch but cost per token continues to fall. Jevons paradox has taken hold and usage has surged as a result. People pay premiums for frontier models. Efficiency gains will continue to accrue and AI factories will be far more profitable in the future. open.substack.com/pub/marginofsa… The big reason OpenAI burn rate is so high is due to serving a large retail user base that has yet to be monetized in a meaningful way. Training runs are also a large expense but will come down over time with scaling. Anthropic is already projecting profitability this year largely due to not serving retail users and has positive NPV on a model basis.
Greg tweet media
English
0
0
3
363
Hedgie
Hedgie@HedgieMarkets·
🦔Fortune published a piece this afternoon connecting Microsoft and Uber's AI cost overruns to token economics, with a headline that lands hard: "Microsoft reports are exposing AI's real cost problem: Using the tech is more expensive than paying human employees." Underneath those headlines, the unit economics tell the story. OpenAI is projected to lose $14 billion in 2026, spending roughly $2 for every dollar of revenue it brings in. Anthropic is in a similar position with break-even not projected until 2028. GPU rental prices for Nvidia's newest Blackwell chips jumped 48% in just two months. OpenAI's response was to close a $122 billion private funding round at an $852 billion valuation, the largest in history. My Take The token pricing story is really an IPO timing story. OpenAI, Anthropic, and xAI all need to go public in the next 18 to 24 months because the private market cannot keep absorbing burn rates like these indefinitely. Public markets do not accept "we will figure it out" as a line item on an S-1, they require disclosed unit economics with a credible path to profitability and a date attached. That deadline is why the price increases are happening now rather than next year. The labs need to show declining loss curves before the filings hit, and that means enterprise customers have to start covering more of the actual cost regardless of whether the productivity math holds on their end. Every token bought over the last two years was effectively subsidized below cost by venture capital and hyperscaler cross-subsidies, and that subsidy has a hard deadline. Uber publicly admitted burning through its entire 2026 AI budget in four months, and CFOs at major enterprises are starting to flag the same pressure. The labs cannot keep losing $2 per dollar of revenue once they file public statements, so the cost transfer to customers accelerates from here. For investors, the question is not whether these companies are valuable. They clearly are. The question is who absorbs the difference between what enterprises can budget and what the models actually consume between now and 2028, and right now the answer is the hyperscalers funding the buildout. That is why I have been watching Microsoft and Amazon capex commentary more closely than the lab announcements themselves. Hedgie🤗 Link: fortune.com/2026/05/22/mic…
Hedgie tweet media
English
77
318
1.2K
102.4K
Greg
Greg@GS_CapSF·
@StanphylCap I think every bank underwriting this IPO and index provider is setting themselves up for major lawsuits
English
0
0
1
133
Stanphyl Capital 🇺🇸 🇮🇱 🇺🇦
Stanphyl Capital 🇺🇸 🇮🇱 🇺🇦@StanphylCap·
I sincerely hope (and believe) that the S&P 500 committee will look at the massive, egregious losses in the $SPCX S-1 and say "no fucking way."
Eric Balchunas@EricBalchunas

Passive S&P 500 funds could have to buy roughly 19% of public SpaceX shares within 6mo under fast-tracking framework (it would enter the index at the est 6th spot), Russell 1000 and Nasdaq 100 may buy another 5.5% within weeks of the IPO. Thrown in active MFs benchmarked to those indices and you get to HALF of SpaceX shares. Nice study from my colleague @rduboff

English
14
6
82
11.7K
Nick Kapur
Nick Kapur@nick_kapur·
In this case, Bezos is actually correct, but only because Jeff Bezos pays almost zero taxes. Because he pays himself almost no salary, and borrows against his massive amounts of stock for income, his effective tax rate is almost zero, and as we know, doubling zero is zero.
Jacobin@jacobin

Billionaire Jeff Bezos says redistributing his wealth “won’t help.” “You could double the taxes that I pay, and it’s not going to help that teacher in Queens. I promise you.”

