Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Telltales

Telltales artwork

Anthropic's 85% is really 42%

Nov 11, 202635 minAudio
ShareXLinkedInFacebookWhatsAppEmail

Mike Nicoletti, Hunt Lawrence, and Jason Wallace walk this week’s Cash Flow Memo: oil averaging $90 through a Hormuz stalemate, whether Jensen Huang survives the spy story, the accounting inside Anthropic’s IPO, and whether you can run Lilly with 100 people. The Cashflow Memo Key Takeaways * Hunt puts post-February oil at about $90 on average and wants the near month near $85 for a stretch to hold that average; 30 ships moved through Hormuz in a recent 24 hours (5 hit by drones) against ~20 million barrels a day of pre-war crude-plus-products flow, the 2028 strip printed $73, and he reads a stalemate (IAEA 24/7 inspections and moving 60% material out of Iran) with WTI able to settle around $80 versus the low $60s before the war. * Of the 18 Bcf/d of extra US gas supply from 2021 to 2026, 13 Bcf/d is Permian associated gas, so high oil is a gas-supply event; Exhibit A then turns to a December spending-authority deadline, an 8% healthcare-cost print, and Hunt’s case that a Sanders-style Medicare for All that keeps private insurance could pass and that the $600 billion federal Medicaid add could fund it if an independent Medicare board, Fed-like, makes the service trade-offs politicians will not. * Jason expects Jensen Huang to still be Nvidia’s CEO: the board’s IR statements describe an investigation of a foreign-espionage attempt against him, not a vote on whether to keep him, and the White House has known since March while keeping him in the inner circle; Mike calls the stock the cheapest since 2016, flags Formula One-style key-man risk from Tae Kim’s The Nvidia Way, and reports Kim saying there is an unnamed internal replacement. * Anthropic’s IPO price is down, Q3 is out, and the deal is oversubscribed; Mike says the 85% gross margin is more like 42% once inference is costed at peak diurnal load rather than as a share of nameplate capacity, maybe worse after the hyperscaler revenue share, and Jason’s S-1 read is 27% inference / 73% training in the first half because they size for the midday peak, with ~6% free-user compute booked as training - treat compute as COGS unless it is explicitly R&D, and 45% is a working gross-margin guess. * Hunt’s Anthropic-founder aside (one-employee agent startups) is the lead-in to last week’s Lilly question: Jason says no, even Anthropic’s one-man biotech still contracted the wet lab, and AI is producing more chemistry work, not fewer heads; BLS productivity ran 3.5-3.6% last quarter against a 2.2% trend, software +12%; then a $3 trillion horse race, SpaceX at $1.8 trillion last Friday versus Anthropic ~$1.3-1.4 trillion pre-money plus a $60-70 billion raise, with Mike picking SpaceX on $50/watt contracts and Hunt saying 8 GW of land compute gets SpaceX close. Show Notes [00:29] Oil Averages Ninety Hunt puts post-February oil at about $90 and wants the near month near $85 to hold that average. Thirty ships moved through Hormuz in a recent day; five were hit. [06:37] Associated Gas Of 18 Bcf/d of extra US gas since 2021, 13 Bcf/d is Permian associated supply. High oil is a gas-supply event. [07:09] Medicare for All Spending authority runs out at the end of December. Healthcare costs are up 8%, and Hunt sees a path for single payer that keeps private insurance if Trump would sign it. [12:43] Jensen Stays Jason reads Nvidia’s board statements as an espionage investigation, not a CEO vote. The stock has not been this cheap since 2016; key-man risk is the other half of the tape. [21:04] Anthropic’s Peak-Load Margin The 85% gross margin is more like 42% once inference is costed at the midday peak. Training is 70% of compute because they size the fleet for that peak. [26:37] Run Lilly with 100? An Anthropic founder told Hunt’s colleague the one-employee startup is coming. Jason says no: the wet lab, the clinic, and the sales call still take people. [31:03] SpaceX vs Anthropic SpaceX at $1.8 trillion, Anthropic around $1.4 trillion pre-money. Hunt wants a running score to $3 trillion on future Wednesdays. Download this week’s Cash Flow Memo at telltales.us. Hunt, Jason, and Mike are back Wednesday. Cashtags $LLY $MSFT $NVDA $TSLA $ANT This post and the information herein are intended for informational purposes only. The views expressed herein are the author’s alone and do not constitute an offer to sell, or a recommendation to purchase, or a solicitation of an offer to buy, any security, nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the author nor any of his employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty, express or implied, is made as to, and no reliance should be placed on, the fairness, accuracy, timeliness or completeness of this information. The author and all employers and their affiliated persons assume no liability for this information and no obligation to update the information or analysis contained herein in the future. This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit telltales.substack.com

