WEBVTT - Meta Doubles Down on AI Spending, Plans New Class C Shares

0:00:02.529 --> 0:00:05.370
<v Speaker 1>Bloomberg Audio Studios. Podcasts.

0:00:05.730 --> 0:00:06.250
<v Speaker 2>Radio.

0:00:06.710 --> 0:00:07.110
<v Speaker 3>News.

0:00:12.990 --> 0:00:16.520
<v Speaker 1>Bloomberg Tech is live from the heart of Silicon Valley

0:00:17.050 --> 0:00:23.280
<v Speaker 1>with Ed Ludlow in San Francisco. This is Bloomberg Tech

0:00:23.300 --> 0:00:26.380
<v Speaker 1>coming up. Meta doubles down on AI, boosting its spending

0:00:26.460 --> 0:00:29.980
<v Speaker 1>outlook as Zuckerberg says its massive investments are starting to

0:00:30.020 --> 0:00:32.729
<v Speaker 1>pay off. Plus, Tesla beats on profit, but a big

0:00:32.770 --> 0:00:36.830
<v Speaker 1>cut to its spending plans raises new questions about Musk's AI,

0:00:36.909 --> 0:00:41.570
<v Speaker 1>RoboTaxi and robotics ambitions. And AI reshapes cybersecurity for both

0:00:41.590 --> 0:00:45.409
<v Speaker 1>defenders and hackers. We speak with OpenAI President Greg Brockman

0:00:45.710 --> 0:00:48.530
<v Speaker 1>and Cloudflare CEO Matthew Prince. Let's get right to our

0:00:48.550 --> 0:00:51.330
<v Speaker 1>top story, and that is Meta. Shares down almost 7%

0:00:51.330 --> 0:00:53.870
<v Speaker 1>in the session, on track for their biggest drop since July.

0:00:53.890 --> 0:00:56.240
<v Speaker 1>I show you the two-day chart because going into earnings,

0:00:56.260 --> 0:00:58.960
<v Speaker 1>there was a lot of buoyancy. It's dragging down technology

0:00:59.000 --> 0:01:01.360
<v Speaker 1>more broadly. A lot of the chip names also a

0:01:01.400 --> 0:01:04.120
<v Speaker 1>big drag at the index level. Meta is creating a

0:01:04.160 --> 0:01:07.640
<v Speaker 1>new class of non-voting shares. And part of the reason

0:01:07.780 --> 0:01:11.000
<v Speaker 1>is Mark Zuckerberg. Mehta says the CEO was considering leaving

0:01:11.060 --> 0:01:14.840
<v Speaker 1>California over a proposed wealth tax that could leave him

0:01:14.910 --> 0:01:17.230
<v Speaker 1>with a multibillion dollar bill. The new class he shares

0:01:17.510 --> 0:01:20.690
<v Speaker 1>gives Zuckerberg stock he can sell for cash without immediately

0:01:20.750 --> 0:01:24.310
<v Speaker 1>reducing his voting control. In return, he's agreed to stay

0:01:24.350 --> 0:01:28.110
<v Speaker 1>in California or wherever Mehta is headquartered for eight years.

0:01:28.470 --> 0:01:30.830
<v Speaker 1>as long as he remains in a full-time leadership role.

0:01:30.870 --> 0:01:33.890
<v Speaker 1>That was one story. Let's dive deeper into what else

0:01:33.940 --> 0:01:37.179
<v Speaker 1>we learned from Meta's earnings report, which was yesterday. Bloomberg's

0:01:37.240 --> 0:01:39.610
<v Speaker 1>Alex Levine is with us. Let's be completely frank about it.

0:01:39.640 --> 0:01:44.910
<v Speaker 1>The Class C shares story was communicated in a difficult-to-understand way,

0:01:45.310 --> 0:01:47.590
<v Speaker 1>partly on the call, partly in the earnings statement, but

0:01:47.610 --> 0:01:49.590
<v Speaker 1>really in a filing. And then there were lots of

0:01:49.650 --> 0:01:51.990
<v Speaker 1>numbers from earnings. We'll start there. What do we actually

0:01:52.030 --> 0:01:53.710
<v Speaker 1>learn about how it's going for Meta right now?

0:01:54.380 --> 0:01:57.840
<v Speaker 4>Well, we know that Meta delivered really strong third quarter results,

0:01:58.060 --> 0:02:02.180
<v Speaker 4>really strong fourth quarter guidance. They beat analyst estimates on those.

0:02:02.470 --> 0:02:07.770
<v Speaker 4>And I think that the decision to issue this non-voting

0:02:07.810 --> 0:02:11.429
<v Speaker 4>stock and Zuckerberg's decision to stay in Silicon Valley to

0:02:11.490 --> 0:02:14.480
<v Speaker 4>sort of see everything through is a really important statement

0:02:14.700 --> 0:02:18.600
<v Speaker 4>as Meta enters this next chapter. It continues to accelerate

0:02:18.620 --> 0:02:20.620
<v Speaker 4>into this next chapter. And I think it also just

0:02:21.220 --> 0:02:24.210
<v Speaker 4>goes to signal that it is confident that its AI

0:02:24.250 --> 0:02:26.770
<v Speaker 4>bets are paying off because some of the decision to

0:02:26.810 --> 0:02:29.530
<v Speaker 4>do this with the shares was, of course, also, as

0:02:29.570 --> 0:02:33.130
<v Speaker 4>the company had said, to help cover some of this

0:02:33.230 --> 0:02:34.590
<v Speaker 4>enormous AI spending down the line.

0:02:34.940 --> 0:02:38.100
<v Speaker 1>Yes, Meta's CFO Susan Lee was asked on the analyst

0:02:38.160 --> 0:02:40.920
<v Speaker 1>call about the class she shares, and she said it

0:02:40.980 --> 0:02:45.820
<v Speaker 1>is a way of diversifying and funding and financing the

0:02:45.860 --> 0:02:48.260
<v Speaker 1>AI spend that's to come. But she didn't mention the

0:02:48.300 --> 0:02:51.780
<v Speaker 1>Zuckerberg part of it at all. CapEx is narrowed for

0:02:51.800 --> 0:02:55.459
<v Speaker 1>this year, 135 billion at the bottom end to 145

0:02:55.460 --> 0:02:58.620
<v Speaker 1>billion at the top end, which we knew. There seems

0:02:58.660 --> 0:03:00.820
<v Speaker 1>to be like some signals on the direction of travel

0:03:00.930 --> 0:03:05.260
<v Speaker 1>for Meta. AI seems to be also really boosting the

0:03:05.320 --> 0:03:07.530
<v Speaker 1>core ad business. What did they say? What were the

0:03:07.560 --> 0:03:09.630
<v Speaker 1>numbers that they'd point to to evidence that?

0:03:10.550 --> 0:03:15.070
<v Speaker 4>Well, just in the last month, Meta launched these two

0:03:15.150 --> 0:03:18.230
<v Speaker 4>new products. You had its personal AI agent, Muse, that

0:03:18.250 --> 0:03:20.660
<v Speaker 4>was debuted in September. And then you had Muse Wonder,

0:03:20.750 --> 0:03:23.960
<v Speaker 4>its latest AI model that debuted earlier this month. And

0:03:24.180 --> 0:03:27.400
<v Speaker 4>what Meta reported, it was really strong early adoption, strong

0:03:27.500 --> 0:03:30.440
<v Speaker 4>early traction in all of this. And I think that

0:03:30.560 --> 0:03:33.290
<v Speaker 4>even though investors do still seem skeptical, as you said,

0:03:33.389 --> 0:03:38.070
<v Speaker 4>shares fell this morning, even as Zuckerberg and the company

0:03:38.110 --> 0:03:40.210
<v Speaker 4>projected all this confidence, I think that they are really

0:03:40.290 --> 0:03:43.340
<v Speaker 4>pointing to the early traction of these products, and also

0:03:43.680 --> 0:03:47.340
<v Speaker 4>the huge number of businesses that are already using the

0:03:47.400 --> 0:03:51.780
<v Speaker 4>other platforms across their family of apps. Instagram, WhatsApp, Facebook,

0:03:51.810 --> 0:03:53.590
<v Speaker 4>I think the number that they said was more than

0:03:53.590 --> 0:03:56.690
<v Speaker 4>500 million businesses. They're really pointing to these numbers as

0:03:56.770 --> 0:04:01.190
<v Speaker 4>evidence that news adoption will continue to scale and that

0:04:01.230 --> 0:04:03.410
<v Speaker 4>it can actually be monetized, which remains the big outstanding

0:04:03.430 --> 0:04:05.150
<v Speaker 4>question here is how are they actually going to really

0:04:05.230 --> 0:04:06.250
<v Speaker 4>make money off of all of this?

0:04:07.220 --> 0:04:09.600
<v Speaker 1>The shares are down 6.5% on track for the biggest

0:04:09.620 --> 0:04:11.720
<v Speaker 1>drop since July. But what was weird is when all

0:04:11.750 --> 0:04:13.670
<v Speaker 1>of this news came out last night, the stock really

0:04:13.710 --> 0:04:16.969
<v Speaker 1>went up and things really changed overnight. Bloomberg's Alexander Levine,

0:04:17.010 --> 0:04:19.270
<v Speaker 1>thank you so much for being with us. During Meta's

0:04:19.290 --> 0:04:22.409
<v Speaker 1>earnings call, Mark Zuckerberg touted the progress of its AI

0:04:22.510 --> 0:04:23.450
<v Speaker 1>agent tool, Muse.