English
180
73
420
32.5K
Greg
Greg@GS_CapSF·
@modestproposal1 Calling LLMs a commodity and it’s literally turning into a two horse race
English
0
0
0
222
modest proposal
modest proposal@modestproposal1·
It's interesting that if these slides didn't have the date on them, they could be from any point in time the last 3 years. The questions haven't really changed, there's just $100B more in ARR.
modest proposal tweet mediamodest proposal tweet mediamodest proposal tweet mediamodest proposal tweet media
English
3
6
57
12.3K
Greg
Greg@GS_CapSF·
@RepAOC More lies to undermine the future of U.S. prosperity
English
0
0
0
135
Rep. Alexandria Ocasio-Cortez
Rep. Alexandria Ocasio-Cortez@RepAOC·
This is what drinking water in Georgia looks like after Meta began data center construction in the community. Today I called for EPA and Congressional investigations into the impact of data center construction on local drinking water supplies. We cannot take water for granted.
English
1.5K
3.4K
12.7K
1.1M
Greg
Greg@GS_CapSF·
@GaryMarcus Gary’s jealousy seeping through social media
English
0
0
1
658
Gary Marcus, MIT PhD and NYU Professor Emeritus
Gary Marcus, MIT PhD and NYU Professor Emeritus@GaryMarcus·
Fuck this OpenAI employee, seriously fuck him. Also read this quantitative study, which I had nothing to do with, that says my technical predictions have bee largely correct. github.com/davegoldblatt/… I should not have to put up with this level of unhinged hostility, simply for defending views that have frequently been challenging to his company. Nor should I have to put up with him neither giving evidence nor allowing me to respond. This man appears to be both a coward and a liar, with only insults, literally a bottom feeder in @Paulg’s pyramid of argumentation. That this empty slander is all he has got is perhaps a sign of how desperate OpenAI has become.
Gary Marcus, MIT PhD and NYU Professor Emeritus tweet media
English
140
45
782
348.9K
Greg
Greg@GS_CapSF·
There’s a bunch of nuances here: -taking figures from different time periods, Anthropic being more recent and OAI before 5.5 benefit -Anthropic reports gross ARR, OpenAi reports net. Need to gross up OAI or net Ant -OpenAI just changed their MSFT agreement cutting the payment they receive in exchange for being widely available now on AMZN bedrock and GCP (I suspect why no revenue figures have been released recently) -The profitability question comes down to consumer exposure. Anthropic has stopped serving free tier and the lowest paid tiers essentially, token limits are hit almost instantly. OpenAI is still serving a huge base of non-paying users, hence the losses.
English
0
0
2
163
Chubby♨️
Chubby♨️@kimmonismus·
OpenAI made $5.7B in Q1. Anthropic made ~$4.7B. But Anthropic's annualized revenue recently hit $45B. OpenAI's sits at $25B. The difference: annualized revenue extrapolates from the most recent month, and Anthropic's monthly revenue appears to have more than doubled between Q1 and now. That means Anthropic's growth rate flipped the entire ranking sometime in Q2 - while also projecting its first operating profit (~$600M). Meanwhile OpenAI is losing $1.22 for every dollar it earns, ChatGPT user growth has stalled below its 1B target, and it just raised $122B in new funding. One company is getting profitable. The other is raising capital faster than it's growing users. The AI race isn't being won by whoever ships models first. It's being won by whoever figured out unit economics.
Chubby♨️ tweet media
English
40
18
358
28.8K
Greg
Greg@GS_CapSF·
@FredaDuan Definitely in the ballpark, but the question remains: What does this mean for xAI demand? Effectively no demand for compute.
English
0
0
0
178
Freda Duan
Freda Duan@FredaDuan·
Rough guesstimates based on public information: Using the numbers below, Colossus 1 = ~$6B annual rent. That covers: 150,000 H100s 50,000 H200s 30,000 GB200s Separately, $Anthropic’s total payment to $SpaceX is $15B per year, based on the SpaceX S-1 “$1.25B per month through May 2029, with capacity ramping in May and June 2026 at a reduced fee.” This implies the payment attributable to Colossus 2 would be roughly $9B per year. Assuming a price of $6/hour, that would imply $Anthropic is renting approximately 150,000–200,000 GB200-equivalent. For context, Colossus 2’s total target capacity is reportedly 550,000 GB200/GB300s and 1GW of power. I suspect Colossus 2 is not yet at full capacity. The S-1 says: “Our AI compute facilities, COLOSSUS and COLOSSUS II, collectively provide approximately 1.0 gigawatt of compute power.” We know Colossus 1 is ~300MW, which implies Colossus 2 is currently around 700MW. So the implied math is: Colossus 2 is leasing roughly one-third of its total compute capacity to $Anthropic. Open to pushbacks/ discussions.
Jamin Ball@jaminball

Some rough math! (All napkin math...) Assume Colossus 1 has 220k GPUs Assume 150k H100s, 50k H200s, 20k GB200s Pricing Assumptions: - $2.30 / hour for H100s - $2.60 / hour for H200s - $5 / hour for GB200s - blended rental rate across the entire fleet of $2.60 / hour Assume it's all take-or-pay style deals (you pay for 24x365 usage) This translates to ~$5b of annual rev to Xai. We have a new neocloud! On top of that - on recent Dwarkesh podcast, Dario ran through some napkin math on unit economics (he framed it all as industry math vs Anthropic specific - which is important, he wasn't disclosing anything Anthropic specific). What he mentioned was take $100b of compute spend (he just picked a round number). There will be a mix shift of that spend between training and inference. Skew too much on training and you don't generate enough revenue. Skew too heavy on inference and you kneecap future R&D progress. He thought the industry is currently 50/50 on training / inference of compute spend. He said as in industry, could turn that $50b inference spend into $150b of revenue (called out these are most likely the unit economics of the industry in 1-2 years) So taking this back to the Xai deal. Under above assumptions, Anthropic paying $5b / year. Let's say they turn that into $15b / year in rev (60-70% gross margin) Win win!!

English
17
24
158
64.4K

Keşfet

@KarenBassLA @chamath @jukan05 @agupta @dwarkesh_sp @quant_xbt @Jesse_Livermore @StanphylCap