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

SPEAKER_03Welcome to Telltales. I'm Mike Nicoletti. Every week, along with my co-host Jason Wallace and Hunt Lawrence, we bring you the cash flow memo. Get it at telltales.us. For the next 30 minutes, we're going to dive deep into energy, technology, and healthcare.

SPEAKER_00So download the memo now and get started. The following conversation is intended for informational purposes only. You should always do your own work to determine if an investment is suitable for you. Let's get started.

SPEAKER_01I was a little worried about my predictions of oil that would get out of hand, but I feel a little better this week because it's come down and it's kind of averaging out, I think, since any of the war activity in February. I think the oil's probably hovering around $90.00. It needs to be where it is, 85 or so for a bit in order to continue to average 90.

SPEAKER_01But that is the current situation. No change or comment on Exhibit C. More oil is moving through Hormoz. And before the bombs and missiles start flying in late February, early March, it's being done by shuttle tankers. The oil now shuttle tankers for products. It's still contested by the Iranians. I was reading one account where 30 ships going

SPEAKER_01through Hormuz, my shuttle, I mean, they pick up cargoes in Saudi facilities or Abu Dhabi or Kuwait or Iraq. They bring the oil out or the products out far enough from the drones and missiles to be loaded at sea into another ship. And then they go back and pick up war. In a recent 24-hour period where 30 ships went through on the Omani side of Hormuz,

SPEAKER_01five of those ships were hit by drones. The U.S. forces that come patrolled that channel, they call it. With helicopters, I think the purpose of the helicopters are to shoot down drums, to see them coming and shoot down drums. So the Iranians are very active, trying to interfere with the capacity to keep the straits open. And the CENTCOM's doing a pretty good job of keeping the oil flowing. 30 shifts may not seem like a lot, but the amount of crude products going through or most before all this started,

SPEAKER_01was roughly estimated around 20 million barrels a day. Well, the crude ships are 2 million barrel ships. 10 ships get you there. The difference between 30 and 10 is not all the ships are as big as Two million barrels. And then there are products tankers. Of course, diesel, especially, and jet fuels very scarce because we've lost the refining capacity in the Persian Gulf. And then we, of course, the Ukrainian drones wrecked out of Russian refineries.

SPEAKER_01So there's been some refining capacity loss there. In terms of trying to predict what's going on for capital markets, oil markets, I think the best prediction is it's just going to continue as is. The Iranians, in order to get the embargo lifted on Iran, probably after agreeing to transport the material enriched to 16%, which is pretty close to being able to be used for bombs, that has to be moved out of the country.

SPEAKER_01The last time this was done, they moved it to their highly enriched material to Russia. They won't agree to do that. They have to agree to a seven-day-a-week, 24-hour-a-day inspection by this UN agency of their centrifuges, the way they enrich uranium. They don't want to do that. And so while there continues to be Albanese and Pakistanis and Intermediaries attempting to negotiate between US and Iran.

SPEAKER_01I think what we can expect is they'll just in various places in Europe or in Gunnar does quite a lot of it. They'll just continue to talk and negotiate. Question is, will the embargo on Iran create enough of dislocation? So that they won't be able to meet the payrolls for the guards units. I think there are about 200 strong in the guards units.

SPEAKER_01And they're the ones that control the country. You know, so far they've been able to. Obviously, there are other ways all fits out over land routes. There's some Iranian oil that shipped through. Pipes into Iraq and then loaded in Iraqi ports. I think the best prediction or the most likely outcome is this deadline just continues. The president said before the elections that he'd have a final campaign to take out

SPEAKER_01with air, take out infrastructure and make it possible for the Iranians not to agree. I don't think That's feasible. I think that's the consensus amongst people, study this kind of thing. Also, the Iranians said, if you do that, we're going to start lobbing Saudi, go there on refining facilities, loading facilities, and whatnot, if there are missiles. So it would just create more trouble, more dislocation, and wouldn't get an agreement done with the Iranians.

