All right, so let's dive into our news stories and our first one is going to be about Q2 being not so great at the macro level. I think we're still concerned about the Ukraine war, Iran war, the price of oil. Corporate earnings didn't seem to give a damn, man. They just were blown out the door. We had an incredibly good Q2 almost across the board, right? This is the year-over-year growth rate. If we take out the COVID quarter in the middle of 2020 and then the year-over-year comping the lapping of that COVID quarter in 2021, you can see the normalized trend for earnings growth in the mid-teens. We're at 50% EPS growth for the S&P 500 and 88% of the stocks have now reported earnings. So, in aggregate to see that many companies up 50% is kind of astonishing. There is if you look at the comparison just to the estimates, the companies are 29% above estimate. The 5-year average for like beating estimates is 7%. The 10-year average is about 7%. So, when you beat earnings by, you know, by 29% not 7, that's real and if it holds up to the end of the quarter, it's going to be the highest earning surprise reported. This is by FactSet. It's the highest that I've ever seen. It's the highest that FactSet has ever seen beginning in 2008 when they began to track it. And so, it's definitely a strong quarter. A lot of that now is when you are a company like Google and you own shares in, you know, in Anthropic or in SpaceX and that marks you get a one-time increase in earnings and usually that will show up in something called other income. So, if you look at the P&Ls of these companies, you can see the one-time investment increases that they had. But, even if I back that out, I did the math on it and if I just back out these one-time earning surprises, I think you're still talking about a 22% year-over-year growth rate in earnings. So, taking out like the Anthropic markup and SpaceX markup, and then Microsoft reported a big OpenAI bump, and Meta reported scale. So I took out the really big ones, and was still left with like low 20s EPS growth. And the stock market's only up 13% year-to-date. So you put that together, you're like, 'Wow, multiples are actually compressing.' And so that's just a really good quarter.
So I was wondering, do you think that this is being weighted heavily just by the AI-focused companies, or are we seeing non-AI companies also report, you know, positively?
The number one industry was actually not IT, it was not tech, it was not AI. It was the energy sector. Which I guess is benefiting from high oil prices. While the rest of it I've taken on the chin. I have to pay 140 bucks to fill up my car, but and then financial services did great, the banks. I guess the banks got a lot of money for taking SpaceX public. So maybe, you know, you can look at that and say that was also a little bit extra juice in the quarter that may not recur. But even materials, like it's just hard to find any spots in the reporting that was soft. So there's a lot of AI help, but I think even without that, this is just a bang-up quarter.
I think I broke my ankle. But I'm going to put some weight on it and see. So moving on to SpaceX, we covered this a lot last week both in our news segment, and also in our special SpaceX earnings announcement. But after earnings came out, we were kind of expecting there might be a drop in the price because all the shares were being unlocked on Thursday. That did not happen. In fact, quite the opposite. SpaceX kind of went on a bit of a rally last week.
You know, I think what happened was you remember when we reported, we said, 'Geez, that was a good quarter.' You know, the quarter if I look at the top line, it was great. The bottom line EBITDA was great. They moved up their guides. They demonstrated they can grow and be profitable. The one thing they maybe missed was CapEx and maybe that's what the market focused on. I think the market had a couple of days to digest the numbers and realize actually this was a really good quarter. And then the lockup as you said happened on Thursday. Shares hit some funds on Friday, but not all the funds are distributing. There's another 7% lockup happening in 10 days on August 20th. There is going to be some continuous supply, so we'll have to watch that, but you know, I think the fact the stock got hit early on like a lot of these IPOs that have done well in the long run, there are big drawdowns from top to bottom for a lot of companies from Facebook to Twitter to Shopify to Twilio. A 40-50% drop from top to bottom isn't unusual. And I think the market was pricing in the expectation for this additional money coming out of lockup. You know, when it didn't hit the street and didn't cause a lot of selling, maybe there was a bit of a relief rally. And then SpaceX had some good news on Terafame and good news on an Nvidia and I think that also provided something of a positive catalyst in the last couple of days. I had a chance to go back and re-listen to the earnings call and I think there were a couple of things that weren't obvious on the call itself, but when you read through the transcript and listen to the answers, it does become a little bit more clear. So Elon did say at the start of the call and then he also repeated in the middle of the call to one of the answers that he was looking at a hundred billion dollar revenue run rate at the end of the year. Like that is a startling number and I think the market is still trying to figure out how to kind of digest that number, to believe that, but you have to believe I think that connectivity is going to be at close to 20 million subs by the end of the year. They're holding their ARPU flat. So their average revenue per user on a monthly basis flat. They said 66 bucks in Q2, it was 66 in Q1. They had doubled their subscribers over the last year. They're just about to put V3 satellites into space versus V2. That's a step change in performance for Starlink. So, I think that is very credible and that's 60 billion in revenue. Their AI business, they said on the call they were going to be at 2 gigawatts by the end of the year. And if you believe that and you think about the pricing that's going to happen on a per watt basis. So, right now the spot price for compute 30 to 50 bucks per watt. And so, if you have a billion watts, a gigawatt, that's, you know, $50 means $50 per gigawatt. That feels believable. And so, I think the market is also digesting, hey, this $100 number could be real. You notice I've actually zeroed out space. I've no revenues for space. We know they're going to put payloads into space. We know Falcon 9 is launching. We know Starship is coming. I don't think you even need to have any of that stuff to get to the $100 run rate. And the one thing that I also happen to put in here is Cursor. And Cursor's closing now. He said this week. And they're at $10 billion in ARR. So, I think this could be at $110 billion by the end of the year without anything for space. And that's, you know, 12 to 15 times ARR for a company growing at 5X year over year. That now seems believable. And then I think it's giving people a lot more comfort with the stock price.
