Bloomberg Audio Studios, podcasts, radio news. Bloomberg Tech is alive from coast to coast with Caroline Hyde in New York and Ed Lovelow in sentences, go.
This is Bloomberg Tech coming up.
President Trump says he discussed in video's eighty two hundred chips with Jijingping.
Will break down the latest.
Plus we speak with the CEO of Figma after its earnings results, the five fears that AI would disrupt the design.
Stack, and our conversation with open Ai CFO Sarah Fryer, who says the startup may raise more capital after completing its recent fundraising round.
Carry has got the markets.
I have, let's check in on them.
We're currently seeing the n AS that one hundred actually having its worst day since March the twenty seventh. At the moment, edre by one point three percent. Look over the course of the week, it's not nearly so ugly.
We're actually inning down about a tenth percent.
But on this day it is the chip stocks in particular, and I'm looking at Micron on the downside.
That's a key point drag. You're looking at Invidia as well.
Look, this is the day in which we digest what happened over in Beijing.
Yes, and that is reflected in semiconductors in part. So the Philadelphia Semiconductor Index or SOCKS is the main gauge of chip makers, chip equipment makers. It's down almost four percent. You have to take into account the astonishing rally in chip stocks year to date seventy percent or something prior to yesterday's close.
In video at one point in the session, it's off.
Session lows was down on track for its biggest decline since February. In part, as I say, because the market's digesting, what was the net outcome of the meeting between President Trump and President G four Invidia And for chips, his President Trump talking about in video's eight two hundred chips on Air Force one earlier today.
As you know, Jensen was, there's amazing and the video and.
He would be a fla.
You know, they have much higher level than the HQO hundred, but the H two hundred is good Chinadja And so yeah.
President Trump there on Air Force one. Let's get the latest with Bloomberg's Tyler Kendall, who joins us once again from Beijing.
Chips in focus for US.
On Bloomberg Tech Taiwan in focus between the President and President g What else do we need to take away from this historic meeting?
Well ed? At this point, we're actually getting some breaking news on how China views how this visit went between President Trump and Chinese President Jijingping. Chinese state media reporting moments ago that China feels that Taiwan is that number
one issue for US China relations now. President Trump and US officials here on the ground maintained that the US policy regarding Taiwan has not changed, though the President Trump told Fox News earlier today that he would like to see tensions cooled down, in his words, between China and Taiwan. When pressed about future US weapons sales, the President was also noncommittal as Congress waits for him to prove a
fourteen billion dollar package that's been queued up. That was clearly the biggest point of contention amid a summit that was rather cordial, and both sides said that they were prioritizing stability. But we didn't really get a lot of tangibles and deliverables in terms of progress. Right We're still waiting on some key details regarding planned commitments. When it
comes to purchase agreements or new investment deals. Though we can confirm that Boeing officials were still here on the ground in Beijing over the last few hours meeting with Chinese officials and what was considered a positive sign for that deal.
But as you.
Mentioned, there there was a big expectation that perhaps there could be a deal when it came to in Vidia's H two hundred chips, as the Nvidia CEO Dnsen Wong was a last minute addition to the travel here to Beijing. President Trump did confirm that they spoke about the chips, but a timately said that China wants to develop their own and that's why we hadn't seen any purchases go
through Ed and Caroline. The President also mentioned that the two sides did discuss the future of artificial intelligence and where we could see some collaboration in terms of guardrails related to the technology.
For the most Tyler Kendall extraordinary work throughout the past few days over in Beijing. We so appreciated. Look, we've got to get there. For the broader chip industry perspective, what does all of this mean in terms of the United States and its ability to drive up manufacturing capacity and its workforce right here.
Sharry List is with us.
It's a vice president of Global Workforce Development Initiatives at SEMI. It's a global industry association connection professionals worldwide across the chip and electronics design and manufacturing supply chain.
It is wonderful, Chary, to have you here with us. Thank you so much for having so.
If we do see tensions rise, and if we do focus therefore even more on a manufacturing footprint brought back to America where Chips have manufactured their design rather than make so reliant on Taiwan, are we able to do it?
We are?
I mean, I think I speak on behalf of the workforce component involved this. It is one of the bottlenecks for this industry in the US, for sure, but there are a remarkable set of programs that are being built all around the country to meet that need. The Chips investments here in the US are launching an incredible growth here for us on our soil here in this country, and we're going to need another one hundred and fifty years one thousand or so people in this work environment.