SPEAKER_01So I think it's just a stalemate. It's conceivable if we look at the bottom of page 2's price of oil strip for 28 down to 73. I think it's conceivable that despite this ongoing scourishing that is going on and the ships exporting oil out in the shuttle tankers, I think it's conceivable that oil, WTI, settles in at 80. Before all this started, it was low 60s.

SPEAKER_01Is the difference between 80 and 60 to 20 miles per hour going to make a big difference to our economy or economies of China or other developed economies. I don't think so. That's just where we are. Natural gas is a casualty of all this because it's covered in prior Wednesdays. The higher the oil price is, the more associated gas you get and the more gas supply.

SPEAKER_01And as you can see from the table, the 18B spent a day of additional gas supply from 21 to 26. 13 of it came out of the Permian-associated gas. I owe prices and I result for gas supply. Exhibit A, the Democrats now will have to pass a budget. I think spending authority for the government runs out at the end of December. So there's going to have to be some bipartisan collaboration

SPEAKER_01of passing an extension of spending authority. The Trump administration wants more defense spending. I suppose they have to cover the cost of the embargo on Iran and trying to protect ships coming through the streets foremost. I doubt if they'll get that from the Democrats. So there'll be a bit of a push to continue the defense effort. I think the priorities for the Democrats is probably going to be out there.

SPEAKER_01They're going to be tempted to just try to have an increase in subsidies for Affordable Care Act or Obamacare. I think that doesn't get too much good done. It increases cash flow that Insurance companies, hospitals. The Bernie Sanders point of the party is going to push like crazy to do single-payer health care or Medicare for all. I think their version of Medicare for all would allow employers to continue

SPEAKER_01to provide insurance, would allow private insurance. I think there's a chance that that type of legislation might have a chance to be passed. Key question is, would Trump sign it? I think he might. There'll be a lobbying effort. It surpasses any lobbying effort we've seen against that by insurance companies and hospitals and all kinds of other health care providers. But I think I wouldn't predict it, but I think it's a possibility.

SPEAKER_01What else if your issue is affordability? I think we're kind of a long-term stalemate. In Iran, I don't think the Democrats are going to be in favor of, oh, let's bring our carriers and our destroyers back and let Iran dominate the straight and foremost. I don't think there's any support amongst the Democrats to do that. So energy prices are just kind of stuck with some improvement. That's nothing like going back to where it was before Iran started.

SPEAKER_01Food pricing, maybe there'd be a little help there. The main thing they should focus on is healthcare. Mike, Jason, any commentary on that?

SPEAKER_02I think the passion for healthcare cost increases this coming year is 8%.

SPEAKER_00So much higher than inflation. We can't do nothing.

SPEAKER_02We have to try something. The system as is will bankrupt the country. So maybe Medicare for all is

SPEAKER_03an attempt we make. Yep, and all those people that are really pushing for it are going to be disappointed when Medicare for All ends up being like every other government-run healthcare system, which really just means reduced amount of services. But it's still probably the right thing to do

SPEAKER_01so that there is a safety net. I keep thinking that Medicare for All will be more expensive for the federal government, but looking at Exhibit A, we... In government spends $600 billion added to state-run Medicaid programs. If we have, as Mike says, I was thinking after all, why in the world do we need those Medicaid programs? So there are the costs that the Medicare system would go up,

SPEAKER_01but maybe most of that $600 billion can be put towards financing the increased Medicare costs. The thing I think you need is, and Mike and Jason are really informed on this, you'd want to have independence for the Medicare Board, and you'd want to make the Medicare Board look more like the reserve thing for the measure of independence. I'm not quite sure how you do that, but that would have to be part of this,

SPEAKER_01because you wouldn't want to have legislators, senators, and congressmen Managing the Medicare for Old Budget, I think it'd be much better to have a professional push to it with an independent board. As I say, something that approximately the Federal Reserve Board has.