0:04:24.830 --> 0:04:27.690
<v Speaker 2>We are building out the capacity to serve more than

0:04:27.710 --> 0:04:30.780
<v Speaker 2>a billion people over the next few years. In the

0:04:30.820 --> 0:04:34.960
<v Speaker 2>near term, Muse is very compute intensive to run. But

0:04:35.000 --> 0:04:38.000
<v Speaker 2>we expect to improve its cost efficiency by more than

0:04:38.360 --> 0:04:40.480
<v Speaker 2>10x in the next year, which will enable it to

0:04:40.520 --> 0:04:42.690
<v Speaker 2>be profitable once the revenue lines are at scale.

0:04:44.470 --> 0:04:47.150
<v Speaker 1>Let's get more Meta's results with Minda Smiley, Senior Analyst

0:04:47.250 --> 0:04:49.030
<v Speaker 1>at eMarketer. And I want to start with Muse. What

0:04:49.050 --> 0:04:51.510
<v Speaker 1>they also said is that more than 10 million people

0:04:51.550 --> 0:04:54.320
<v Speaker 1>signed up for Muse in the U.S. and Canada in

0:04:54.360 --> 0:04:57.140
<v Speaker 1>its first six weeks. How do you interpret that data?

0:04:58.430 --> 0:05:01.310
<v Speaker 5>Yeah, I mean, that's certainly a good number in terms

0:05:01.330 --> 0:05:03.710
<v Speaker 5>of early adoption. But I do think when you think

0:05:03.750 --> 0:05:06.469
<v Speaker 5>about the fact that a lot of these downloads and

0:05:06.630 --> 0:05:08.870
<v Speaker 5>usage are coming from the fact that Meta has been

0:05:08.890 --> 0:05:14.180
<v Speaker 5>able to promote Muse through its own channels, Facebook, WhatsApp, Instagram,

0:05:14.240 --> 0:05:16.440
<v Speaker 5>quite heavily. And they've also just invested a lot of

0:05:16.520 --> 0:05:19.200
<v Speaker 5>money into marketing Muse as well. So I think it

0:05:19.240 --> 0:05:22.480
<v Speaker 5>kind of waters down those numbers in a way because

0:05:22.720 --> 0:05:24.539
<v Speaker 5>I think it points to the fact that a lot

0:05:24.560 --> 0:05:28.500
<v Speaker 5>of this interest is isn't necessarily organic isn't necessarily genuine

0:05:28.560 --> 0:05:30.190
<v Speaker 5>it's really just coming from the fact that people are

0:05:30.320 --> 0:05:33.250
<v Speaker 5>maybe might see something about muse and instagram click on

0:05:33.290 --> 0:05:35.570
<v Speaker 5>it download it mess around with it and then maybe

0:05:35.589 --> 0:05:37.410
<v Speaker 5>a week later they they delete.

0:05:37.210 --> 0:05:40.039
<v Speaker 1>It from their phone I've been using Muse as much

0:05:40.080 --> 0:05:42.740
<v Speaker 1>as I can, like as somebody interested in technology. If

0:05:42.760 --> 0:05:45.080
<v Speaker 1>you host a technology news show, you should probably use

0:05:45.120 --> 0:05:47.660
<v Speaker 1>the technology, right? But I'm still kind of not really

0:05:47.800 --> 0:05:52.380
<v Speaker 1>using it beyond helping put social posts together, more deep search.

0:05:52.420 --> 0:05:54.700
<v Speaker 1>Maybe it's helping me with my calendar. So I found

0:05:54.740 --> 0:05:57.280
<v Speaker 1>it interesting that Mark Zuckerberg said that they expect most

0:05:57.320 --> 0:06:01.320
<v Speaker 1>people to use the free version. The monetization comes elsewhere.

0:06:01.339 --> 0:06:03.960
<v Speaker 1>Is there any evidence that they've got going on the elsewhere?

0:06:05.250 --> 0:06:06.130
<v Speaker 6>Yeah, well, it's interesting.

0:06:06.190 --> 0:06:08.190
<v Speaker 5>I mean, they talk a lot about how they hope

0:06:08.270 --> 0:06:12.659
<v Speaker 5>to maybe collect fees from Muse as they help with transactions.

0:06:12.700 --> 0:06:14.600
<v Speaker 5>One of the stats they pointed out yesterday was that

0:06:14.990 --> 0:06:16.650
<v Speaker 5>about a quarter of users, or I think they said

0:06:16.670 --> 0:06:20.010
<v Speaker 5>the main use case, at this moment is shopping, which

0:06:20.070 --> 0:06:23.190
<v Speaker 5>of course, you know, as a company that's invested quite

0:06:23.230 --> 0:06:26.320
<v Speaker 5>a lot of money into social commerce, it makes sense

0:06:26.360 --> 0:06:28.080
<v Speaker 5>that they would want Muse to also.

0:06:27.860 --> 0:06:29.159
<v Speaker 6>Be a commerce tool.

0:06:30.000 --> 0:06:32.940
<v Speaker 5>So I think when it comes to monetization, there are

0:06:32.960 --> 0:06:36.300
<v Speaker 5>a lot of opportunities there when you think about the fees,

0:06:36.339 --> 0:06:41.020
<v Speaker 5>when you think about subscriptions, even extending Muse to small businesses.

0:06:41.740 --> 0:06:42.440
<v Speaker 6>But I think a lot of.

0:06:42.440 --> 0:06:45.850
<v Speaker 5>That still remains quite theoretical, right? I mean, there's no

0:06:45.910 --> 0:06:48.409
<v Speaker 5>guarantee that this will take off It has to really

0:06:48.450 --> 0:06:52.380
<v Speaker 5>gain traction with consumers, with businesses before those monetization lines

0:06:52.820 --> 0:06:53.910
<v Speaker 5>can really go into effect.

0:06:54.900 --> 0:06:58.500
<v Speaker 1>It was still the core ads business that drove the beat, right?

0:06:58.520 --> 0:07:00.810
<v Speaker 1>I get the stock is down significantly, but ad revenues

0:07:00.850 --> 0:07:05.370
<v Speaker 1>were more than $ 63 billion, the street expected below 62.

0:07:05.370 --> 0:07:08.570
<v Speaker 1>And the interesting point is price per ad growing at

0:07:08.830 --> 0:07:11.690
<v Speaker 1>a rate of about 14% was also beyond expectations. It

0:07:11.730 --> 0:07:14.530
<v Speaker 1>seems like the AI story really is that they're able

0:07:14.570 --> 0:07:19.280
<v Speaker 1>to drive higher pricing on ads rather than look at volumes. Yeah.

0:07:19.600 --> 0:07:22.360
<v Speaker 5>And it's funny because I think because there's been so

0:07:22.400 --> 0:07:24.570
<v Speaker 5>much focus on, well, where are they going with AI?

0:07:24.630 --> 0:07:27.470
<v Speaker 5>What are their ambitions? Everyone wants to hear about Muse.

0:07:27.510 --> 0:07:29.430
<v Speaker 5>And it makes sense, right? They're spending a lot of money.

0:07:29.550 --> 0:07:31.250
<v Speaker 5>We want to see where they're going with all of this.

0:07:31.610 --> 0:07:34.690
<v Speaker 5>But the reality is that AI is really fueling their

0:07:34.710 --> 0:07:38.080
<v Speaker 5>advertising business, which we all know is the core, the

0:07:38.100 --> 0:07:41.820
<v Speaker 5>bulk of their business model. And it's doing so quite effectively.

0:07:41.840 --> 0:07:44.620
<v Speaker 5>That's why we continue to see this double-digit growth. quarter

0:07:44.680 --> 0:07:47.230
<v Speaker 5>after quarter and they keep refining it, keep building on it.

0:07:47.430 --> 0:07:50.190
<v Speaker 5>And it's helping not only with advertising, but also usage

0:07:50.210 --> 0:07:51.040
<v Speaker 5>and engagement as well.

0:07:52.520 --> 0:07:53.840
<v Speaker 1>Minda, I'm not going to ask you to come in

0:07:53.900 --> 0:07:56.080
<v Speaker 1>here on the mechanics of them issuing a new class

0:07:56.120 --> 0:07:59.780
<v Speaker 1>of stock. The whole point is that it's structured so

0:07:59.800 --> 0:08:02.870
<v Speaker 1>that Mark Zuckerberg can sell 20 percent straight away and

0:08:03.200 --> 0:08:05.550
<v Speaker 1>cover his tax bill if that were to arise. And

0:08:05.550 --> 0:08:08.710
<v Speaker 1>80 percent is staggered over eight years. But that's the

0:08:08.750 --> 0:08:11.330
<v Speaker 1>bit I want to talk to you about. Mark Zuckerberg

0:08:11.370 --> 0:08:14.470
<v Speaker 1>has agreed to stay and enroll basically for eight years

0:08:14.790 --> 0:08:19.170
<v Speaker 1>wherever the company's headquartered, California or otherwise. How important is that?

0:08:19.350 --> 0:08:23.950
<v Speaker 1>You know, the founder, CEO continuing to lead efforts like AI. Yeah,

0:08:24.080 --> 0:08:24.760
<v Speaker 1>it is important.

