원문 보기
Anthropic updated its S1 with third quarter results in a massive computing bill. Let's dig into the details with Evan Schlloman. He's a principal at Neostellar Capital. Evan, thank you for joining us today. Thanks for being with us. I just want to start with a few of those numbers right off the top here.
Quarterly revenue growing from almost 12 billion in Q2 to nearly 17 billion in Q3. However, companies disclosing nearly $900 billion in future purchase and lease commitments. If you're a potential investor looking at this company, how do you square those two figures? Yeah. So, one of the interesting things you're looking at and look, if you look at the topline numbers there, as you said, right, was 45% growth going from Q2 to Q3.
So, you're talking about really, you know, a extremely unique story in terms of the growth at a a pretty high dollar figure in terms of absolutes. So, let's just keep that in mind for the context there. I think when you look at the spending side, one interesting thing is you see the short-term compute which they disclosed there, right?
Going from about 1.5 billion to I think a little over 5 billion um when you look at sort of the they were disclosing in Q2 and then the Q3 update. And so what that's telling us really is that if you look at the pace of growth and what Anthropic may have been anticipating, right, in terms of long-term compute that they were purchasing, that obviously comes at a cheaper price.
And so you would ideally like to have more of your compute bill come from long-term compute. The fact that the company is tripling short-term compute contracts to that over $5 billion value really tells us that the growth rate there is coming even faster than Anthropic may have anticipated when they were originally coming into those contracts and it's showing just the absolute sheer demand that they are needing to pull the trigger in order to satiate right and so that's one thing when we square sort of the compute value and demand is this company is now growing at the clip that you described before right and they are needing to anticipate and sort of built for this and the question is how do you provide that compute satiate that demand in the most cost effective way and that is building up long-term facilities building up long-term compute capacity so that you can you know properly be able to provide the supply that meets the demand that Anthropic is seeing on the other end of that equation
Evan one of the big questions here with Enthropic also with open AI but really with giving the upcoming IPO has been this idea of when the customers of what these labs are producing are going to start seeing some of that ROI, that return on investment. When are we going to start seeing the actual uh you know business production case of all the spending that's going on here?
When you look at Enthropic, the numbers coming out of this company, how do you think through that question and what it could mean for Enthropic going forward? I think you're seeing it right. And so if you look at the composition they disclosed in terms of the customers, you're seeing an increase in sort of diversification.
You're seeing an increase in customers at the you know call it 100k mark at the million in terms of annual spending commitment mark and the reason you are seeing customers spending more and more with these platforms you have you know a what 900% in RR figure if you go back from sort of Q3 to I think a year ago is because customers are presumably getting value out of the products right and so when you think about models that are getting more and more performant you're having a use case for AI that's expand expanding.
So your base is expanding there. I think the reason we are seeing that these customers, one, you're growing your base of customers. You're seeing more and more customers on the platform. But two across the board in terms of AI, you're seeing more and more usage. And the reason you're seeing that usage is because people are determining that the value they're getting out of it is, you know, at least within merit if not sort of value creative for them to use the product.
And so that's I think what's generating some of the expanded use cases here. um when you look at sort of actual usage you're seeing a huge amount of usage coming from recent models right so about 80% I think of current quarter revenue or Q3 revenue was being generated from models that were created within the last 6 months the reason is because these models are more and more performant and so you would expect to see with more and more performant models a broader use case of where sort of the ROI intersects you're able to justify the cost and use of the model um and so I think we're already entering that curve I think there's probably still early innings though in terms of how far we're able to go when we see more and more performance coming out of these models where companies are going to be able to generate value.
Evan, looking at the IPO, the upcoming IPO for Anthropic, we've seen some ripples around the potential valuation for this company. Anthropic, of course, looking for $2 trillion. A story coming out of the Wall Street Journal, you quite literally just minutes ago reporting that institutional investors have been telling Anthropic that the company should be targeting closer to a$ 1.5 trillion$ 1.6 trillion valuation.
Those investors citing a few things. I'm quoting from the journal here, including choppy markets and growing concerns about overexposure to AI bets, political backlash to data centers, uncertainty around how soon hyperscalers will show AI related profits. Looking at that IPO, where is your head at on the valuation for this company and what Anthropic should be aiming for and then what they might actually end up getting?
Yeah, I think you're seeing a few different trends here, right? and they're all sort of intersecting and that's that's part of what you have to locate is one you're seeing if you look at the base story for what's driving AI and what's driving anthropic or these companies you are seeing companies that are quite literally driving the frontier right they are you are talking about technology that is shaping almost every part of a technology stack across the economy right now and these are companies that are driving sort of the bleeding edge of that and so you have that on one end you also have a fundamental case of increasing demand an increasing demand from repeat customers customers as well as new customers coming in the platform.
But that right of course is buttressed up against a for what we see in the public markets extreme you know a volatile market timing. Um right and so if you look at just macro risks even risks that exist within the market uh if you think about sort of any sort of short term of what you're looking at those have to be taken into consideration.
I would also say that if you think about the IPOs, right, Anthropic is seen here as, you know, one company that may be able to open up sort of a broader IPO landscape at Aperture. And so, you would think that this IPO is likely going to be positioned to give it the best chance possible to perform well.
Um, and obviously, if you're putting in money at the current IPO pricing, you have to be underwriting a higher price in order to make money on it, right? no one's, you know, putting in at X value because they think they're going to sell it a year from now at that same exact value. And so you have the confluence of all these events um that are driving sort of uh you know supply demand equilibrium and where they'll find that price.
Evan, that'll certainly be something to watch when we do get this IPO. We thank you for your time. Much appreciated for your insights. Absolutely. Thank you.
Comments 0