SPEAKER_02Yeah, the woman that solved this in Montana is the top choice.

SPEAKER_01Yes.

SPEAKER_03Whatever you set up, this board would have to be responsible for making trade-offs, the tough trade-offs that no politician can or is willing to make. And that is how much money do we spend on someone's health, which means you have to value what's the value of a life. And all that will be, hopefully, give it a budget so they can allocate dollars effectively Through the system

SPEAKER_03and they're going to be tough trade -offs people will not necessarily like it, but you need that ability to

SPEAKER_01make trade -offs, and The founder of NVIDIA, whether or not he'll survive his relationship with this Chinese lady who was supposedly a spy for the Chinese. Mike and I talk 20 minutes every morning, so we covered this in some detail. Jason, how does it look to you?

SPEAKER_01Do you expect Jason to be the CEO of NVIDIA with 60 people reporting to him at the end of the year? Or how does it look to you?

SPEAKER_00Where to start? Short answer is I expect them to still be there.

SPEAKER_02I suggest reading the NVIDIA board's statements that they've provided on NVIDIA's IR site.

SPEAKER_01They kind of contradict how the media is reporting it.

SPEAKER_03The media's suggestion is the board's investigating Jensen and his wrongdoings of supporting a spy and deciding

SPEAKER_02whether he should remain a CEO or not. If you read the board's statements, it's clearly that they're

SPEAKER_00investigating a foreign espionage attempt against their CEO. And then the statement that came out yesterday was that they hired a law firm and some investigators to

SPEAKER_02fulfill the statement that they made on Sunday. So I don't think the NVIDIA board is even looking at removing Jensen. I think they're examining what was the consequence of having a spy target their CEO

SPEAKER_03and what information was exploited or what decisions did she have influence over. Infidelity is unfortunately pretty common. I don't think he's removed just for that. And then otherwise, He is really a victim of a state-sponsored espionage.

SPEAKER_02And Mike and I have a question of when did Jensen know or when did she become a spy?

SPEAKER_03So the CCP has a reach to all of their citizens and they might have known about the relationship

SPEAKER_02after it started and it turned her then. So it could be like a black male situation against her as well. What I think is interesting is it appears that the U.S. government has known about this, that she was potentially a spy for over a year. And certainly the reporting of the timelines I've seen is that the White House has been aware of it since March.

SPEAKER_01And Jensen's still been in the inner circle of the Trump administration after March.

SPEAKER_02So they didn't kick him out. They didn't distance themselves from him. And you have to ask why. So if they knew this was happening, they're either not informing Jensen and they're going to use the opportunity To watch him very closely

SPEAKER_03I think there was no danger to the US, and I kind of think there's no danger to his job, really. NVIDIA's on sale. Yeah, it hasn't been this cheap since 2016. It's a great news cycle. It's been... Well, it should be journal on the front page every day.

SPEAKER_02Yeah, and I think it'll be very uncomfortable for Jensen for a while. He's going to be subpoenaed and called in front of Congress. This morning, I saw people saying... What's the potential for jail time for him?

SPEAKER_03I think he's going to have a rough go, but in the end, I think he's really a victim as well. There is a small probability there's a big problem. Yeah. But I think that the big picture is the news and the media are grabbing onto that very remote, small probability and trying to make that story and simply not that logical. Yeah. Yeah. I think the easier case to make is this is the CEO of the most important company in the world.

SPEAKER_02Why would the CCP not try to either spy on him, influence him, whatever the situation? How would you replace him if something else happened to him?

SPEAKER_01He does seem maybe comparable to us, but he does seem irreplaceable. Part of NVIDIA's activity. Would you have to go outside, do you think, to get a prominent person? The person I'm thinking about is the CEO of Microsoft, other than the theory that Microsoft may have some people that are on Microsoft, and they might have enough stature to stand in for Janssen

SPEAKER_01because It's not only NVIDIA and its stockholders and employees. It's just so much dependence on NVIDIA through the tech industry that he's not being there. The only other person with a similar kind of

SPEAKER_03Significance, I think, is Musk. And I suppose you could argue, where would we?

