Bloomberg Tech is live from the heart of Silicon Valley, with Ed Ludlow in San Francisco. Ed: This is Bloomberg Tech. Coming up, OpenAI aims to raise $30 billion from investors in a new round of funding after pushing back its plans for an IPO. Plus, President Trump supports using outside auditors to assess the safety of AI systems after meeting with Silicon Valley leaders yesterday. And Apple keeps pushing the envelope under new CEO John Ternus and plans to make its foray into the smart home market. Good morning from Los Angeles. We'll get to why I'm here later in the show. Let's get straight to our top story in today's Big Number.
$1.4 trillion, that's the valuation OpenAI is seeking as it looks to raise at least $30 billion in a new funding round, all according to sources. The potential raise would give the ChatGPT maker fresh capital after pushing back plans to go public. Bloomberg's AI editor joins us now. Seth, you worked on this story with me yesterday. Let's be very clear. This is Bloomberg's reporting alongside and parallel with OpenAI having its Dev Day, and it sounds very much, based on our reporting, this is the investors pushing for this, right? Seth:
Yeah, that's certainly the impression we got here. That said, I think OpenAI at this point, there's no clear timetable for when this company might go public. Our understanding is it's looking potentially for a bridge round to give it additional capital, while it continues to push forward on these very cost-intensive infrastructure initiatives and model development to give it the bandwidth to continue to build without going public in the next few months.
We had two critical conversations yesterday, which we will play out later in the program. OpenAI's CEO, Sam Altman, and OpenAI's CFO both addressed the IPO timing. And in Sarah's case, they had a lot of capital on the balance sheet, and she responded to our reporting a little bit. It was your job to take that, put it in the story, kind of summarize it. Seth:
Yeah, I think she said a couple of important things to you. The first is this is a very well-capitalized business. They raised $122 billion. Certainly, as this company looks to push ahead with new data centers and its own chips endeavor and other efforts to compete for top talent, there might also be an opportunity and a need to bolster its coffers, so to speak.
The headline of yesterday morning, before the Dev Day kicked off, how is OpenAI's business doing? There were numbers. What did we learn about the health of OpenAI? Seth:
The narrative in the last quarter has been that business is accelerating. There was a perception, probably more true than not, that the company was on the back foot as Anthropic was more ascendant. But the company has narrowed its focus to AI coding and enterprise tools and started to regain momentum of its in-house models. It seems to be experiencing renewed consumer and business adoption, including, again, on the consumer side, probably took a long time to have a billion users after slowing a bit. Now they're saying they've hit 1.2 billion just a couple of months after announcing one. On both sides, they're starting to fire again on all cylinders, and that's a good story if they're raising more capital.
Seth Fiegerman, thank you very much. As billions more pour into AI, tech leaders are making the case for opportunities to come with that investment, while continuing to warn about the risks. Here's what they had to say after meeting with President Trump at the White House.
Whoever wins AI, wins. I think that's very important. But I think the technology has very real risks. You know, the mechanism, how we address those risks, is still under discussion. But we all need to work together to make sure that we win. If we do this right, work with the president, everyone here, we can win.
This is the largest industrial revolution in history. And we're going through the single largest infrastructure buildout, so it's creating a million jobs here in the United States. We have chip plants coming up, computer plants coming up, factories coming up, and so it's creating just enormous amounts of jobs.
The basic idea is that we want to give the American people and our customers confidence that the technology works in the way that we intend it to.
We're obviously going through change, but it can be a computer. It used to be a job. It used to be people doing calculations. But now you have digital computers to do those tasks. So I don't think anyone is pining to go back to doing calculations all day, which is what they used to do.
Let's get more from Washington. Our senior tech editor, Mike Shepard, joins us now. There is an agreement in the way in which the CEOs in the world are going to work together, and there is a position from the government. Explain it. Mike:
Well, Ed, this agreement emerged after a nearly two-hour lunch between President Donald Trump and some of Silicon Valley's biggest names, and we all just saw them there on the screen. Now, this agreement is not something that is binding legally, but the president said it is morally binding. What it calls on is for companies to employ independent auditors to assess the safety of models. It calls for them to strengthen their internal controls of those AI systems, to make sure that they don't go rogue, and it also calls on the companies to collaborate more and to talk about safety and meet more regularly. So we could be seeing a repeat of this, perhaps not with the president at the White House the way we did yesterday, but it does suggest that there will be further conversations on this. The fact that the president was willing to entertain this talk at the White House is really a sea change for him. He's rejected a lot of these doomsday fears about AI as a hoax. So to see him welcoming all these executives there for this discussion itself marked a big move by him, perhaps a change in direction for the administration too.