0:08:24.820 --> 0:08:26.920
<v Speaker 5>And it's a question I've been getting recently. You know,

0:08:27.280 --> 0:08:29.480
<v Speaker 5>what does this all say about Zuckerberg and his.

0:08:29.400 --> 0:08:30.400
<v Speaker 6>Commitment to the company?

0:08:30.420 --> 0:08:33.099
<v Speaker 5>If when will he step down? And I think it

0:08:33.140 --> 0:08:37.380
<v Speaker 5>just goes to show that he is committed to this vision. And,

0:08:37.390 --> 0:08:39.260
<v Speaker 5>you know, to what extent it will play out, of course,

0:08:39.320 --> 0:08:42.939
<v Speaker 5>is a different story. Obviously, we see that we saw

0:08:43.020 --> 0:08:45.280
<v Speaker 5>what happened with the metaverse. We see that he's going

0:08:45.660 --> 0:08:48.340
<v Speaker 5>heavily into smart glasses as well, which doesn't get as

0:08:48.540 --> 0:08:52.160
<v Speaker 5>hasn't gotten as much coverage lately with the launch of Muse.

0:08:52.240 --> 0:08:55.090
<v Speaker 5>But I think it goes to show that he doesn't

0:08:55.110 --> 0:08:56.770
<v Speaker 5>plan on going anywhere anytime soon.

0:08:58.000 --> 0:09:00.240
<v Speaker 1>Minda Smiley, Senior Analyst at eMarketer. It's been great to

0:09:00.260 --> 0:09:02.000
<v Speaker 1>have you back on the program. Thank you very much.

0:09:02.420 --> 0:09:05.360
<v Speaker 1>Coming up, the other earning story, which was Tesla beats

0:09:05.420 --> 0:09:08.420
<v Speaker 1>on profit, but cuts billions from its spending plans. What

0:09:08.470 --> 0:09:13.410
<v Speaker 1>that means for Musk's AI, robo-taxi, and robotics ambitions. We're

0:09:13.429 --> 0:09:20.480
<v Speaker 1>going to look at Tesla next. This is Bloomberg Tech. Okay,

0:09:20.500 --> 0:09:23.380
<v Speaker 1>shares of Tesla down, but way off session lows. They

0:09:23.400 --> 0:09:25.420
<v Speaker 1>were down more than 3% a little earlier. The company

0:09:25.500 --> 0:09:29.560
<v Speaker 1>beat on profit, but it's the spending outlook that's getting investors' attention.

0:09:29.620 --> 0:09:33.020
<v Speaker 1>It is cutting its CapEx plan by more than $ 5 billion,

0:09:33.080 --> 0:09:36.860
<v Speaker 1>raising questions about the pace of its AI, Robotaxi, and

0:09:36.920 --> 0:09:40.380
<v Speaker 1>Optimus Ambitions. Joining us now is Ben Harwood, Newstreet Research

0:09:40.400 --> 0:09:43.619
<v Speaker 1>Technology and Space Infrastructure Senior Analyst. We'll get to the

0:09:43.679 --> 0:09:47.180
<v Speaker 1>SpaceX-Tesla merger idea of it all. I mean, yeah, it's

0:09:47.230 --> 0:09:52.150
<v Speaker 1>really interesting. capex is being trimmed maybe that's just like

0:09:52.190 --> 0:09:54.890
<v Speaker 1>the reality of how quick they didn't deploy does it

0:09:54.950 --> 0:09:56.370
<v Speaker 1>signal something bigger to you.

0:09:57.670 --> 0:10:00.920
<v Speaker 7>No Hi, thanks for having me on. And the client

0:10:00.960 --> 0:10:03.400
<v Speaker 7>CapEx I see is just mainly a timing issue. And

0:10:03.520 --> 0:10:06.830
<v Speaker 7>I expect them to ramp up CapEx, you know, coming

0:10:06.850 --> 0:10:10.930
<v Speaker 7>into 27 and into 2028. And you mentioned what they're

0:10:10.950 --> 0:10:14.090
<v Speaker 7>doing on the AI front with FSD, with RoboTaxi. Actually,

0:10:14.130 --> 0:10:17.429
<v Speaker 7>everything we saw from that perspective was very positive. So

0:10:18.170 --> 0:10:21.219
<v Speaker 7>What we saw is that FSD adoption reached again an

0:10:21.300 --> 0:10:25.260
<v Speaker 7>all-time high. 75% of vehicles in the US are now

0:10:25.300 --> 0:10:29.140
<v Speaker 7>being sold with FSD. And that's really showing that the

0:10:29.210 --> 0:10:31.730
<v Speaker 7>new technology that you have with FSD is driving demand.

0:10:31.770 --> 0:10:36.250
<v Speaker 7>And RoboTaxi is the fastest scaling RoboTaxi business out there.

0:10:36.290 --> 0:10:38.590
<v Speaker 1>Ben, let me jump in. Because the audience knows I

0:10:38.670 --> 0:10:42.000
<v Speaker 1>use FSD every single day for my 60-mile round-trip commute.

0:10:42.140 --> 0:10:45.280
<v Speaker 1>I use it. I pay for it myself. But a

0:10:45.340 --> 0:10:47.420
<v Speaker 1>new Tesla, and I know a lot of the numbers

0:10:47.460 --> 0:10:51.150
<v Speaker 1>that quarter were existing Tesla owners upgrading. But if you

0:10:51.170 --> 0:10:54.170
<v Speaker 1>give Tesla somebody FSD for free, they might use it

0:10:54.210 --> 0:10:57.010
<v Speaker 1>for a while. The question is, do they stick with it?

0:10:57.640 --> 0:10:57.920
<v Speaker 6>Right.

0:10:58.059 --> 0:11:01.420
<v Speaker 1>The take rate, the actual long term usage. Do you

0:11:01.440 --> 0:11:03.790
<v Speaker 1>have a good answer on that? Yeah, exactly.

0:11:03.809 --> 0:11:06.530
<v Speaker 7>So that's a good question. So every Tesla owner in

0:11:06.550 --> 0:11:10.490
<v Speaker 7>the US gets a month for free. What the CFO

0:11:10.510 --> 0:11:12.959
<v Speaker 7>clarified on the call yesterday when we asked the question

0:11:13.000 --> 0:11:18.300
<v Speaker 7>is actually 75% of people are buying the technology even

0:11:18.380 --> 0:11:20.439
<v Speaker 7>when they get it for free. So they'll take it

0:11:20.500 --> 0:11:23.720
<v Speaker 7>in that first month. So it's very clear that there

0:11:23.840 --> 0:11:26.990
<v Speaker 7>is a lot of demand for this product. Word is

0:11:27.020 --> 0:11:31.069
<v Speaker 7>getting out, like I said, 75% of new users, new

0:11:31.090 --> 0:11:35.490
<v Speaker 7>vehicles are being sold with FSD. That's up 20 points sequentially.

0:11:35.510 --> 0:11:38.950
<v Speaker 7>And that's going to be up 40 points year on year. So,

0:11:39.910 --> 0:11:42.370
<v Speaker 7>performance of FSD, as you would have noted, Ed, has

0:11:42.429 --> 0:11:46.330
<v Speaker 7>improved over the past 12 months. And that's really spreading

0:11:46.350 --> 0:11:49.130
<v Speaker 7>the word. And it's now driving vehicle demand in the U.S.

0:11:49.510 --> 0:11:52.290
<v Speaker 7>You know, it's only a U.S. phenomenon now. Hopefully, in

0:11:52.309 --> 0:11:54.250
<v Speaker 7>the first half of next year, that technology will be

0:11:54.309 --> 0:11:57.120
<v Speaker 7>approved in Europe. And then we'll start to see similar

0:11:57.150 --> 0:11:59.580
<v Speaker 7>numbers coming out of Europe and a strong growth in

0:11:59.640 --> 0:12:00.439
<v Speaker 7>vehicle demand there.

0:12:01.420 --> 0:12:03.880
<v Speaker 1>Ben, very quickly, did you change your price target or

0:12:03.900 --> 0:12:06.600
<v Speaker 1>call on the stock after last night? No.

0:12:06.720 --> 0:12:09.220
<v Speaker 7>So we're still seeing at 600. So it's a very

0:12:09.280 --> 0:12:12.660
<v Speaker 7>healthy upside from where we are today. But everything that

0:12:12.700 --> 0:12:17.980
<v Speaker 7>we saw on the quarter was backing up our price target.

0:12:18.020 --> 0:12:21.480
<v Speaker 7>It's all based upon Tesla becoming an autonomy leader and

0:12:21.820 --> 0:12:23.839
<v Speaker 7>exactly what was announced yesterday.

0:12:24.910 --> 0:12:28.830
<v Speaker 1>Okay, let's talk about special committees at both Tesla and

0:12:28.850 --> 0:12:31.829
<v Speaker 1>SpaceX looking at the idea of a merger. Last night,

0:12:31.870 --> 0:12:34.599
<v Speaker 1>Tesla IR said they will not be taking questions on

0:12:34.640 --> 0:12:38.119
<v Speaker 1>the subject, but 24 hours before that, Musk disclosed it

0:12:38.179 --> 0:12:40.400
<v Speaker 1>after he was asked on SpaceX's call. You cover both.

0:12:40.440 --> 0:12:44.090
<v Speaker 1>You cover the area. What are you modeling for? That transaction,

0:12:44.120 --> 0:12:48.250
<v Speaker 1>how it looks like, who acquires who, and the prices involved? Yep, exactly.