SPEAKER_01Musk has had a CEO at SpaceX for this whole period of time, who I guess could certainly run, should be able to run the launch business. Maybe somebody can run Starlink, or maybe somebody can run Starlink. So Musk is a similarly active and Competent, credible leadership capability does that backup, just at least based on a read of the proxy statement and people's assessment

SPEAKER_01of who works there is unusual in that you have to go outside for someone very prominent and competent to replace them, I think.

SPEAKER_03Mike Jason had a set

SPEAKER_01-to reverse assessment of the implications of losing up close if he was in a car crash or something, what do you do

SPEAKER_03So the analogy that was described in the book, The NVIDIA Way by Taikim is that NVIDA is like a Formula One car and it's all custom built around Jensen -Huang. And there have been a few podcasts this week that have talked about that and basically to reiterate the point that, oh look he's irreplaceable. This key man risk is way too high. That said, I was in a chat group with Ty Kim.

SPEAKER_03This week, and this is the exact topic that was brought up, and his response

SPEAKER_01was that there is a replacement for him at NVIDIA, but he didn't want to name the person because it would cause too much commotion. He's probably the most outside person that knows NVIDIA very well, so that's just one observation.

SPEAKER_03I don't have a I think it'd be hard to put an outsider in to run it, is the conclusion that I've come to.

SPEAKER_01But I also don't think, I think Colette does a great job as CFO, but I don't

SPEAKER_03think you're going to put her in that role either. I think it has to be a technical person.

SPEAKER_01Anything else on page one, two, three, and four that has been overshadowed by all the news about Jensen?

SPEAKER_03Talk about Anthropic?

SPEAKER_01The Anthropic IPO. Yeah, we have to leave a little time for healthcare. We probably should have switched into the Anthropic ICO five minutes ago. Let's do Anthropic.

SPEAKER_03Well, we got the Q3 numbers and they've lowered the IPO price. I think we've got some news that it's already over-scribed.

SPEAKER_01The thing that Jason and I have been discussing the last few days is how their inference costs are so low.

SPEAKER_03And we figured a bit out, and it has to do with the way that they account for their inference costs. When you stand up a data center, it doesn't just run 100% of the time generating money. Your profitable time is when tokens are processed. Tokens get processed throughout the day and follow

SPEAKER_01the diurnal pattern within each time zone during the working day is the primary time of use.

SPEAKER_03There's obviously spikes, different patterns among different usage groups, but broadly you can think of it as a bell-shaped curve that peaks in the middle of the day. The way that Anthropik's accounting for that cost is they're taking the total cost of the data center. And the percentage of the capacity that's used by inference and applying it that way. When in fact, the value of that compute capacity

SPEAKER_03during that diurnal period is many times greater than that of the dead zone. They can utilize the dead zone nights and weekends doing some reinforcement learning and Post-training, all that kind of stuff gets used by their R&D staff as best they can. But it's fairly well documented that these companies have been trying to figure out how to

SPEAKER_01shift workloads for their customers off of that diagonal pattern.

SPEAKER_03So when we think about and look at their data center capacity planning, a lot of that has to do with the fact that they need to Budget capacity for the peak of the diurnal pattern. I think there's a lot of room for optimization there. Scale adds a lot to one's ability to manage that. So if you have your own data centers operating in different places around the globe in different time zones, you're able to move and shift workloads around perfectly.

SPEAKER_03Which maybe partly explains what's happening, spending so much money on compute. But I think the big point here is to show that the gross profit is not as good as it looks. Instead of 85%, it's more like 42%. And maybe a little worse, but you account for the revenue share from the hyperscalers.

SPEAKER_02Yeah. One of the questions that prompted this and really eating at us was why the training budget 70% of the compute

SPEAKER_03and inference only 30%. And it really is as Mike was saying. Budgeting for the peak load of inference And the volume above the bell curve is greater than under the curve. So the first half of the year, Anthropics S1 said it was

SPEAKER_0227% inference, compute dedicated to inference, 73 to training. And they call the compute a fungible pool of compute. So it's really the same machines doing inference and training. One thing we think is happening, is about 6 % of their Compute load goes to free users of the product. And that as well is documented

SPEAKER_02as the training and internal research of other. So really that is inference cost that's not being included in

SPEAKER_03the gross margin of providing the product. The more honest... Counting treatment would be to treat the compute cost as cost of goods sold unless it's explicitly used for RDD.