You know, after seeing those numbers come out last week, they were, you know, even better than expected. And I think we're just going to see a ton of demand for AI compute. You know, Elon's really shifted dramatically in the last year from buying all that compute for expectations around xAI maybe not hitting.
But then pulling in Google and Anthropic as customers, maybe the Defense Department's customers, Reflection AI and other people. I think Elon is back to selling picks and shovels to everybody else who needs AI. We're hearing that AI, you know, revenue from OpenAI and Anthropic is probably north of 100 billion and probably will be north of 100 billion for Anthropic by the end of the year just by itself.
If he can get to that level of revenue and that is what he's good at, right? Like we know that he can do that. I think you're right that Grok was a little bit of a miss and but that's why they acquired Cursor and got a great deal on that. I think 60 billion dollars if they were at 10 billion ARR at the end of the year that's 6X revenue so you got to be end up being one of the best acquisition deals I think of all time. Like I really do think that was, you know, kind of a steal almost in hindsight. But I think that the story that people are maybe not quite recognizing is a lot of SpaceX's projected AI revenue is going to be coming from other AI companies' revenue. If Anthropic does well, SpaceX does well, too.
And Google and he signed a deal with Reflection and is probably, you know, there's probably more coming. It seems like he has found that's the bottleneck in the infrastructure stock right now and if he can work with those three or four partners. It's interesting that he's, you know, at odds he's been insulting Anthropic and OpenAI and been fighting with everyone in the space but now they're customers so he said I would imagine that's going to tamp down and they're all going to be buddy-buddy and, you know, and you're right those are the big customers. The fact that they're growing at 5X year over year and profitable that seems to feel like hey if I'm paying 12 to 15 times ARR for SpaceX from their end of year run rate that doesn't feel terrible. It's still a rich valuation but it begins to get into an area where I think, you know, institutional investors will start picking up the stock now if they actually hit those numbers.
You know, I'm kind of reminded in the movie The Matrix when Morpheus is talking about Neo, he's starting to believe. I think the market is actually starting to believe that he loves the projection maybe for real. Well, another announcement that was made this week was Elon talking about a big partnership with Nvidia and as a result of that I think SpaceX is going to be front of the line for Nvidia chips or some predetermined percentage of Nvidia's chips and that's a big benefit to them, right? Come on.
It is interesting that they're working with Nvidia. Elon typically is pretty notorious for, you know, wanting to build his own internal picks and shovels and supply. Like Tesla does that. They don't really outsource to third parties. The fact they're working with Nvidia and giving them an exclusive it sounds like. I mean that's really big for Nvidia. And it also seems like if they are competing with, you know, Coreweave and other neo clouds, that exclusivity gives them a first dibs on what Nvidia produces. That seems like a win-win. That is a big deal though.
Elon on the earnings call had said that there was a non-zero chance of hitting their trillion dollar annual revenue milestone although earlier than expected. They previously talked about that in 2030 but now talking about that potentially happening in 2029 and even like you said before that hundred billion dollar run rate being hit by the end of this year. I guess we're seeing Elon spend more and more on CapEx that's really going to be utilized as much by other providers, by other compute users than just SpaceX itself.