0:12:48.490 --> 0:12:51.150
<v Speaker 7>So, we think the acquisition is going to go ahead.

0:12:52.370 --> 0:12:54.630
<v Speaker 7>It will likely happen by the end of next year

0:12:54.809 --> 0:12:56.839
<v Speaker 7>or be announced by the end of next year so

0:12:57.740 --> 0:13:02.239
<v Speaker 7>that both the companies and Musk himself can take advantage

0:13:02.320 --> 0:13:06.329
<v Speaker 7>of the political backdrop that we have right now. I

0:13:06.370 --> 0:13:08.730
<v Speaker 7>think the rose to reveal last week really goes to

0:13:08.809 --> 0:13:11.570
<v Speaker 7>show the fact that there's a lot of collaboration going

0:13:11.630 --> 0:13:15.880
<v Speaker 7>on between the businesses. I think you'll have SpaceX acquire Tesla,

0:13:16.130 --> 0:13:19.200
<v Speaker 7>given the fact that Musk has a much higher voting

0:13:19.260 --> 0:13:25.219
<v Speaker 7>control in SpaceX versus Tesla. The Tesla shareholders will want

0:13:25.280 --> 0:13:30.510
<v Speaker 7>a premium from for the acquisition, but there are clear synergies.

0:13:30.530 --> 0:13:32.390
<v Speaker 7>The Roadster event, as I said, was one of them,

0:13:32.850 --> 0:13:35.650
<v Speaker 7>but Tesla is a power company. What does SpaceX want?

0:13:35.710 --> 0:13:38.570
<v Speaker 7>SpaceX wants to roll out data centers. It needs power

0:13:38.590 --> 0:13:40.340
<v Speaker 7>for that. It will leverage Tesla Megapacks.

0:13:40.390 --> 0:13:43.140
<v Speaker 1>That's the bit that I'm trying to understand. I'm sorry

0:13:43.160 --> 0:13:46.640
<v Speaker 1>to jump in. Tesla's data center near-term ambition is 400 megawatts.

0:13:47.900 --> 0:13:50.500
<v Speaker 1>The reach goal for SpaceX next year is 20 gigawatts.

0:13:50.520 --> 0:13:54.610
<v Speaker 1>Like those are different scales of commitment. You know, but

0:13:54.650 --> 0:13:56.569
<v Speaker 1>he talks about the shared goals that they have. It

0:13:56.610 --> 0:13:58.059
<v Speaker 1>seems like one supporting the other.

0:13:59.120 --> 0:14:03.020
<v Speaker 7>Yeah, I would say it's almost two ways. So, Tesla

0:14:03.080 --> 0:14:07.740
<v Speaker 7>can leverage the compute of SpaceX. SpaceX will give Tesla

0:14:07.780 --> 0:14:12.260
<v Speaker 7>compute that will enable Tesla to improve FSD even faster

0:14:12.300 --> 0:14:16.540
<v Speaker 7>and ramp up the robo-taxi fleet and performance even faster.

0:14:17.000 --> 0:14:22.070
<v Speaker 7>Tesla will give the Megapack expertise, solar expertise, so SpaceX

0:14:22.110 --> 0:14:25.430
<v Speaker 7>can roll out its data centers faster. So even behind

0:14:25.470 --> 0:14:29.450
<v Speaker 7>the scenes on the rocket front, Tesla provides power components

0:14:29.510 --> 0:14:35.880
<v Speaker 7>to ensure that they can solve problems on the rocket

0:14:35.890 --> 0:14:39.540
<v Speaker 7>side of things. But of course, a big objective is

0:14:39.880 --> 0:14:42.060
<v Speaker 7>for Musk to control one entity as well. But we

0:14:42.080 --> 0:14:44.900
<v Speaker 7>do see synergies playing out between the two businesses.

0:14:46.230 --> 0:14:48.130
<v Speaker 1>Ben Howard of New Street Research. Great to have you

0:14:48.150 --> 0:14:49.580
<v Speaker 1>on Bloomberg Tech. Thank you very much. I want to

0:14:49.620 --> 0:14:52.060
<v Speaker 1>say on that potential merger, Bloomberg's learned there is a

0:14:52.080 --> 0:14:54.760
<v Speaker 1>push to announce a deal by the first half of

0:14:54.820 --> 0:14:57.540
<v Speaker 1>next year with Musk and his advisors looking to get

0:14:57.580 --> 0:15:00.380
<v Speaker 1>it done while President Trump is still in office. That's

0:15:00.420 --> 0:15:03.220
<v Speaker 1>all according to sources. And it's the big focus of

0:15:03.260 --> 0:15:06.300
<v Speaker 1>today's big take. Bloomberg's Craig Trudell, who leads our coverage

0:15:06.340 --> 0:15:10.990
<v Speaker 1>of both companies, is with us. That's the idea, right?

0:15:11.020 --> 0:15:14.250
<v Speaker 1>It's the environment, as Ben just outlined, that the companies

0:15:14.290 --> 0:15:17.150
<v Speaker 1>find themselves in. And there is kind of a deadline

0:15:17.170 --> 0:15:19.770
<v Speaker 1>to get something done. Yeah, that's absolutely right, Ed.

0:15:19.930 --> 0:15:21.450
<v Speaker 8>And I think, you know, if we take a step

0:15:21.490 --> 0:15:24.710
<v Speaker 8>back here, these are two boards stacked with a lot

0:15:24.730 --> 0:15:27.010
<v Speaker 8>of directors who are quite close with Elon, who go

0:15:27.050 --> 0:15:30.150
<v Speaker 8>back a long time. Tesla's board, you know, has managed

0:15:30.170 --> 0:15:34.260
<v Speaker 8>to pull off lining up companies. pretty incredible pay package

0:15:34.660 --> 0:15:38.120
<v Speaker 8>for its CEO only just a year ago. So if

0:15:38.180 --> 0:15:40.920
<v Speaker 8>and when Elon says jump, there's been a sort of

0:15:40.980 --> 0:15:43.200
<v Speaker 8>track record there of them asking how high.

0:15:44.040 --> 0:15:46.740
<v Speaker 1>So is that an issue? Probably not.

0:15:47.040 --> 0:15:50.140
<v Speaker 8>The question I think here is, will there be an

0:15:50.200 --> 0:15:54.250
<v Speaker 8>administration to sort of review the prospect of a merger

0:15:54.310 --> 0:15:57.350
<v Speaker 8>involving such big companies that is going to be supportive

0:15:57.370 --> 0:16:00.510
<v Speaker 8>of And I think Musk understandably wants to try and

0:16:00.690 --> 0:16:02.490
<v Speaker 8>make that push while Trump is still in office.

0:16:03.230 --> 0:16:05.030
<v Speaker 1>Well, even near a term, right, we're a couple of

0:16:05.080 --> 0:16:07.340
<v Speaker 1>weeks away from the midterms. And so from sort of

0:16:07.400 --> 0:16:11.580
<v Speaker 1>congressional oversight, the makeup of Congress could change. Like we

0:16:11.620 --> 0:16:12.760
<v Speaker 1>talk about that in the big take.

0:16:13.660 --> 0:16:16.910
<v Speaker 8>Yeah, that's exactly right. And it's not necessarily the case

0:16:16.930 --> 0:16:19.450
<v Speaker 8>that Congress has all that much of a say here.

0:16:20.080 --> 0:16:23.800
<v Speaker 8>But it absolutely is the case that if you're Musk,

0:16:23.820 --> 0:16:26.680
<v Speaker 8>you're also bracing for the possibility that this could take

0:16:26.800 --> 0:16:29.270
<v Speaker 8>quite some time. The U.S., of course, will have a say,

0:16:29.710 --> 0:16:31.510
<v Speaker 8>but it's not the only country that will have a say.

0:16:31.630 --> 0:16:35.470
<v Speaker 8>And the fact that China looms so large for Tesla

0:16:35.910 --> 0:16:39.290
<v Speaker 8>with such a presence there, and there's some potential conflict there, right,

0:16:39.360 --> 0:16:41.590
<v Speaker 8>where The fact that SpaceX is a U.S.

0:16:41.950 --> 0:16:42.770
<v Speaker 1>Contractor, a U.S.

0:16:42.790 --> 0:16:47.050
<v Speaker 8>Defense contractor, Starlink doesn't necessarily align with how the CCP,

0:16:47.170 --> 0:16:49.840
<v Speaker 8>you know, kind of wants their Internet to work. Will

0:16:49.860 --> 0:16:52.280
<v Speaker 8>they have notes and have a view as to whether

0:16:52.300 --> 0:16:53.820
<v Speaker 8>or not this merger should go through?

0:16:53.880 --> 0:16:56.520
<v Speaker 1>I think that's one of the big questions here. I

0:16:56.540 --> 0:16:58.990
<v Speaker 1>would point out in a regulatory filing, they confirmed this

0:16:59.030 --> 0:17:00.630
<v Speaker 1>morning in the special committees. I know it was discussed

0:17:00.650 --> 0:17:02.110
<v Speaker 1>on the call, and they had the same language, like,

0:17:02.130 --> 0:17:03.390
<v Speaker 1>this isn't a certainty.

0:17:03.790 --> 0:17:03.950
<v Speaker 2>Just.

0:17:04.700 --> 0:17:07.000
<v Speaker 1>The other thing, of course, is the voting power differences.

0:17:07.020 --> 0:17:08.500
<v Speaker 1>We're going to get to that in the coming weeks.