So we're building programs all over the.
Country away from the geopolitics. You know, the news of the week highlights just capacity reliance. Right, So the one thing that Taiwan's really good at is the brutal economics of semiconductor manufacturing, and talent's a key component of that. Of course, there's all this plan on paper to build more fabs, have more foundry capacity in the United States, Do we have the people with the skills to make them run and not just make them run, make them hum right, brutal economics.
Yes, I mean I think that's what we're all trying to work towards. Absolutely, had we didn't have as many programs established in the US anymore because we weren't manufacturing here. With the investments that are happening here, programs are launching all over the country. In fact, under the Chips investment, there's a two hundred million dollar workforce This.
Is the Chips Act. A result of the Ships actually, sorry.
As a result of the Chips Act, there is a two hundred million dollar investment in workforce through the National Science Foundation in concert with the Department of Commerce, to invest in building a national infrastructure around workforce development so that we can fund regional nodes around the country to build what's needed regionally in workforce and how to feed that into a national infrastructure.
Not just jump in ofs.
I mean Tim Cook very famously said, didn't he that China could fill a sports stadium with tooling engineers and that type of talent in America would struggle to fill a meeting room. What kind of roles are we talking about here? Show that you're trying to skill up.
This nation on Yeah, I mean, I think we need everything across the industry, from technicians and operators to fill the fabs, to be on the fab floors, to all sorts of engineers across electrical engineering, mechanical engineering, chemical engineering, to our researchers, our PhDs. We need everybody. We need marketing talent, we need finance talent. So I think the challenge right now in the US is actually the image
and awareness of our industry with students. So we do a lot of work in the space of getting students excited or passionate about this industry or.
Even to know it.
Because kids walk around all day every day with their phones, the phones, their iPads, their computers were in cars that are driven through chips. We are using appliances, everything we use all day long, and most people don't know that, parents don't all know you know, So it's it's an educating of the country. Really.
Well, it's been so interesting is the method by which we start to up our fabrication footprint. Now, in some ways it's about leaning into intel, into local plays, but a lot of it's been about how drawing TSMC and saying please build here in Arizona. How much in the past have we relied on sort of TSMC talent coming here. Is that a way that's been reskilled and we've learned from others, or is it really about teaching them in our own.
Education system, seeing others being brought up through the STEM education perspective, rather than learning from talent abroad.
There's certainly been a mix of both. I think we have, of course, relied on talent abroad. As an industry. This is a global industry. This is a really intense, intricate, complicated industry, and we need talent from everywhere, right, that's clear. What we're trying to do now is make sure that we can build the workforce in the US with US citizens to get US jobs and to fill all of these roles. So we are still learning of course, we're
all learning from each other. We all have different strengths across the world, so you know, hopefully we'll be able to meet those needs here in the US with all the programs that are being established.
We're out of time.
But actually Caroen I didn't even think about that, right. What did not come out of the meeting again between the two presidents probably the issue of visas right and talent here in Silicon Valley, San Francisco. How long has that been a story, you know, talent from China coming to universities, particularly in the world of AI, particularly in the world of AI. Shery list is Semi. It's great
to have you on Bloomberg Tech. Thank you very much. Listen, guys, tune in Monday for an exclusive interview with the CEOs of Dell and Nvidia. From the sidelines of Dell Technology's world. Will be jennying over to Las Vegas for what is could not be a more timely conversation. Coming up, Cerebras surges in its trading debut, turning summer Silicon Valley's earliest backers into billion dollar winners.
US.
Next, this is Bloomberg Tech.
We have got to check in on Cerebras.
What a debut yesterday extraordinary sixty eight percent higher close at one point eighty nine percent. Unsurprisingly, there's been a profit taking company today more by some four point eight percent, let's call it on the day as we see the rest of the industry being under some pressure across the AI in chip spectrum because of concerns about really where
the US China relationship goes. But three of the chip makers earliest venture capital backers al cerebra said Benchmark, Eclipse Foundation Capital, they're poised to make billions from their bets following Cerebras's IPO here for more has been a bouth VC and startups reporter Rebecca Times, So how early do they back this company and what sort of rewards are they going to be able to give LPs?
Yeah, so two of the are sorry, three of the four biggest backers in Cerebras of their biggest outside backers came in a decade ago. In twenty sixteen, they invested in Cerebras's earliest round on the order of twenty five million dollars, and they now stand to make billions of dollars each at the IPO. Benchmark is among them inter firms with the biggest steak now around eight percent, and this would have you know, this was its first hardware investment in over a decade at the time, and that
big swing has really paid off for them in spades. Obviously, the stock trading up massively from its IPO price of one hundred and eighty five dollars per share rebackgau.
I don't know if you noticed yesterday, but when Andrew Feldman, the Cerebris CEO, was on the show, le Or Susan, the CEO of venture firm Eclipse, was hanging over his right shoulder every couple of seconds gave a little cheeky glance down into the camera lens. Eclipse another firm you know, regulars on this show that made a lot of money.
Here.
It's really fun reporting on this with you. It's like, you know, it's a really important IPO day. Look what else do we need to know? And just reflect what your week was like covering this blockbuster listing.
Absolutely so.
We also learned in the days leading up to the IPO that Semikin Doctor Company Arm and its majority backer SoftBank, had made an attempt to acquire Cereubris in the weeks before it's listing. Those offers were ultimately rebuffed, but there's huge competition in this market. There's obviously tons of interest in the private markets and now the public markets as well. In aichip companies an Area Infrastructure, Eclipse and Foundation Capital,
one of the other largest free risk backers. This sort of hardware, this infrastructure is very much their bread and butter, and so they've got, you know, some of the things in the pipeline sort of that follow this general theme. And certainly there are more investments being made all across stages in the private markets, very much in line with this theme. So expecting to see much more activity here.
Invoted for Rebecca Torrance, big week, thank you very much. From the excitement around Srebris's debut to the massive AI spending that's still ongoing at the hyperscale, is investors increasingly focused on whether companies like Alphabet, Meta and Amazon can justify the soaring capex with sustained revenue growth. On the other side, let's get more on that with Eric Sherid and Goldman sax Co Business Unit, leader of the Technology,
Media and Telecommunications Group in Global Investment Research. Some weeker, I mean that's the formula. You look at the hyperscalers, you get the capital expenditures number. We get to the end of earning season and video reports next Wednesday. Probably the largest beneficiary of that capital expendit show. If we're being honest, you sit here on a Friday morning, Eric, what's your conclusion of the week's news flow and how it impacts those bigger names that you cover well.
I think the main takeaways for us is that we remain in an infrastructure led cycle, so CAPEX continues to have an upward bias, albeit the bias to the upside was more muted this quarter than it was last quarter, which we think investors generally received positively. The second element would be that the revenue backlogs, or the future revenue that could come from this capex is now over nine hundred billion dollars combined, spread across both Alphabet and Amazon's
cloud computing divisions. That is giving investors increased confidence that there's revenue that will follow allbeit one, two, three years after the capex is spent. And interestingly, the margins in these cloud segments also surprised to the upside because non AI workloads are accelerating structuring data is accelerating, so therefore the long term view of earning a return on a
larger revenue base gave people more confidence. And that's why you've seen Amazon and the Alphabet over the last one three months act very very well as stocks as there's been a greater appreciation for that.
I mean, Eric at one point this week we wondered if Alphabet was going to eclipse in video as the world's most valuable company. And we think about that vertical integration that is just helping with this flywheel, the fact that they are able to have such prowess when it comes to TPU and the amount that they're able to begain inefficiencies. And I think if Amazon actually using cerebras share stock. More broadly, they've been using Cerebris hardware alongside
some of their own in house chips. How are you seeing this world of benefit from using your own self made chips alongside those from others.
Custom Silicon or TPUs from Alphabet and Amazon, we think continues to be one of the most underappreciated narratives in the market. Firstly, it continues to drive workloads into these cloud ecosystems so they capture more revenue overall. Secondarily, because they designed the custom silicon, they garner more of the margin by doing it. And the performance of TPUs, while not quite where GPUs are on an absolute performance basis, if you measure it on price to performance, they're actually
quite competitive. So this is something where you can go to your customers offer a price to performance ratio that looks very attractive and they benefit from garnering more revenue and more incremental margin. That's another theme that we think is gaining in prominence across this landscape.
Eric, you lead the team at that is covering Amazon and Alphabet right, and I'm just saying that's point out the obvious. That's your focus, but you must track Anthropic so closely. Both companies have significant financial interests in anthropic. Both have some kind of competition with Anthropic at the model level. In Amazon's case, Bedrock is the marketplace for Claude. That's very tangled as a web. How do you untangle it?