SPEAKER_02Yeah.

SPEAKER_03And we don't have a way to back that out. We could approximate it, but I think if you assume 45% first margin, that's probably not out ask questions, and

SPEAKER_02you can have a window into what they're planning to roll out based on their

SPEAKER_03Contracted spending on new data center. Very clearly they see the peak usage of tokens spiking in the coming years. And we have a couple ideas of how that happens. Talked ourselves back into the situation that, yes, at

SPEAKER_02peak load, all this compute does get used. Right now, today, we're only doing text generation and we're looking at problems that are solved by basically writing reports.

SPEAKER_03But a lot of the stuff they're working on and what's coming... We saw it teased by OpenAI have their Sora video product.

SPEAKER_02There's a lot like graphic design type Tools being made. That's the future. That's going to get bigger as well. And that's much more token intensive.

SPEAKER_01Right. We've got to do some help here. That gives me a good lead-in to where we were last Wednesday. Could you run lowly with a hundred people? A hundred technical people. It's our most valuable healthcare company. And just as a lead-in, Jason, I assume the answer is no, but just as a lead-in, Holly Durbin was in a meeting with one of the founders of Adropic. It was related to the roadshow.

SPEAKER_01It was prior scheduled. It was last week. And my friend sat next to him, and he said that he thought That there would be a great number of startups. He said, it's all about scale. A great number of startups that literally would have only one employee. There'd be an idea of the person with agents would be able to do everything. If you take that as a prediction, one of the anthropic seven founders that are going to share 50% voting,

SPEAKER_01why couldn't you run globally with 100 people rather than thousands?

SPEAKER_02Lily has a lot going on. So this Anthropic did purchase a one man company

SPEAKER_03in the biotech space and he was using AI to help him design molecules that he then contracted other researchers to do the wet work for.

SPEAKER_02They're doing the chemistry, they're doing some testing. So even though it was a one man company, there were

SPEAKER_03Many more people involved

SPEAKER_02doing the physical science. Until robots can do that. That's the job of people today. Lilly is a huge organization. They have many drug discovery programs going on, many clinical research programs.

SPEAKER_03I think the sales cycle is definitely human intensive.

SPEAKER_02You look at the reps going to clinicians' offices and bringing donuts. That's how they're making sales. So I don't think, I think everyone involved becomes more efficient and maybe you compress timelines of how things are documented and productivity improves. But I don't see them really cutting headcount any significance based on this, they might accelerate.

SPEAKER_01Mike, do you agree?

SPEAKER_03Oh, my experience is you've had AI to a human and just do more work. It's a personal experience.

SPEAKER_02It's a personal experience. The researchers I speak to say AI is helping them come up with more ideas faster, but like Mike said, that leads to more work. More people are doing the chemistry then to test these ideas. Yeah.

SPEAKER_03They were very much second industrial revolution level of growth coming. We had some interesting data out of

SPEAKER_00BLS last week around productivity numbers.

SPEAKER_02The historically productivity advances at an annualized rate of 2 .2%. This last quarter, it was I believe 3 .5 or 3 6%. And then they break it down by industry. A couple of them were surprising, but it's no surprise that the top one was software industry where productivity increased 12%. So yeah, they're the first industry to really fully adopt AI tools and we've already seen in

SPEAKER_00it where AI

SPEAKER_03kind of made the job faster.

SPEAKER_02Not kind of, it greatly improved the speed of the job. There were some layoffs initially and now there's more hiring than layoffs today. More engineers are required and productivity is rising double digits.

SPEAKER_03Right. Law is coming behind that closely. Accounting. There's a lot of fields that

SPEAKER_01will benefit greatly.

SPEAKER_02Yeah.

SPEAKER_01Yeah. Well, it's absolutely clear that Jason and Mike are not tumors. Definitely not. They're the ones who are predicting that jobs will be gone and the world will land and all that kind of stuff. How... If we set up a contest between SpaceX and Antropic in terms of who's going to reach some number of equity value, trillions of dollars with equity,

SPEAKER_01say three trillion dollars, and assume that SpaceX doesn't get there by acquiring Tesla, who gets there first, SpaceX or Antropic?