0:17:08.540 --> 0:17:10.659
<v Speaker 1>Bloomberg's Craig Trudell, really great to have you with us.

0:17:10.700 --> 0:17:14.160
<v Speaker 1>Thank you. Now, coming up, IBM, another earning story. Delivered

0:17:14.200 --> 0:17:17.700
<v Speaker 1>better than expected third quarter results overall, but there's still

0:17:17.740 --> 0:17:19.760
<v Speaker 1>some concern out there. The stock a little weaker, down

0:17:20.060 --> 0:17:22.219
<v Speaker 1>seven-tenths of 1%. We've got more on those earnings when

0:17:22.240 --> 0:17:39.449
<v Speaker 1>we return. This is Bloomberg Tech. IBM delivered better than

0:17:39.510 --> 0:17:42.130
<v Speaker 1>expected sales growth in its third quarter earnings report, but

0:17:42.210 --> 0:17:45.929
<v Speaker 1>consulting revenues missed estimates. Ultimately, the numbers have failed to

0:17:45.950 --> 0:17:48.910
<v Speaker 1>impress investors. There's some concern out there about this shift

0:17:49.290 --> 0:17:51.790
<v Speaker 1>in corporate technology spending. Joining us to talk more about

0:17:51.810 --> 0:17:55.270
<v Speaker 1>the earnings is Fatima Bulani, Managing Director at Citi. There's

0:17:55.290 --> 0:17:57.330
<v Speaker 1>a buy rating on the stock with a target price

0:17:57.650 --> 0:18:01.150
<v Speaker 1>of $ 280. It's interesting because a decline of 7 tenths

0:18:01.210 --> 0:18:03.119
<v Speaker 1>of a percent is not a lot to write home about.

0:18:03.170 --> 0:18:05.940
<v Speaker 1>And a lot of what IBM presented was quite good.

0:18:06.840 --> 0:18:08.919
<v Speaker 1>There seems to be some worry out there in the

0:18:08.940 --> 0:18:10.760
<v Speaker 1>market that IBM is just not it right now in

0:18:10.780 --> 0:18:12.040
<v Speaker 1>the world of technology spending.

0:18:13.920 --> 0:18:16.159
<v Speaker 9>Yeah, let me unpack a lot of that, because I

0:18:16.180 --> 0:18:18.960
<v Speaker 9>think what you're alluding to is IBM is actually a

0:18:18.980 --> 0:18:23.409
<v Speaker 9>very polarizing stock. Now, just the forensic analysis on the

0:18:23.630 --> 0:18:28.450
<v Speaker 9>three key print last night, really the headline beats, you know,

0:18:28.510 --> 0:18:33.260
<v Speaker 9>coming from the infrastructure and software segment's of the overall businesses,

0:18:33.320 --> 0:18:36.710
<v Speaker 9>which in totality are two thirds of the overall revenue mix.

0:18:37.250 --> 0:18:39.609
<v Speaker 9>Those are important points to highlight because we saw a

0:18:39.650 --> 0:18:44.379
<v Speaker 9>re-acceleration in those segments. And they indicate to us one

0:18:44.430 --> 0:18:49.120
<v Speaker 9>very important thing, that the mainframe business is at the

0:18:49.160 --> 0:18:52.600
<v Speaker 9>very least stable. Now, in a lot of ways, we

0:18:52.660 --> 0:18:56.560
<v Speaker 9>thought this was a little bit of a surprising redemption

0:18:56.640 --> 0:18:59.180
<v Speaker 9>quarter on the back of what was a pretty monumental

0:18:59.280 --> 0:19:03.300
<v Speaker 9>miss for this business back in 2Q. So the fact

0:19:03.320 --> 0:19:07.480
<v Speaker 9>that this company was able to recapture and recoup two

0:19:07.580 --> 0:19:11.000
<v Speaker 9>thirds of these slipped transactions, you know, within a matter

0:19:11.100 --> 0:19:11.820
<v Speaker 9>of 90 days.

0:19:12.100 --> 0:19:13.379
<v Speaker 6>We do think that's impressive.

0:19:13.740 --> 0:19:16.320
<v Speaker 9>But, you know, in the same breath, this is not

0:19:16.359 --> 0:19:20.570
<v Speaker 9>sort of a decisive victory for the bulls, where there

0:19:20.609 --> 0:19:22.790
<v Speaker 9>is still going to be a lot of work and

0:19:23.090 --> 0:19:25.350
<v Speaker 9>numbers and execution that IBM is going to have to

0:19:25.390 --> 0:19:28.450
<v Speaker 9>put up to sort of compel the bearers to back

0:19:28.540 --> 0:19:32.260
<v Speaker 9>off from this narrative around, hey, the mainframe business is

0:19:32.380 --> 0:19:37.760
<v Speaker 9>under structural duress. and existential duress from the AI onslaught

0:19:37.800 --> 0:19:38.540
<v Speaker 9>from the labs.

0:19:38.960 --> 0:19:41.680
<v Speaker 1>That's my question. What is IBM in the AI age?

0:19:41.720 --> 0:19:45.470
<v Speaker 1>Bloomberg's Romain Bostic got on the phone with IBM CEO

0:19:45.530 --> 0:19:49.770
<v Speaker 1>Arvind Krishna, and he presents that in the shift to inference,

0:19:49.920 --> 0:19:53.820
<v Speaker 1>IBM is primed for the AI age. Is that a

0:19:53.840 --> 0:19:54.520
<v Speaker 1>part of your model?

0:19:55.710 --> 0:19:57.619
<v Speaker 9>Yeah, look, there's a couple of angles to this, right?

0:19:57.720 --> 0:20:02.200
<v Speaker 9>Our framing here has been IBM is not quite in

0:20:02.240 --> 0:20:05.639
<v Speaker 9>the AI loser camp, nor are they decisively in the

0:20:05.740 --> 0:20:06.420
<v Speaker 9>AI winner camp.

0:20:06.440 --> 0:20:07.580
<v Speaker 6>They are an AI survivor.

0:20:07.880 --> 0:20:11.370
<v Speaker 9>And we have to remember that IBM is so pervasive

0:20:11.450 --> 0:20:15.869
<v Speaker 9>in some of the largest and most complex, most heterogeneous

0:20:16.630 --> 0:20:18.610
<v Speaker 9>IT organizations.

0:20:17.930 --> 0:20:18.470
<v Speaker 6>In the world.

0:20:18.810 --> 0:20:21.490
<v Speaker 9>So it gives them a pretty important seat at the table.

0:20:21.119 --> 0:20:25.060
<v Speaker 1>In terms of enabling. You said AI survivor, right?

0:20:25.710 --> 0:20:28.410
<v Speaker 9>Ai survivor it's not a bad place to.

0:20:28.369 --> 0:20:31.520
<v Speaker 1>Be that's not about you no not a bad place

0:20:31.540 --> 0:20:33.379
<v Speaker 1>i mean i'm looking at like you and your industry

0:20:33.420 --> 0:20:36.160
<v Speaker 1>colleagues and what in aggregate you're saying you know red

0:20:36.220 --> 0:20:37.879
<v Speaker 1>hat i drove quite a lot of growth on the

0:20:37.900 --> 0:20:40.550
<v Speaker 1>software side um in the quarter gone i think a

0:20:40.570 --> 0:20:42.830
<v Speaker 1>lot of people are basically saying how sustainable is the

0:20:42.869 --> 0:20:46.990
<v Speaker 1>growth that they just printed going further afield what's your

0:20:47.010 --> 0:20:47.850
<v Speaker 1>view on.

0:20:47.670 --> 0:20:50.219
<v Speaker 9>That So the view here is just to kind of

0:20:50.359 --> 0:20:53.719
<v Speaker 9>close the thought on this AI survivor notion. They are

0:20:53.880 --> 0:20:57.929
<v Speaker 9>very deeply embedded in very important and consequential parts of

0:20:57.950 --> 0:21:00.850
<v Speaker 9>the IT stack, right? So while they're not necessarily getting

0:21:00.930 --> 0:21:04.710
<v Speaker 9>the torque that you would see from, you know, Other

0:21:04.830 --> 0:21:06.910
<v Speaker 9>AI winners that have been anointed, you know, we think

0:21:06.950 --> 0:21:07.629
<v Speaker 9>of Datadog in.

0:21:07.590 --> 0:21:08.609
<v Speaker 4>Our coverage universe.

0:21:08.670 --> 0:21:10.750
<v Speaker 9>We think of a Cloudflare in our coverage universe and

0:21:10.770 --> 0:21:11.869
<v Speaker 9>even some of the cybernames.

0:21:11.930 --> 0:21:12.130
<v Speaker 10>Right.

0:21:12.170 --> 0:21:15.910
<v Speaker 9>So there isn't a clear and unequivocal tether to how

0:21:15.930 --> 0:21:20.850
<v Speaker 9>they're benefiting from AI. But certainly there's opportunities here from

0:21:20.890 --> 0:21:26.010
<v Speaker 9>the expanse of IBM's portfolio to help. the average organization

0:21:26.210 --> 0:21:29.970
<v Speaker 9>operationalize and productionize AI. And that's certainly been the message

0:21:30.010 --> 0:21:32.710
<v Speaker 9>from the company. And where some of that does fall

0:21:32.780 --> 0:21:37.080
<v Speaker 9>short in terms of the reality is the organic software business.