In that photograph that we're showing on the screen right now, there is an absentee, which is Sam Altman, the CEO of OpenAI. He was with me in San Francisco at Dev Day. Greg Brockman, the president of OpenAI, there. As far as I can tell, a lot of the focus was on what Dario Amodei had to say in the press pool, calling for him to come forward, talk a bit about how he felt the meeting had gone. Mike:
Well, Dario was really the focus of this meeting. In a way, he's the one who helped to bring it about with that essay in September, warning of the existential risks, promising to slow development of the company's most advanced models, calling on the industry to do the same, and actually winning support from Sam Altman and Elon Musk to perhaps slow the roll a little bit on the most advanced AI models, to understand their risks better. But all of that has put him in conflict with President Donald Trump, who wants to see no slowdown whatsoever. He sees AI as critical to his economic agenda and to keeping a strategic edge over China. So hearing Dario talk about the risks in front of the president in the same way, but then to hear the president also say, 'Dario, you're doing a fantastic job,' it was really quite a moment, because the two have been at such odds for really months. But they managed to patch things up, perhaps with this dinner on Sunday night. We didn't get a full readout of it, Ed, but it really did seem like Dario is in the fold and acknowledged and tolerated as maybe that voice of dissent that they need to help propel a document forward, like the one that was signed yesterday.
The president also mandates we now call it superintelligence, not artificial intelligence. Mike Shepard, thank you very much indeed. Coming up in the program, we'll be joined by the CoreWeave CEO on the company's latest lineup of tools for developers. Don't call us the neo-cloud. This is Bloomberg Tech.
CoreWeave is out with a new set of development tools. They look to diversify its lineup. The product is designed to help clients train and improve their AI models. They announced it earlier today at the first user conference in San Francisco. Delighted to say Mike Intrator joined us from there. This is interesting, right? I know that there's a discussion you and I need to have about don't call you a neo-cloud, but what is the business case here for CoreWeave offering this kind of software solution? Mike:
So, look, first of all, thank you for having me. Always a pleasure to be on the show. I think what you're seeing from us is the systematic pull by the clients that use our infrastructure up the stack, and they are looking for a cloud solution to the distinct and unique problems that are associated with artificial intelligence. And CoreWeave is the company that is leading the delivery of the environment that is required to be able to build the next generation of artificial intelligence tools, everything from, you know, the frontier labs through 100-year-old companies like Caterpillar, where, you know, people need access to the tooling that allows them to confidently do what they need to do for their companies. And that's really what Forge is there to do.
CoreWeave is the original neo-cloud. Why should we not call you a neo-cloud? Mike:
So we were the original neo-cloud, and now we are the original AI cloud, right? It really has to do with the fact that, you know, a neo-cloud is something that exists at a point in time, but what we have built is the cloud solution for artificial intelligence. And when you build from the ground up, when you build specifically to serve a particular type of use case, you begin to build the tooling that allows for the optimal delivery of the service, and that's really what we've done. And we don't think that's a neo-cloud, and we don't think it's bounded by time. We think it's going to be a long-term, very successful business that is going to be required by the world as it increasingly leans on artificial intelligence.
Very quickly, this will be meaningful to you financially. This will be accretive to margins. Mike:
Yeah, it's going to be wonderful. It's really going to allow an entire class of companies that have struggled to integrate artificial intelligence, either because they don't have the skill sets to manage the infrastructure internal to their company or otherwise. It's a developer toolset that allows folks to move into artificial intelligence quickly, efficiently, effectively, at the right pace and cadence for their companies. That's really important. We meet them where they are and provide them the tools.
Mike, help me understand in real terms what pacing the frontier of AI means for an AI cloud. When this all started, the assumption was if you pace the frontier, there's a drop-off in activity, so that demand story for compute changes. What have you actually seen? Mike:
I think that narrative is misunderstood, deeply misunderstood, actually. Let's kind of bound it, right? The leader in artificial intelligence, it is imperative that we remain at the forefront of artificial intelligence. And so when you're talking about pacing, it's a rather narrow definition of what it is, right? You need to be able to maintain your position at the forefront, but you also need to build guardrails to ensure that you can deliver the best quality safely, and that's really what pacing is about. The net effect of that is not in the demand that we are seeing as companies continue to procure the infrastructure, the software, the cloud services that they require to build their businesses over time.
What is the political threat or headwind here right now? We're so close to the midterms. That has been debated everywhere. Is there something tangible you can tell me about how that does impact you? Mike:
Yeah, look, we have really been the leaders on operational and the infrastructure for artificial intelligence. My last earnings call, I talked about how we brought on 500 megawatts worth of active power into our infrastructure. There's an enormous, enormous amount of scale. And part of that comes from our approach to how you go about building data centers, where you enter the communities early. You work with the communities. You make sure that there are community benefits. You make sure that their voices are heard. It also is a recognition that not every community is going to have a data center and that we don't have to force the issue. Fundamentally, moratoriums do not alter the demand for compute. They simply move where the compute is going to be built. You know, we're not trying to force this on anybody. We're trying to work with the communities where we build our infrastructure, and we think we do a really good job of that. We think we get a lot of credit from the communities for the work that we do there.