SPEAKER_03I think SpaceX. The new contracts, assume they're going to sign it at $50 a watt, is going to increase the value of the company. And Anthropic will need to convert, at some point, they're going to have to convert the capacity into revenue and the revenue into free cash flow.

SPEAKER_02The formula is much more simple at SpaceX.

SPEAKER_03Right.

SPEAKER_01Well, just for the record, Space Jackson is a trillion eight as

SPEAKER_03of last Friday.

SPEAKER_01And it looks like the free money ICO value for Antropocids, one four, one three, one four, something of that range. The money being raised, that's $60 or $70 billion, will be added to that. So this will be something we'll probably keep track of. It'll be when the topic's finally trading, it'll be kind of like a trillion eight for SpaceX and a trillion four for our topic.

SPEAKER_01So it'll be a horse race.

SPEAKER_03Yeah.

SPEAKER_01Three trillion from a trillion fours doubled. And it's fair to say that with all the count requirements that their proper chance have their own compute, doing your double layers come the order where SpaceX, combination of Monday Night Rays, cash flow from launch and Starlink, they did succeed in adding their 18 watts of land-based compute. I think that would get those three trillion or get them closer anyway.

SPEAKER_01Something to focus on on future Wednesdays. Anyway, we're way over. Everyone be well and stay healthy. We'll be here in exactly seven days from now.

SPEAKER_00And thank you very much for listening. Take care. The views expressed on this podcast are the hosts alone and do not constitute an offer to sell or a recommendation to purchase, or a solicitation of an offer to buy any security nor a recommendation for any investment product or service. While certain information contained herein has been obtained from sources believed to be reliable, neither the hosts nor any of their employers or their affiliates have independently verified this information, and its accuracy and completeness cannot be guaranteed. Accordingly, no representation or warranty expressed or implied is made as -to

SPEAKER_00and no reliance should be placed on The fairness accuracy timeliness or completeness of this information. [music] Thank you for joining us this week.

SPEAKER_00Please tune in again next week as we will be back on Wednesday. As a reminder nothing on this podcast should be considered investment advice. You should always do your own work to determine if an investment is suitable for you.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Anthropic Lists, Jensen Waits

    Nov 18, 202634 minAudio

    Mike Nicoletti, Hunt Lawrence, and Jason Wallace walk this week’s Cash Flow Memo: an $80 oil stalemate, oil producers at 10x free cash flow, Anthropic’s first day of trading, the fight over Jensen Huang’s return, and a...

    Credit the people on this episode

  2. The Day After the Blue Wave: Oil, Deficits, Chips and Drugs

    Nov 4, 202637 minAudio

    Recorded the day after the election, Mike Nicoletti, Hunt Lawrence, and Jason Wallace walk this week’s Cash Flow Memo: what a blue wave means for oil, the deficit, data centers and healthcare, plus the economics inside...

    Credit the people on this episode

  3. Weekend Update - W2644

    Nov 1, 20268 minAudio

    ▶ Explore this week’s Tape - live, sortable, drill-down → The Tape - W2644 Universe of 94 cashflow-memo names, snap dates 2026-10-23 → 2026-10-30. Composite is rank-sum percentile of FCF Yield + NTM Revenue Growth...

    Credit the people on this episode

  4. Anthropic Burned Under $10B. Now It's Betting Hundreds of Billions.

    Oct 28, 202635 minAudio

    Mike Nicoletti, Hunt Lawrence, and Jason Wallace walk this week’s Cash Flow Memo, from $90 oil and a grinding 10-year to the long-awaited Anthropic S-1, with Montana joining on how pharma will use AI. The Cashflow Memo...

    Credit the people on this episode

  5. Weekend Update - W2643

    Oct 25, 20269 minAudio

    ▶ Explore this week’s Tape - live, sortable, drill-down → The Tape - W2643 Universe of 94 cashflow-memo names, snap dates 2026-10-17 → 2026-10-23. Composite is rank-sum percentile of FCF Yield + NTM Revenue Growth...

    Credit the people on this episode