Well?
I think there's a lot without getting into any one company in their relationship with another. I think the world overall is becoming more interdependent. When it comes to AI, you're going to have foundational model companies that need compute, You're going to have hyper scalers that can deliver that compute.
You increasingly are going to have hyperscalers who are effective partners that allow the foundational model companies to come to market and connect with enterprises like Goldman Sachs, and there is going to be a lot of scale that benefits driving incremental growth in this landscape. The truest measure of technology computing shifts in my career has been that only a handful of companies on both the infrastructure and the
platform layer earn excess returns on capital. So there's only going to be a handful of companies that are enterprise platform companies. There's only going to be a handful of companies that are consumer platform companies. And if you come back to the earlier point that you led with, which is the capital need for this entire cycle, there's only a handful of companies that have the access to capital
to be able to build to this. So there is going to be an interdependence that comes from just a handful of companies that can actually build at this level of scale.
Alek Sheridan fascinating.
Not having you on the show has always come back soon of Goldman sachs.
Now coming up.
Tensions rise between Apple and Open Ai. Why the AI startup is weighing possible legal action against the iPhone maker.
This is Blomberg Tech.
And it's time now for talking tech and first up. Samsung management is making a rare eleventh hour visit to union leaders to a massive chip factory strike, but the world stop memory maker is facing an eighteen day walkout that could cost the company seven hundred million dollars a day and good stall critical AI chip production. Plus Bill Ackman's Pershing Square has taken a new course take in Microsoft, with Ackman arguing that the market is unestimating the tech chanswer resilience now.
The move comes as Microsoft.
Continues its aggressive push into AI, reclaiming its status in Agwin's portfolio and Elon Musk's XAI, and is officially entering the coding agent race with the launch of grock Build in an attempt to catch up with down public explaud Musk is racing to close that gap in the developer market. It's coding agents become the next multi billion dollar frontier in AI.
Ed Let's chat about Open AI. You sat down with CFO Sarah Fryer last night, and she you had the opportunity to basically say, good timing. Sam Altman's under away in a very big case against Elon Musk. And then there's the breaking news last night about Apple and the relationship flu part.
Just reflect on.
It, I mean an extraordinary not saying planfo.
Great timing though.
A joy a bounty of news when you're sitting down with a key executive. And Sarah fry is very clear about what she needs at the moment. She needs money for compute that continues. He has already got more than one hundred and twenty billion dollars of it, and she's got plenty of optionality.
We talked about her relationship with Sam Altman.
She's very What's so interesting is she's saying, Look, if you want your CEO and CFO to be, you know, always agreeing on absolutely.
Everything, you're sort of getting her wrong steer. That shouldn't be how it works. But they have a really good working relationship.
Look, she was just in California at his ranch on the weekend because they're working extra time on things like Compute.
Here is a close relationship.
But yes, at times they have to disagree, so that was an interesting discussion, particularly when we've had some heman's own sort of way in which he presents himself and the bill in the business being under the coals and under the limelight.
But I think was notable.
I asked her about the open Ai and Apple relationship, and of course she couldn't comment. But this is a company that depends on partnerships to get their technology into people's hands.
Apple is key for.
The CHATCHBT, but it was meant to be better integrated into the overall Apple Intelligence experience.
We're going to get the detail on that just a set.
We're going to hear more from Kara's conversation with Sarah Fryer later this hour. Stick around for that. Here's the details Apple and Open Area's relationship is framed. Open Ai is weighing possible legal action against the iPhone maker, arguing it hasn't seen the expected benefits from the partnership and it's Apple's use of its technology remains limited and hard for users. Define let's get out to Bloomberg Senior Tech
editor Dana Wolman incredibly detailed report from the team. I suppose what was that agreement initially?
So the two companies teamed up and really open Aiyes, saw this as an opportunity to, as you said, get its product as big of a name as it is in front of even more people Apple's huge user base. So open Air's technology would be built into Apple's platforms, and for a lot of users, especially users who are perhaps less tech savvy, it might have been their first
exposure to open AI's technology and open air. I was hoping that this would result in billions of dollars worth annually in new subscribers.
Let's talk about whether or not they have to be forced with some legal action to get recompense here, but more broadly, it's a time where Apple's own Apple Intelligence is not working as they hope, having to depend on Google to help build siry.
More broadly, are they just having to go to more players in the ecosystem?
I'm sorry, can you repeat the question? I don't know if I fully heard you.
How are we thinking about Apple's own dependence on other offerings for its own Apple Intelligence that isn't working?
So Apple is going to be opening up its own platforms to other developers as well, which surely is not helping with the dynamic with open Ai. It's something that Marker mentioned in his report. That said, and this has come up in other of Bloomberg Technology's reports as well, is that Apple is sort of benefiting from the investment that all of these other AI developers have invested over
the years in AI. Now it's getting to sort of integrate this menu of different increasingly advanced tools, as you said, as it continues to build out its own much delayed product of its own data.
Wollman, we so appreciate you, thank you for coming on regarding open AI and Apple.
Coming right up, we're discussing AI in another area of Figma.
It is flipping the script on the AI disruption narrative. CEO Dylan Fields joining us next after earnings were a blowout, much better than expected, and so much for that disruption vis Blue meg Tech. Welcome back to Bloomberg Tech, and we've got to focus in on the company Figma. It is defying those fears that AI would disrupt the design stack. The company has just reported a massive first quarter with revenue growth accelerating to forty six percent. It successfully begins
to monetize new AI features. Joining us now Figma founder CEO Dylan Field. Dylan, Look, there has been many many a question on software providers about the disruption that will come from the likes of large language model frontim makers, anthropic being one. What is driving your revenue growth and the ability to exceed guidance as well well?
First of all, thank you for having me and good to see you. And yeah, a strong quarter. We had. Revenue accelerates forty six percent year every year, and our net dour attention for customers that are over ten k of ARR is now at one hundred and thirty nine percent. And also strong cash flow with non up margin of sixteen percent in the quarter in free cash flow of twenty seven percent.
We also raise our guidance. So we're very glad with the results.
And I think in terms of taking a step back around what is behind the quarter and also the moment you're mentioning as AI commanitizers code, it makes it so that code is easier than ever to write, you know, the layer above code as that gets commoitized is fine.
I got an audience question for you on that in just a moment. But for me like that, I don't know what you would call it, Like the credit caps or the way you charge on usage was so fascinating. So in March you started charging customers a fee to use AI, in particular beyond a certain limit, and the response is varied. Right when people hit that cap, many
were willing to pay for more credits. There was some drop off though, like a small group was saying, well, if that's the case, I won't use a Figma product anymore. Where do you see that netting out?
Well, we have many products and you're always welcome to you know, as a free user use one of our free surfaces, or as a paid user of the seat, use our.
Traditional design tool.
But yes, if you want to use the product we have Figma make, we want to use AI features in Figma design. What we did was we essentially added some number of free credits to paid seats to make sure that people had a way to try these features out. And then we also made it to that if you want to buy additional credits you can, And for a long time we actually made it said everything was free. But you know that's not exactly the move that us do that forever and it does cost real money.
We also have.
Thima Weave, and weave is extremely exciting and with weave, what you can do is essentially create a workflow, which is a note through notebased editing tool where you connect up different outputs from models I think images, videos, three D models and more, and then you can push them through a workflow so that you're able to really mold those model outputs like clay. So for example, NBBJ, a architecture firm. One of the customers we mentioned Durens Call.
They used to do these very extensive customer shoots where.
They would go out to site and they would.
Really understand what is the different lagging at different times and they would then superimpose the three D model of the building. And with Thigma weave they can do that all in a workflow where they can just really easily control all sorts of different parameters and it saves them a ton of time and gets better results for the client. So that's another one where we also see AA friend SCIPT kick in.
You bring real anecdotal evidence to bear, and the anecdotical elevidence we hear is like like if you ever try and st wave figma from those that use it when in.
The workforce, they'll not leave.
Employees will like march out the building because they so love the product.
But how do you then fight this narrative?
Then investors just want to sell first, ask questions later, and has put your stock under pressure since the IPO.
I mean, I think we control the inputs and we need to deliver for our customers, as simple as that. And so we're working very hard always on making sure that we're doing the right thing for the long term. And I think that the long term is thankfully aligned with our strategy. We're in the best position we think we are basically can be. And you know, is I
becoming more important than ever? And I think that one area when it comes to maybe the market or the world, as we're seeing design go more broad in these companies and also be more appreciated as the way that you win, but also what you break through a very head of information landscape.
You have to really people what design is.
It's not just you know, creating something that you know you think is beautiful because you know, many people have different aesthetics. It's how it works, it's ux, it's forum, its function.
Yeah, and we have.
To I think, really help people understand the many different facets of design and Enfigma, design is not always just you know, how it looks, how it works. It's also the thinking process out there, and I think in this world, Dyaling, Yes, sorry, I don't.
Mean to cut you off.
We're running out of time, and I want to get that audience question to you because, as you know, I think it's really important regular of the show, Ben, how do you see inference and token costs affecting your margins going forward?
You kind of alluded to it earlier.
Yeah, we talked about in our Orange call yesterday and how if we see an opportunity to go really big and have a ton of growth, we will take it and we will push hard. But I think there's a short term, there's the long term, and sometimes there's ways to push hard on short term and that might create downward pressure and margins. But if you're going for a massive TAM in the long term, I think that's the right move in what our investors should be cheering us
on to do. And the TAM is very large both for design, for sculpting, you know, advertising marketing and breaking through noise, and I think that in general, if we're able to deliver on that and able to win this increasingly competitive landscape which is growing so fast of design when it's the new code. I think that puts us in an amazing edition for the future. So very excited
to bringing more people into the design processes. What we're seeing with our customers, it's not just designers, it's many others as well, but also level them up on design.
Figmacia didn't feel back on Bloomberg Tech.
Thank you very much for joining us.
Now coming up, Figure is putting their humanoid robots out the lab and onto the live stream. CEO Brett Adcock with us next. This is Bloomberg Tech. Figure says its humanoid robots just completed more than twenty four hours of continuous package sorting autonomously, a live stream watched by millions across YouTube and X three FO three robots worked in shifts, scanning, flipping, sorting packages at roughly human speed, all powered by its in house AI software running directly on board the robots.
But that demonstration did spark some skepticism was it real? Joining us now, as Brett Adcock founder CEO A Figure, that's where we start, you can say, Brett definitively. Over the twenty four hours or more, there was no teleoperation. A lot of people in the comments, as you know, pointed to the idea, and I think we had video a bit that the three on shift kept gesturing to the head, which is a tailtale sign in robotics of teleoperation.
Your pledge that there was none.
There's absolutely no telly operation into this. The robots are all operating fully autonomously using an onboard neural network redesign called Helix two. Sometimes when the robot takes a turn to left to grab packages, it moves its left hand out of the way upwards. You'll see this behavior happen every single time the robot turns for packages.
But we've been.
Running autonomously now for close to fifty hours, the robots operating shifts. There's been basically almost no downtime on the belt. We've pushed over close to sixty thousand packages and we're just going to keep going now and see how far this can go.
So this was live streams, right, and that was one reason I really wanted you to come on the program, because there's the bit people don't see what's happening behind the scenes, like in shift changes, where does the robot go?
Does it need maintenance?
For the most part, the robots operate on a four hour battery life. After the battery is low, the robot messages another robot to come out to take its place. The robots didn't do a swap. The robot has just left. The conveyor system is going to go charge wirelessly, understand while the other robot continues to do work. If there are issues, say we have hardware software issues, the robots can basically walk off into maintenance and call another robot
to take its place. The goal is to be able to enlistit twenty four to seven operations with basically no like no failures on the on the use case itself, which we haven't had today. So the robots basically the conveyor system has been running twenty four to seven since like middle of this week. I think we're now we're approaching fifty hours of just full like every single hour since since we've launched, the robots have been basically I've been doing work now on this line, which.
I think is like wow, which I think is crazy.
You know, Figure wants to build like you know, we want to build like I robot. You know, we want robots everywhere in the world in the commercial market, Like this is like the first large step to doing that.
Okay, So what's harder now robots getting faster or making them even more reliable, because at the moment you seem to be doing both.
The robot that you're seeing here is roughly operating around human speech just about three seconds of package. That's the requirement to operate on this logistics line. So we're at like we're at human parity and speed. The goal is also to have like ninety percent success rate on the bark the package flips for barcode scanning. We're in that as well. In that requirement, the robots are also like
getting extremely reliable. Like part of this whole process of running this twenty four to seven with no downtime is to show how reliable humanoid robots are. And four years ago when I started the company, humanoid robots were falling. They were extremely unreliable systems. We've designed the systems and engineered this now to a point where the robots are I think extremely reliable. I think we're showing that now
in the live stream to the entire world. The big focus for us is is like how do we solve for a truly general purpose machine, and then how do we manufacture unprecedented volume similar to cell phones? Today, so like, so this week back que our manufacturing facility will manufacture anywhere between like sixty and seventy humanoid robots just this week, right, and we do it right next door on the Figure campus.
So brat, I want to get into the idea, this is full stack. I've been hearing a lot growing speculation that open ai could get back into robotics. And you have a history white Way, you had a partnership, you decided on the software side you could do better yourself.
You know, what do you make of that?
That idea that ultimately a big party is going to want to own the entire stack that powers the humanoid robot.
To really do this right, like if we want to really build like I robot like the movie, yes, you know we basically you have to design the entire hardware system almost yourself, like motors like stater rotors, the electromagnetics work, like, you have to do all the battery systems work. You have to do all the actuator design, sensor design, kinematics like structures like which we do all in house now
here at Figure. We also manufacture the robots, so we're like and then we also test them and we do all the AI data collection and all the AI and neural net training ourselves here in house. So basically, this is a full end to end vertically integrated system that we now have out doing real use case work that humans do, and we can do this at human speeds. And we're doing this now for like you know this example, most of these shifts run like this eight hours a day.
We're doing this like twenty four to seven, just to show how reliable the systems are and how like mission ready these things are to get out at scale. So anyway to solve this you have to have like a truly vertically integrated approach from top to bottom.
What about the bottlenecks? View is in our money do you want to go public?
Brat?
What is?
What are you needing to get this out?
More broadly, our largest two bottlenecks are like data for pre training, our helix, neural net and UH in manufacturing, we're spinning up manufacturing here our manufacturing facilities called BATQ. We've we're now doing we're now out of several like thousands of run rate annually production that is continuing to scale up.
UH.
You know, I think UH on the data side, we're collecting and training, uh, kind of unprecedented models for like our AI stack here internally that we've ever done.
Uh.
So I think we're like we're and we have you know, we have well over a billion dollars of cash in the balance sheet today. So I think from a from a financial perspective, we're in a good spot. We're manufacturing at pretty much unprecedented volumes for ourselves, and we're we're building like next generation AI models that I think are uh,
to be honest, are just completely mind blowing. And so the goal is like the goals to solve the data problem and the manufacturing problem, to get humiliar robots out of scale.
And maybe come up with even more names for these robots currently getting them from online fans. I think one's called Bob and Gary Brett Gadcock. We so appreciate your time today, Founder and CEO of Figure.
Now let's send our attention to Annie Jasse. It was once Jeff Bezos's deputy.
Remember, now five years into his tenure as CEO, Jesse is steering the company through some of.
Its greatest changes. That's the focus of today's big take most. Matt Day joins us.
With more You went to see sort of the new real focus point for Annie Jasse. It's data centers, and this is a company that is now just scaling so much in a vertically integrated manner. Matt, what did you learn from visiting the hubs and what it says about Andy?
But it says that there's just such sprawl to Amazon these days. You think you know Amazon from the package is showing up at your doorstep. So much of what they're spending money on it's data centers. It's a supply chain behind data centers. That means chips, that means hardware engineering, that means software at the large language model level.
They're really just all over the place.
And it's kind of an unfathomable sort of things they got going on, and really the sort of single organizational principle behind it is Andy at the top of it. He's making all these calls on where they got to shift money with the comple and stuff out of what they're putting into. It's really really an impressive machine.
So Matt, knowing that he's watching right now and he Jesse is glued to Bloomberg Tech.
There is a good chance that's true.
What do we learn about him is the Amazon CEO versus Jeff Like you and I've talked about this in the past, But what's different about him? Does he lean hard into the AWS thing?
You definitely did.
And if you look at where their bets are today, you know a lot of them are, and what can Amazon do uniquely and what can they do differently when it comes to their core retail business. So they've left some opportunities on the table there. That's not the case in AI. They want to be everywhere in AI. They want to sprinkle AI through all of their product lines, and that is where they're spending a ton of their their capex right now. So he's definitely brought some of
that AWS pedigree with him. We've also just learned he is he is an extreme detail guy. It's not that that was not a Jeff Bezos trade, but for the last few years of his leadership at Amazon, Jeff was checked out and some of the business. You know, Andy hasn't let go of anything since he showed up. He's all over the details of this business.
Blue Most Matt Day with the big take on Andy jase Era at AWS and Amazon must coming up arguments in the trial between Open AI and Elon Musk wrapped up yesterday. We're going to get to the latest and some more of the conversation we've been having with open Ai the past twenty four hours. This is Bloomberg Tech.
Opening Eyes chief financial Officer Sir Fran talked about the demand the company is seeing and the compute needs it's up against. She sat down with me at the Cancello Spark summit in Napa.
Take a listen.
Compute itself has clearly been a bottleneck. I don't think any of us, even Sam and Greg, who I think we're incredibly prescient overall, in terms of the need for compute, I don't think they could have foreseen just the sheer demand that we have right now in twenty twenty six, and so that it's the energy that sits behind it.
We're seeing memory out of Southeast Asia where there's all of these chok points in the supply chain, and this is why it behooves us to get ahead of that, strike those partnerships earlier when people are maybe not seeing just the sheer scale.
Let's talk about partnership and about some of the people you just talked about, Quite Brockman and Sam Altman.
They've had a tough week.
In terms of being in trial, having their relationships sort of cast.
Into the limelight. What is this like working with Sam?
There's been some questioning as to how you work alongside him, Do you work well?
Is it a strong bond? Is he trustworthy?
Is it?
Seemed to have been grilled by certain lawyers of Elon Musk this week.
So I think mister Musk is very much out to just distract, and we are just staying the course, like building our technology, putting it into customers' hands, and really creating agi for the benefit of humanity. Working with Sam is great. We get on incredibly well. I think we have a wonderful partnership, super good ying and Yang Sam push us.
He's super curious.
My job is to take that curiosity and create optionality. So just as we talked about whether it's making sure we have enough compute, making sure we have the funds to do it, makeing sure we build the business strongly. And I spend an an ordinate amount of time with customers, as does Sam, and we get to compare a lot of notes on that front, as well as sometimes our nuts on parenting. So it goes the whole gamut frankly Open.
AI CFO Sarah Fry speaking with Caroline about her relationship with the company's CEO, Sam Altman. The relationships and choices of open ais founders have been front and center of the lawsuit brought by Elon Musk over the charitable status of the company. Arguments in the trial wrapped up yesterday. Bloombergs Madelin Meckelberg has been there every step of the way, and so this is where we're at. Closing arguments we
wait for. I think jury deliberation Monday give us everything that we need to know at this point.
That's right. I think one thing that we learned from closing arguments, it's really relevant to the conversation you just had with Sarah, is that this case, for both parties, really boils down to credibility. That's the message that they left jurors with on Thursday when they had closing arguments. Musk's attorneys are trying to cast Sam Altman as someone
who's deceptive, who's untrustworthy. Meanwhile, lawyers for Open Ai were trying to point to the evidence that's been presented so far in the case, saying Musk's story doesn't add up with what we've seen in the documents, what you've heard from witnesses, and so now it's up to the jury to decide whose version of events they believe about this saga of the formation of open Ai, the dissolution of this relationship and the transition into rivalry between Musk and
Altman that we see playing out today.
What happens if Musk wins.
That's a big question. So at this point, the jury is going to start deliberating and they're going to issue a verdict that's advisory, so it's a non binding recommendation to the judge in this case, who's going to have the final say on whether or not Musk has enough to prove his claims here. And Musk is asking for a lot. He's asking for one hundred and thirty four billion dollars in d images that he says he's going
to donate to the Open Ai Foundation. He's asking for Altman and Brockman to be removed from their jobs, and he's asking for the open Ai come up company that's been formed to be transitioned back to a nonprofit. These are huge asks. The judge is also going to hear arguments on Wednesday about which one of those might be realistic, and we're just going to kind of have to wait and see how she comes down on the main question
before we get to that. But the stakes are huge and ex essential essentially for open.
Ai b mags Mann, Mackelberg, fascinating.
Thank you.
I mean, we wait with bated breath for the rest of the week and what indeed happens in that court case.
But he'say, there's some.
Confidence world is part about it is that court case happened in the context of the week, all the other news flowing. It's almost kind of I don't want to say buried, but we do wait for it. It for this sedestion of Bloomberg Tech on Monday, speaking with the CEOs of Dell and the video two thirty pm Eastern eleven thirty m Pacific. If you do not want to miss it, check out the pod for the recap. What a week this is Bloomberg Tech