In your risk factors, in different regulatory filings, you might need to look at non-NVIDIA solutions. I sat down with some of the specialized platforms that are out there. Where are you in your journey on non-NVIDIA or non-NVIDIA GPU-based architectures? Have you done anything there yet? Mike:
Yeah, we monitor the full spectrum of accelerators that are available, that can drive artificial intelligence companies forward. Our job is to deliver the best solution in the world, on the best technology in the world, through the best software in the world. And right now we don't have enough infrastructure to stand up enough of the NVIDIA infrastructure to meet the needs of our clients. We are overwhelmed. We are largely sold out. We are sold out over time for several years. It's really a function of being the best at the job that we do, which is deliver the best solution in market, which is the NVIDIA solution from an accelerated perspective. We do a great job of it. We have 5,000 people. The conference was sold out. It's people that want to get onto our cloud and build with us on our infrastructure, which is really grounded in the NVIDIA GPU accelerant.
CoreWeave, great to catch up with you. Thank you for making time with us. Coming up, we're going to speak about breaking ground on its Texas shipyard. This is the defense tech. This is Bloomberg Tech.
The way things used to work between the U.S. government and defense contractors was pretty straightforward. The U.S. military would define specifics. Several companies bid for it. One of them would be selected, spend years developing capability. Now Lockheed Martin says things have changed with the new administration. Listen to this.
What is different now is that the current leadership in the Pentagon has decided to take the straitjacket off the companies, which allows us to take the straitjacket off our engineers and scientists, to allow us to be flexible, to propose solutions, to do prototypes. We go out in the field, and the field could be the Pacific Ocean, of course, for us. And we test prototypes on actual ships with sailors that are stationed on the ships, running the prototype to see if it works.
That was Lockheed Martin's CEO. Tune in on Friday. More of that conversation during Bloomberg's Defense Tech Special, 11:00 a.m. Eastern, live from Skunk Works, Lockheed Martin's advanced aircraft development in Southern California, one of several reasons I'm here in L.A. today. Saronic is breaking ground today on what it says will be the largest and most advanced shipyard in America. It's in Brownsville, Texas, and it's at the center of the company's push to rapidly scale production of autonomous ships. Joining us now is Saronic co-founder and CEO, Dino Mavrookas. Dino, great to have you back. Just summarize why this is such a critical moment. Dino:
This is a revitalization of a critical industry here in the United States. I mean, you've seen the administration call for the revitalization of shipbuilding here in this country. The vice president is actually joining us today at this historic groundbreaking. This is the first new shipyard built in the United States since World War II. And we are investing billions of dollars. We are going to create thousands of jobs. And we are going to start building ships again right here in this country. This is American ships, American jobs, and it's the revitalization of an American industry.
Dino, speed. In your field, what Lockheed Martin wants to do, speed. So how quickly does that facility actually operationalize and get that autonomous out there?
Speed is everything. We started Saronic four years ago. In four years we raised $2.5 billion, and we're standing in Brownsville, Texas, getting ready to break ground on the largest shipyard in the country. Speed is everything, because our adversaries aren't waiting. The shipbuilding industry in the United States has declined dramatically. China has invested heavily. We need this now. So we're going to be operational here in Brownsville at Fort Alpha by the end of 2027, with our first ships off the line by the end of 2028, and we already have orders for those ships, delivering LCE vessels to the Navy and Marine Corps.
The orders are there. What are your capital needs to make that happen, run production, get it out? Dino:
We've already committed over $3 billion to this project. We've raised $2.5 billion in private capital to date. We are putting capital to work. We will raise more capital as needed. And we will keep moving quickly ahead of demand to make sure that the capacity is online to deliver not just for defense, but ultimately for the commercial market as well.
Your platforms have been deployed very recently in the context of the war with Iran. Is there any update to that, how the U.S. is using your platforms? Dino:
I mean, I obviously can't get into the specifics of how they're being used, but what you're seeing is autonomy being proven out real time on the battlefield. You mentioned the platforms that are overseas right now. That's a 24-foot, fully autonomous speedboat. We can build thousands of those vessels every single year, and we've already delivered hundreds to the Navy. And what you're seeing, by using these, we rescued two downed Apache helicopter pilots in the Strait of Hormuz. It's the trust and the confidence from the military to use these autonomous systems in the highest-stakes environment and keep actual additional people out of harm's way.
Speed is one key factor. What is the economic proposition per unit of your platform? If you think about the Air Force's program, they want $20 million per unit on an aircraft fighter. Is there something comparable where you can say, look at the economics that we offer versus other alternative solutions, legacy solutions? Dino:
