Bloomberg Tech is live from the heart of Silicon Valley with Ed Ludlow in San Francisco. OpenAI is looking to embrace fresh funding with a $1.25 trillion valuation. Anthropic could deliver blockbuster gains for some of the biggest investors with billions of dollars on the line. SpaceX is taking another step toward becoming another major U.S. wireless carrier, buying access to at least 80,000 cellular sites nationwide. We have made it. It's Friday.
Technology stocks are higher but headed for two straight weeks of declines. Back-to-back weekly drops on the Nasdaq 100 since July. Not as bad this week and not as deep of a drop as last week. A lot of the news is in the private market. I want to get today's top number, 1.25 trillion dollars. That is the valuation that OpenAI is targeting. Joining us for the details, Rebecca Torrent. We reported that this round is underway. 30 billion dollars raised. It looks like the valuation is coming off of it. What else do we need to know?
We are reporting that Code 2 Management is in talks to co-lead this round. The total size is roughly $30 billion, but the price has come down a little bit. We are looking at a $1.25 trillion valuation, that is down from the previous price, which is reflecting the price of the broader market.
When we first broke the story, people were like, why is AI doing this? We have reporting on when they might go public. It all played out where we got to speak to Sam Altman about it. It seems like this is the investors wanting to do this rather than OpenAI needs the money.
That is how it is being positioned. I will say they do have these commitments. The work they are doing is not cheap. They can certainly use the money. This money would be a bridge to as early as this year, but now they are looking at next year at the earliest. This capital will get them to the next milestone, ideally.
They have already raised $122 billion. What they told me when I asked if they were doing a new round, a lot of that prior round's capital sits on the balance sheet. Let's frame this in the race with Anthropic. What was Anthropic's last public valuation? Where are we at?
$1.25 trillion — Anthropic's was $965 billion. This private round for OpenAI would push it above Anthropic for the highest valuation in tech. That said, it looks like Anthropic is going to go public as soon as November, in which case it would no longer be a private company and OpenAI would be the highest-valued company.
Thank you very much indeed. Anthropic's upcoming IPO will bring multibillion-dollar gains to investors. Iconic owns a stake of $17 billion in the company. They could see the biggest returns. This is one of the most read on Bloomberg today. The reporter is with us. There are lots of big names in Anthropic. Some of them have been there for a long time. Some of them haven't. Who is going to win?
This is definitely a story of the biggest investments leading to the biggest returns. Concentration is a word that we hear a lot about this year, and in this case many of the firms — Iconic, Sequoia, GIC — made their largest investments of all time in this company, expecting their largest returns. We have to give credit to the early-stage firms. Anthropic did not get institutionally backed until 2023, when Spark Capital had its largest investment at the time, $75 million. Now we expect the stake to be valued at $7 billion.
We just showed you the backers of Anthropic. For the first time in 2023, it wasn't that long ago. What we call strategics — Nvidia, the cloud computing players. Tell me about more recent history in those backgrounds.
It is where you see the biggest returns for an investor that just backed it last year. With Iconic, it had passed on Anthropic before. With GIC, we reported that it has invested upwards of $4.5 billion in the company, with the first check starting last year. Period of time. We are both anxiously waiting. We don't expect them to be in the 5% ownership bucket. It's important for us to get the story out to understand what they have before then.
A $17 billion stake, that was based off of Anthropic's last valuation.
Correct. We will have to correct this when we do get prices. $2 trillion has been discussed. We are actively chasing that.
I wish you a peaceful and happy Friday. Whether they come or not.
SK Hynix scrapped its plans after it faced opposition from regulators and shareholders, making it the latest firm to hold an IPO due to growing investor concerns. The announcement, writing as much as 4.5% of shares outside of the United States. SpaceX taking aim at the giants with plans to access tens of thousands of cellular sites across the U.S. That's next. This is "Bloomberg Tech."
SpaceX is looking to secure 80,000 cellular sites across the U.S. to build a fourth nationwide wireless carrier. A person has told us that the company is in discussions with one of the U.S.'s largest owners of cellular towers. This is the market reaction. Shares have surged. There is movement in the carriers as well. Kelcee broke the story. This all started — SpaceX's first earnings call was a big moment where they dropped the bomb about what they wanted to do. Can you go through what you are reporting on their latest move?
Yes. This detail is significant. It's the first indication that SpaceX is serious about building out this ground-based network. Up until this point, they discussed strategies around incorporating satellite and some mobile nodes into their consumer home internet terminals as a novel way to put together a network. Now the reporting shows that they are actually planning to do things the way that the big wireless companies do things. That doesn't mean they won't bring in some novel elements like coverage from space, but the towers are a traditional play for a traditional wireless network.
The audience is probably more familiar with Starlink, which is constellation-based internet, space-based internet. They are looking at a path to Starlink mobile, but they are doing it very much on Earth. Why is the cellular tower reporting important? What does that tell us?
That tells us that SpaceX is looking to scale up in the same way and on the same level as these wireless competitors, AT&T, T-Mobile, Verizon. They are trying to get on their level in terms of ground-based infrastructure. Until this point, I don't think we have seen that. Again, in earnings we did hear from Elon Musk and others that they do plan to build out this ground-based terrestrial network, in addition to requiring spectrum and building their satellite network from space. This is the first indication that they are really seriously looking at this.
At one point in the session, shares were up 16%, on track for the biggest jump since 2008. Thank you very much. SpaceX is not just thinking big with wireless. Elon Musk has ambitions for AI infrastructure, another sign that the buildout is not slowing down. Investors are increasingly trading on AI vibes and not fundamentals. Jordan is back with us on the show. I'm trying to make sense of the trade. Explain how you see this as being vibes and not fundamentals.
Ultimately, fundamentals should drive stock performance, but it is the price determining the narrative. We are in the midst of earnings season. As an analyst, I think myself and my clients are trained, and their job is to understand what these companies are saying, the forecast and the estimates. Those are looking good, but the stocks are not trading as such. They are selling good beats and raises, yet maybe they are selling off. I think everyone's nervous that there might be smoke in the movie theater, and they don't want to be caught getting burned out if things get lower faster.
Can I ask you about SpaceX? In the context of what we learned, bigger picture, their goal for next year is to have 20 gigawatts of compute capacity. A lot of people would say, do we extrapolate out from what SpaceX said in the same way we would do with hyperscalers to give us the confidence level of what the next 12 months looks like? What was your take on that?
Even their base case of adding 10 additional gigawatts is probably 80% — it's a lot more than people thought. Whether they can get 20 is another story. That depends more on the availability of the power source as opposed to anything else. I think it sets a bullish tone. It probably sends a message to the industry that we are going all in. Mark Zuckerberg at Meta is doing something similar, where he would rather have more compute. Both SpaceX and Meta and these guys are working on new capabilities, things like agentic AI. They know those things are going to be compute-hungry. They need plans in place to get this compute set up now, or else they could fall behind and not be able to monetize these capabilities.
There are some warning signals in credit markets. For example, there has been sustained pressure over two weeks on the AI trade. I know you have said this is being done on fundamentals, but what is the source of the anxiety around the AI trade right now that you are hearing each morning?
It's a couple of things. First and foremost, people are nervous that these massive amounts of spending are unsustainable. You cannot finance it. People watch rates for these compute vendors, like the neo-clouds. Secondly, why would anyone build all this and spend the money if there was not enough monetization or revenue potential, right? It is still an economic decision. You don't invest in something where you are not sure you are going to get the return. So much of AI is new, novel, involving. It is hard for investors to see that far out and say if they spend $2 trillion per year, they can get a return on that. Every three to five months, basically, we see new features and capabilities that AI can come in and do or save. I think that changed the dynamic. That is why people are so laser-focused on these frontier model revenue run rates, the Anthropic and OpenAI ARR numbers. If those don't continue to accelerate and grow, there is going to be somewhat of a mismatch between the spend and the return.
Let's keep talking about that. The more imminent thing is an Anthropic IPO. The reporting is they would like to pull it off before Thanksgiving. Again, how does that impact your day today and your clients, in terms of something they are bracing for and think will happen to equity markets more generally?
That is a good question, and it is twofold. They need to see this deal, which is the most high-profile AI frontier model company, be successful. By that, there are various definitions of successful. I think in the first 30 days, it has got to trade well and not basically go down from the first opening price. There has to be demand. There has to be people thinking it is a good investment. That is going to be dependent upon the valuation. I think we will know the fundamentals by the time it goes public. Secondly, and this is not to be underestimated — if you are a portfolio manager looking at should I buy this or own this new issuance, you have got to find room in your portfolio to fund that. People are going to want to know, does that come at the expense of another competitor, a SpaceX, a Tesla, some of the semi companies? It is unknown. It is going to be pretty big. If you are trying to keep up with a benchmark, you might need to have exposure. If you are a hedge fund, you don't have to own it unless you think it is going to give you alpha.
SpaceX fast-track inclusion — those managers had the same thought before it went public. Happy Friday. Great to have you back on the show. Apple's $2,000-plus iPhone duo is on sale. People around the world lining up at stores overnight with preorders, or to try their luck on a walk-in purchase. Large crowds have been rare in recent launches, but inventory for the foldable phone is drawing people to stores. They are placing orders on the Apple website, and the quoted delivery time in the U.S. is the end of January. The same for me here in California. Coming up, a startup raises 250 million dollars to fund efforts to remake the 50-year-old GPS system. We will speak with the CEO next.
It is time for Talking Tech. Jensen Huang announcing another AI investment, this time into an AI video startup. The San Francisco-based company is on track to surpass a billion dollars in annualized revenue. This investment will help fund the global infrastructure and hiring. Plus, approval to deploy models overseas. A Chinese state media affiliate posting on social media that the company has completed the new registration process that Chinese AI companies must complete to operate globally. Much of Chinese AI is similarly going to remain free on export. An examiner stepping up scrutiny of valuations in AI, focusing on fees, evaluations, and private funds, as well as how to prepare for AI-related cyberattacks. These are the latest signs of regulatory concerns about AI and private markets. A startup working to improve GPS with a new purpose-built commercial satellite navigation network has raised $250 million. The funding round was led by a longtime SpaceX board member and valued the company at 2.5 billion. The co-founder will join you in Talking Tech.
That is right. Brian Manley is with us. Congratulations on the round. Big numbers. What I would say is let's start with the problem you are solving for. People will be listening, going, GPS, I thought we cracked that. What is the problem here?
Appreciate you having this on, and that is a question other investors are asking. With GPS, it works. It is something I use every single day. What are the challenges? The world we came out of was the autonomous car world, and realizing the current navigation systems do not support the precision, the power, the protection that is needed for the world to evolve from the hands-on world it is today to a hands-off world. That extends beyond cars. That extends into agriculture, drones, robotics, defense. All these different things that depend on location and time need a better capability than today.
We're talking about computing at the edge, and that can mean anything in this day and age. So what is the technology that you are developing that serves as the fix?
There were two big breakthroughs we made as a company that really enabled us to do what we do. The first one is that historical GPS satellites are these big, multi-hundred-billion-dollar monsters. We figured out how to provide high performance from small satellites we can build for hundreds of thousands of dollars at scale instead of hundreds of millions of dollars, right here in the factory I'm standing in. That enables us to go from satellites that are 20,000-plus kilometers away from Earth down to small satellites that are 20 times closer to Earth, and enable us to get massively better performance than you can with legacy systems.
Let's talk a little bit about the round. It is a big Series D. Congratulations on it again. What is the priority for you? Where are you going to deploy that capital to get you moving on this?
It is coming off a huge week for us. We launched six new satellites actually on Tuesday morning. These are some of the first satellites that will start providing the world's first commercial satellite navigation service. At this point, we have a pretty massive backlog of customers, both on the commercial side as well as the government side, that are really just waiting for us to get the satellites up in the air. The new capital is going toward building more satellites and getting them up in the air to deploy the service globally and expand our capability.
In August, you had a big approval from the FCC. Can you explain the significance of that moment, why it was unique?
Spectrum rights is one of the biggest bottlenecks for any company. Spectrum is a very limited national resource. We have been coordinating for the last five-plus years to figure out how could we share the GPS spectrum efficiently without causing any sort of interference. That is what really enables our service to be able to integrate into billions of GPS devices around the world without changing the hardware, just updating the software. We have done massive amounts of engineering. In August, we got the full approval for our full network, which is the last puzzle piece that gives us the regulatory approvals that we need to go deploy the service at scale now and start turning on customers.
What is at stake here — what I find fascinating is that if there was a disruption, it could cost the U.S. a billion dollars a day, and that is the thing you are fighting to solve. Great to have you on the program. Brian Manning raising a $250 million Series D at a $2.45 billion valuation. Coming up, we talk about the latest growing competition in early-stage investing, what we are seeing in venture and private markets, and an innovation right now. This is what markets look like this Friday. In aggregate on the week, we are down for the second consecutive week on the Nasdaq 100. Lots more to discuss. Half-time in the city of San Francisco. Happy Friday. This is Bloomberg Tech.
When you think about a bubble, you think about building capacity. With AI, you are building capacity, and you are hoping you can keep up with the demand.
Two-thirds of the U.S. economy's capital expenditures right now are on technology, whether it is cloud or AI. And if we slow down on the build economy...
I don't think that we are living in a bubble. The fear that we are in front of a bubble is justified.
These fumbles often coincide with these giant transformations, and the companies on the other side that troubled by the market leaders for decades to come. Today, it is really an indisputable shortage of compute, and it's hard to deliver that compute, and that is what is preventing some of the access as seen in the past, but they call bubbles.
Welcome back. That was some of the top figures in finance addressing fears of an AI bubble this week on Bloomberg Television, and there was a range of views also reacting to fears of an AI bubble, credit markets. The highest point in 17 months, or at least, the inverse relationship between U.S. government bonds and AI-focused stocks has strengthened to its highest level in a year. I mean, are fixed income markets reacting to something that is genuinely a concern, real and in front of them, or is this markets and technicals? It's not really my strongest point.