0:21:38.119 --> 0:21:43.610
<v Speaker 9>That has definitely been a flashpoint for a lot of investors.

0:21:43.630 --> 0:21:45.670
<v Speaker 9>That's part of the reason why I've mentioned that this

0:21:45.730 --> 0:21:48.179
<v Speaker 9>is still a pretty polarizing stock. in spite of the

0:21:48.220 --> 0:21:52.740
<v Speaker 9>fact that they put up a pretty strong beat, reaccelerated sequentially,

0:21:53.300 --> 0:21:56.439
<v Speaker 9>and sort of proved to you that they took what

0:21:56.480 --> 0:21:59.010
<v Speaker 9>was theirs in terms of the mainframe business, right? So

0:21:59.030 --> 0:22:01.869
<v Speaker 9>a little bit of a redemption quarter, but not out

0:22:01.890 --> 0:22:04.810
<v Speaker 9>of the woods yet as it relates to some of

0:22:04.830 --> 0:22:08.170
<v Speaker 9>the bearish concerns that, hey, maybe they've won the skirmish,

0:22:08.330 --> 0:22:09.050
<v Speaker 9>but not the war.

0:22:10.170 --> 0:22:12.230
<v Speaker 1>The stock had been down almost 5%, you know, when

0:22:12.250 --> 0:22:14.630
<v Speaker 1>the market opened up this Thursday morning. It's now off

0:22:14.650 --> 0:22:18.260
<v Speaker 1>by a percentage point as people digest everything that Fatima Boulani,

0:22:18.340 --> 0:22:20.100
<v Speaker 1>managing director of Citi, just went through. Great to have

0:22:20.119 --> 0:22:22.100
<v Speaker 1>you on the show. Thank you so much. Coming up

0:22:22.400 --> 0:22:24.699
<v Speaker 1>in the program, Alison Porter of Janus Henderson joins us.

0:22:25.000 --> 0:22:27.240
<v Speaker 1>Meta is a top holding of the firm's tech leaders, right?

0:22:27.260 --> 0:22:29.480
<v Speaker 1>And we'll get her take, yes, on the AI spending.

0:22:29.520 --> 0:22:32.040
<v Speaker 1>But I want to talk about this new Class C

0:22:32.140 --> 0:22:36.560
<v Speaker 1>share issuance and the Zuckerberg of it all. Half time

0:22:36.619 --> 0:22:39.920
<v Speaker 1>here in San Francisco. Stay with us. This is Bloomberg Tech.

0:22:44.540 --> 0:22:47.200
<v Speaker 1>AI risk, I would say, is kind of the prevailing

0:22:47.300 --> 0:22:50.419
<v Speaker 1>theme of the market this morning. The Nasdaq 100 is

0:22:50.500 --> 0:22:53.619
<v Speaker 1>off by 2.4% at its session low. It was on

0:22:53.660 --> 0:22:56.659
<v Speaker 1>track for its biggest drop, I think, since July. The

0:22:56.680 --> 0:22:59.669
<v Speaker 1>semiconductor index, or the SOX, also is kind of where

0:22:59.690 --> 0:23:01.550
<v Speaker 1>the brunt of selling is right now. I'd say also

0:23:01.570 --> 0:23:04.470
<v Speaker 1>you have NVIDIA, AMD, etc. as the big drags at

0:23:04.490 --> 0:23:07.270
<v Speaker 1>the index level. But the earnings story, the big one

0:23:07.330 --> 0:23:10.770
<v Speaker 1>is Meta. Down again on track for its biggest drop

0:23:10.810 --> 0:23:15.760
<v Speaker 1>since July, off 6% or so, off session lows. Narrowed

0:23:15.780 --> 0:23:19.859
<v Speaker 1>the CapEx range, strong top line growth, the ad business outperforming,

0:23:19.900 --> 0:23:22.869
<v Speaker 1>good signals for 27%. but they issued a new class

0:23:22.910 --> 0:23:26.890
<v Speaker 1>of stock. There's also a warning in the credit market.

0:23:26.910 --> 0:23:29.850
<v Speaker 1>I'm looking at Meta 5-year CDS, which, as you can

0:23:29.910 --> 0:23:32.800
<v Speaker 1>see on the right-hand side of that screen, is at

0:23:32.820 --> 0:23:36.020
<v Speaker 1>a record level since trading began. The credit risk for

0:23:36.040 --> 0:23:39.320
<v Speaker 1>Meta is becoming a story. We can get into the specifics.

0:23:39.660 --> 0:23:42.900
<v Speaker 1>Alison Porter, Portfolio Manager at Janus Henderson. on the global

0:23:42.940 --> 0:23:45.390
<v Speaker 1>technology leaders team. I'd looked at some of your holdings,

0:23:45.430 --> 0:23:48.369
<v Speaker 1>like Meta is up there, right? I would love to

0:23:48.410 --> 0:23:51.590
<v Speaker 1>start with this new class of stock. It has economic value,

0:23:51.869 --> 0:23:54.909
<v Speaker 1>but no voting power. And the reality of it is

0:23:54.970 --> 0:23:58.070
<v Speaker 1>that the company is doing this to give Mr. Zuckerberg

0:23:58.090 --> 0:24:00.810
<v Speaker 1>some liquidity to cover a tax bill if here in

0:24:00.850 --> 0:24:05.109
<v Speaker 1>California Prop 40 passes, but it also keeps him focused

0:24:05.170 --> 0:24:07.570
<v Speaker 1>and imposed. What do you make of all that?

0:24:09.470 --> 0:24:13.080
<v Speaker 11>Yeah, I think the strength that Meta has seen on

0:24:14.600 --> 0:24:18.740
<v Speaker 11>revenue needs to be funded from CapEx. And the market

0:24:18.820 --> 0:24:23.850
<v Speaker 11>has been expecting that we would see a capital raise

0:24:24.050 --> 0:24:28.480
<v Speaker 11>at some point. And this new share class structure you know,

0:24:28.710 --> 0:24:33.610
<v Speaker 11>is designed to allow them to focus on potential acquisitions,

0:24:33.730 --> 0:24:37.430
<v Speaker 11>partly for hiring for employee comp. We saw CJ Desai

0:24:37.510 --> 0:24:44.290
<v Speaker 11>hired from MongoDB and also to fund CapEx. So it's multi-purpose.

0:24:44.330 --> 0:24:45.510
<v Speaker 11>And we do think this is going to be an

0:24:45.590 --> 0:24:48.840
<v Speaker 11>overhang on the stock through until we see that capital

0:24:48.880 --> 0:24:52.590
<v Speaker 11>raise actually occur. So I should say sometime in Q1.

0:24:53.900 --> 0:24:54.110
<v Speaker 3>Right.

0:24:54.270 --> 0:24:56.890
<v Speaker 1>You know, it's a 90 day process. And I should

0:24:56.910 --> 0:24:59.070
<v Speaker 1>say that in the earnings statement and on the call,

0:24:59.950 --> 0:25:02.129
<v Speaker 1>Meta CFO Susan Lee framed this very much as a

0:25:02.190 --> 0:25:06.030
<v Speaker 1>mechanism to get flexibility in financing the AI spend. But

0:25:06.070 --> 0:25:08.490
<v Speaker 1>in the regulatory filing, it's also detailed that there's an

0:25:08.550 --> 0:25:12.050
<v Speaker 1>ad hoc founder led agreement with Mr. Zuckerberg. Is it

0:25:12.090 --> 0:25:15.070
<v Speaker 1>important to you that you have the founder CEO sort

0:25:15.109 --> 0:25:19.510
<v Speaker 1>of locked in to lead an increasingly AI focused strategy

0:25:19.530 --> 0:25:20.010
<v Speaker 1>at the company?

0:25:21.600 --> 0:25:24.399
<v Speaker 11>Yeah, I think it's very important that we do see

0:25:25.609 --> 0:25:29.550
<v Speaker 11>Mark Zuckerberg and we do see him continue to lead,

0:25:29.570 --> 0:25:33.990
<v Speaker 11>like all founder-led businesses, including Tesla. That's where we've seen

0:25:34.850 --> 0:25:38.440
<v Speaker 11>founders be able to have that vision and invest against

0:25:39.080 --> 0:25:44.500
<v Speaker 11>future growth. Obviously, Mark Zuckerberg's been core to that story

0:25:45.280 --> 0:25:49.320
<v Speaker 11>and having the vision on you is driving towards the

0:25:49.400 --> 0:25:53.480
<v Speaker 11>personal intelligence product. But we do think it's really important

0:25:53.600 --> 0:25:56.640
<v Speaker 11>that they focus on shareholder returns. And on the call

0:25:56.680 --> 0:25:59.340
<v Speaker 11>last night, that was a really big feature, focusing on

0:26:00.300 --> 0:26:05.810
<v Speaker 11>spending on CapEx, where you actually would see a return

0:26:05.880 --> 0:26:09.610
<v Speaker 11>and a lot of focus on granular monetization as part

0:26:09.650 --> 0:26:11.390
<v Speaker 11>of that. But yeah, it is important that we see

0:26:11.450 --> 0:26:14.710
<v Speaker 11>Mark continue to lead Meta forward.

0:26:15.980 --> 0:26:19.730
<v Speaker 1>You mean monetization in the sense of monetizing the AI

0:26:19.830 --> 0:26:23.790
<v Speaker 1>tools and AI products specifically, or you're looking for evidence

0:26:23.910 --> 0:26:26.580
<v Speaker 1>that the core business, which is the ad business, is

0:26:26.640 --> 0:26:28.700
<v Speaker 1>improving because of their work in AI?

0:26:30.680 --> 0:26:33.180
<v Speaker 11>Both of those things. And I think obviously we're seeing

0:26:33.220 --> 0:26:36.400
<v Speaker 11>the improvement in the core business already. And they talked

0:26:36.440 --> 0:26:39.990
<v Speaker 11>about strength and growth into next year. There's already over 20%

0:26:39.990 --> 0:26:44.070
<v Speaker 11>growth expected into next year. And we've seen a pickup

0:26:44.190 --> 0:26:47.740
<v Speaker 11>in not just user engagement, but also in growth. in

0:26:47.960 --> 0:26:52.320
<v Speaker 11>pricing and in ad volume across a number of their

0:26:52.400 --> 0:26:56.300
<v Speaker 11>different products. I think it is very important, though, that

0:26:56.340 --> 0:26:59.240
<v Speaker 11>for products like Muse, that we start to see a

0:26:59.290 --> 0:27:03.510
<v Speaker 11>monetization path. And that's where it's much less certain. And

0:27:03.690 --> 0:27:08.850
<v Speaker 11>he did talk about typically Meta would focus on first-drive users,

0:27:08.890 --> 0:27:12.630
<v Speaker 11>first-drive engagement, then monetization. But they've talked about moving much

0:27:12.710 --> 0:27:16.000
<v Speaker 11>faster on Muse to be able to show that monetization

0:27:16.060 --> 0:27:18.560
<v Speaker 11>at a much earlier stage. And that's going to be

0:27:18.619 --> 0:27:22.919
<v Speaker 11>important if you expect investors to actually pay for the

0:27:22.980 --> 0:27:26.330
<v Speaker 11>CapEx and they want to do an equity raise to

0:27:26.430 --> 0:27:31.850
<v Speaker 11>show that they are delivering on returns from that MetaMuse product.

0:27:33.119 --> 0:27:35.740
<v Speaker 1>The last sort of metadata point that I'm looking at

0:27:36.280 --> 0:27:37.960
<v Speaker 1>is what came out in the filing this morning, which

0:27:38.040 --> 0:27:43.180
<v Speaker 1>is the commitments, the spending commitments have gone to $ 825 billion.

0:27:43.220 --> 0:27:46.110
<v Speaker 1>I think in July it was about $ 700 billion. So

0:27:46.140 --> 0:27:50.189
<v Speaker 1>they didn't guide for CapEx for 27 yet, but the

0:27:50.230 --> 0:27:53.810
<v Speaker 1>direction of travel is pretty clear. Most people comfortable with

0:27:53.850 --> 0:27:56.389
<v Speaker 1>that or that number is getting a bit scary with

0:27:56.410 --> 0:27:56.949
<v Speaker 1>how big it is?

0:27:58.460 --> 0:28:02.180
<v Speaker 11>I think, you know, the overall commitments, it will depend

0:28:02.240 --> 0:28:05.090
<v Speaker 11>on how they show monetization. And I think that's very

0:28:05.140 --> 0:28:10.090
<v Speaker 11>important that they continue to see leverage and strong revenue growth.

0:28:10.490 --> 0:28:13.530
<v Speaker 11>They've also talked about one thing that Meta doesn't have

0:28:13.550 --> 0:28:17.070
<v Speaker 11>that Amazon and Google have is this third party cloud

0:28:17.130 --> 0:28:20.210
<v Speaker 11>business that gives them visibility on the capex that they have.

0:28:20.630 --> 0:28:23.530
<v Speaker 11>And Meta just don't have that visibility to spend against.

0:28:23.550 --> 0:28:24.889
<v Speaker 11>And I think that's why this is going to be

0:28:24.930 --> 0:28:29.350
<v Speaker 11>an overhang on the stock market. And through the capital,

0:28:29.730 --> 0:28:32.220
<v Speaker 11>through this period, until we see the capital raise, until

0:28:32.260 --> 0:28:35.260
<v Speaker 11>we get more colour on the size of that capex.

0:28:35.300 --> 0:28:39.610
<v Speaker 11>And we really have to drive down compute efficiency because

0:28:39.620 --> 0:28:43.430
<v Speaker 11>the unit economics on MUSE are not attractive at present.

0:28:44.420 --> 0:28:47.100
<v Speaker 1>I think that's interesting because I'm already obviously thinking about

0:28:47.140 --> 0:28:49.370
<v Speaker 1>next week. And if you say, OK, SpaceX is a

0:28:49.380 --> 0:28:54.530
<v Speaker 1>neocloud and meta is meta. What did we learn about

0:28:54.590 --> 0:28:57.900
<v Speaker 1>what might come in the balance of earnings season when

0:28:57.940 --> 0:28:59.720
<v Speaker 1>it comes to the hyperscaler capex story?

0:29:01.260 --> 0:29:04.980
<v Speaker 11>Well, the strength that we're seeing from Meta, I think

0:29:05.050 --> 0:29:10.150
<v Speaker 11>particularly even the early strength in Muse, the compute power needed,

0:29:10.170 --> 0:29:14.530
<v Speaker 11>you need a lot of CPUs for that agentic portion.

0:29:15.490 --> 0:29:19.450
<v Speaker 11>Leveraging agentic, that very much plays into Amazon's Gravitron and

0:29:19.490 --> 0:29:24.730
<v Speaker 11>CPU product. And we've seen from the leverage, we had

0:29:24.790 --> 0:29:28.070
<v Speaker 11>some disclosures from Anthropic last week as well, that that

0:29:28.090 --> 0:29:32.700
<v Speaker 11>should also bode very well for the TPU business at Alphabet.

0:29:33.160 --> 0:29:35.760
<v Speaker 11>So we would expect to see continued strength in the

0:29:35.840 --> 0:29:41.610
<v Speaker 11>cloud divisions of both Amazon and Alphabet. next week and

0:29:41.630 --> 0:29:45.600
<v Speaker 11>that continued build on their backlog, especially from that vertically

0:29:45.690 --> 0:29:51.560
<v Speaker 11>integrated setup that they have showing through to the supply

0:29:51.620 --> 0:29:53.760
<v Speaker 11>chain that Meta doesn't have just now.

0:29:55.760 --> 0:29:59.830
<v Speaker 1>The story for quite a long time, or for Big

0:29:59.910 --> 0:30:03.750
<v Speaker 1>portions of this year was to be long semiconductors and

0:30:03.810 --> 0:30:06.770
<v Speaker 1>short software. And there have been different parts of the year,

0:30:06.810 --> 0:30:08.630
<v Speaker 1>like the coming out of the summer, where that may

0:30:08.670 --> 0:30:13.150
<v Speaker 1>be changed, if not reversed. Alison, where's your thinking right

0:30:13.190 --> 0:30:15.850
<v Speaker 1>now on the right place to be in the market?

0:30:15.870 --> 0:30:22.230
<v Speaker 11>This is a great question. cycles, past compute waves. And

0:30:22.250 --> 0:30:24.630
<v Speaker 11>we think we're moving out of that phase of just

0:30:24.770 --> 0:30:30.390
<v Speaker 11>blindly buying physical infrastructure for AI, just buying all of

0:30:30.510 --> 0:30:34.910
<v Speaker 11>the semiconductor, going long semis, short the software index. We

0:30:34.930 --> 0:30:37.970
<v Speaker 11>think that's far too blunt. And going forward, investors are

0:30:37.990 --> 0:30:40.750
<v Speaker 11>going to have to be very active because this volatility

0:30:40.790 --> 0:30:43.510
<v Speaker 11>is here to stay. And also, they're going to have

0:30:43.550 --> 0:30:46.710
<v Speaker 11>to be much more stock specific. There are lots of

0:30:47.000 --> 0:30:53.160
<v Speaker 11>still really attractive opportunities in semis and physical infrastructure, but

0:30:53.200 --> 0:30:56.120
<v Speaker 11>you have to pick your points in terms of valuation

0:30:56.660 --> 0:31:01.220
<v Speaker 11>and product cycles. And then the market as this AI

0:31:01.420 --> 0:31:06.140
<v Speaker 11>investment goes. phase evolves, we're really looking for those companies

0:31:06.200 --> 0:31:10.570
<v Speaker 11>that can drive applications and drive revenue growth from that.

0:31:10.630 --> 0:31:13.490
<v Speaker 11>So we're focusing more and looking for those names and

0:31:13.570 --> 0:31:17.470
<v Speaker 11>leading applications and leading platforms that can drive growth. We've

0:31:17.490 --> 0:31:21.810
<v Speaker 11>seen some You know, really strong numbers out of Spotify today.

0:31:21.910 --> 0:31:25.450
<v Speaker 11>We're seeing a number of companies where we're seeing actual

0:31:25.760 --> 0:31:30.620
<v Speaker 11>leverage on AI. We've seen some headcount reductions come through.

0:31:30.680 --> 0:31:32.940
<v Speaker 11>And we think next year could be another attractive year

0:31:32.980 --> 0:31:36.740
<v Speaker 11>of efficiency for tech. As we see companies, we're really

0:31:36.780 --> 0:31:42.120
<v Speaker 11>looking at companies that can drive margin expansion in 2027.

0:31:42.160 --> 0:31:44.440
<v Speaker 1>The margin expansion is really important. Just on Spotify, it

0:31:45.000 --> 0:31:47.700
<v Speaker 1>hit 788 million monthly users and basically had some fighting

0:31:47.760 --> 0:31:51.380
<v Speaker 1>talk about how it's not going to get displaced by AI.

0:31:51.640 --> 0:31:55.120
<v Speaker 1>The stock's up significantly. It's US-listed shares. On margin expansion,

0:31:55.900 --> 0:31:59.160
<v Speaker 1>where does the free cash flow story for any given

0:31:59.220 --> 0:32:02.739
<v Speaker 1>company or actually an aggregate feature for you? A lot

0:32:02.770 --> 0:32:06.989
<v Speaker 1>of people seemed really calm when some absolute cash machines

0:32:07.390 --> 0:32:10.670
<v Speaker 1>tipped into negative free cash flow, albeit you know, temporarily

0:32:10.730 --> 0:32:13.710
<v Speaker 1>or on a longer-term basis, it's hard to gauge where

0:32:13.750 --> 0:32:14.830
<v Speaker 1>the concern is on that.

0:32:16.150 --> 0:32:18.390
<v Speaker 11>Yeah. And I think the important point is free cash

0:32:18.410 --> 0:32:24.040
<v Speaker 11>flow is not disappearing. It is transferring around the ecosystem. So,

0:32:24.080 --> 0:32:26.620
<v Speaker 11>we have seen the hyperscalers continue to spend. But if

0:32:26.660 --> 0:32:29.780
<v Speaker 11>you look at the free cash flow of NVIDIA, for example,

0:32:30.840 --> 0:32:34.100
<v Speaker 11>we expect that that will continue to grow. And NVIDIA

0:32:34.500 --> 0:32:38.860
<v Speaker 11>has been reinvesting that into the ecosystem You know, very similarly,

0:32:38.880 --> 0:32:42.630
<v Speaker 11>you know, we saw Apple back after the iPhone launch.

0:32:43.290 --> 0:32:45.310
<v Speaker 11>They saw huge free cash flow growth and they had

0:32:45.350 --> 0:32:48.450
<v Speaker 11>to invest that into the supply chain. And that's what

0:32:48.470 --> 0:32:51.750
<v Speaker 11>we're seeing with NVIDIA. So I think it's important investors

0:32:51.770 --> 0:32:55.690
<v Speaker 11>don't think about free cash flow disappearing, but it's circulating

0:32:55.710 --> 0:32:58.910
<v Speaker 11>in the system. And also the point at which we

0:32:58.990 --> 0:33:01.930
<v Speaker 11>expect to see some of these hyperscalers actually reach a

0:33:01.970 --> 0:33:06.650
<v Speaker 11>free cash flow inflection point as CapEx actually peaks. which

0:33:06.710 --> 0:33:11.220
<v Speaker 11>is likely to be into 2028. And then we expect

0:33:11.260 --> 0:33:15.010
<v Speaker 11>to see some free cash flow inflection positively for them

0:33:16.030 --> 0:33:17.190
<v Speaker 11>in the back half of 27 and 28 as well.

0:33:21.310 --> 0:33:24.270
<v Speaker 1>Alison Porter, Portfolio Manager at Janus Henderson. We covered a

0:33:24.450 --> 0:33:26.970
<v Speaker 1>very broad range of subject matter. Appreciate your time. Thank

0:33:26.990 --> 0:33:30.270
<v Speaker 1>you very much indeed. Now coming up, AI uncovers a

0:33:30.450 --> 0:33:34.830
<v Speaker 1>major security flaw hiding in millions of websites. But could

0:33:34.850 --> 0:33:37.890
<v Speaker 1>the same technology also give hackers an edge? Greg Brockman,

0:33:37.910 --> 0:33:41.780
<v Speaker 1>OpenAI President, and Matthew Prince, Cloudflare CEO, joining us next.

0:33:42.160 --> 0:33:51.660
<v Speaker 1>This is Bloomberg Tech. AI helped uncover a potentially catastrophic

0:33:51.700 --> 0:33:56.590
<v Speaker 1>vulnerability hiding for years in PHP. It's a programming language

0:33:56.710 --> 0:34:00.390
<v Speaker 1>used to build and run millions of websites and web applications.

0:34:00.890 --> 0:34:06.060
<v Speaker 1>OpenAI's technology helped researchers find the flaw. Cloudflare helped protect

0:34:06.120 --> 0:34:10.719
<v Speaker 1>websites before that flaw was publicly disclosed. But AI could

0:34:10.739 --> 0:34:13.219
<v Speaker 1>also help hackers exploit it faster. The concern is that

0:34:13.239 --> 0:34:16.620
<v Speaker 1>the flaws could potentially let a hacker steal data or

0:34:16.680 --> 0:34:20.020
<v Speaker 1>seize control of websites. Joining us now, OpenAI President Greg

0:34:20.060 --> 0:34:23.210
<v Speaker 1>Brockman and Cloudflare CEO Matthew Prince. I'm grateful to both

0:34:23.270 --> 0:34:25.989
<v Speaker 1>of you for your time. Greg, I want to get

0:34:26.050 --> 0:34:30.230
<v Speaker 1>specific here. So what happened was Calif used... GPT 5.6

0:34:30.230 --> 0:34:33.090
<v Speaker 1>Sol to find a version of this bug back in August.

0:34:33.570 --> 0:34:37.660
<v Speaker 1>And then OpenAI independently kind of found a more complete

0:34:37.719 --> 0:34:41.760
<v Speaker 1>version in September. What is it that OpenAI's models actually

0:34:42.260 --> 0:34:45.420
<v Speaker 1>did or do that allowed you to find something that

0:34:45.440 --> 0:34:48.690
<v Speaker 1>I think researchers have not found for like years and years?

0:34:50.190 --> 0:34:53.670
<v Speaker 3>First of all, thank you for having me. You're exactly right.

0:34:53.910 --> 0:34:57.870
<v Speaker 3>Calif had a security researcher use our models very guided

0:34:57.930 --> 0:35:02.730
<v Speaker 3>by that researcher's expertise and intuition. A month later, we

0:35:02.810 --> 0:35:05.950
<v Speaker 3>used our latest internal model that fully autonomously found this

0:35:05.969 --> 0:35:11.670
<v Speaker 3>more complete version and very independently found the full version

0:35:11.690 --> 0:35:15.390
<v Speaker 3>of what was released. The thing to realize about the

0:35:15.450 --> 0:35:19.840
<v Speaker 3>AI models is that fundamentally, they're empowering tools. And that

0:35:19.860 --> 0:35:23.020
<v Speaker 3>they're able to work across a lot more context than

0:35:23.060 --> 0:35:26.000
<v Speaker 3>people are. You know, I myself actually have found a

0:35:26.140 --> 0:35:29.239
<v Speaker 3>security vulnerability way back in 2010 in an open source

0:35:29.260 --> 0:35:32.530
<v Speaker 3>piece of software. And I remember finding a small function

0:35:32.610 --> 0:35:35.350
<v Speaker 3>that looked like it wasn't quite implemented correctly, but all

0:35:35.370 --> 0:35:38.430
<v Speaker 3>of the Usages of it had been protected, but then

0:35:38.530 --> 0:35:41.549
<v Speaker 3>actually someone added some code later that didn't protect in

0:35:41.730 --> 0:35:43.529
<v Speaker 3>quite the right way. And so it's this kind of

0:35:43.890 --> 0:35:46.250
<v Speaker 3>sleuthing that you do as a person. But it's very

0:35:46.350 --> 0:35:48.730
<v Speaker 3>hard to keep all that context in your head, to

0:35:48.750 --> 0:35:51.859
<v Speaker 3>have expertise across many different disciplines, and to think about

0:35:51.920 --> 0:35:54.160
<v Speaker 3>how the software might be used in practice. And so

0:35:54.380 --> 0:35:58.180
<v Speaker 3>that capability, which is what makes AI so useful, that

0:35:58.219 --> 0:36:00.299
<v Speaker 3>it means that it can help us solve problems, it's

0:36:00.320 --> 0:36:02.700
<v Speaker 3>also something that helps us find flaws in the software

0:36:02.719 --> 0:36:03.720
<v Speaker 3>that people have already written.

0:36:04.530 --> 0:36:06.390
<v Speaker 1>And Greg, as a follow up, a specific question I

0:36:06.410 --> 0:36:08.009
<v Speaker 1>got for you is how much of the work in

0:36:08.050 --> 0:36:13.310
<v Speaker 1>this case study was literally done by AI versus a researcher?

0:36:14.469 --> 0:36:15.770
<v Speaker 6>So for the Calif side.

0:36:15.830 --> 0:36:17.850
<v Speaker 3>I'd say, I don't know how to put an exact

0:36:17.910 --> 0:36:20.620
<v Speaker 3>percentage on it, but the researcher and their deep expertise,

0:36:20.960 --> 0:36:25.620
<v Speaker 3>let's call it 75%, 80%. Whereas for the OpenAI internal version,

0:36:25.940 --> 0:36:28.980
<v Speaker 3>that was really, we set up an environment and we

0:36:29.160 --> 0:36:31.560
<v Speaker 3>took our most capable model with a lot of compute

0:36:31.580 --> 0:36:34.719
<v Speaker 3>behind it. And that autonomously was able to find the

0:36:34.739 --> 0:36:35.420
<v Speaker 3>full exploit.

0:36:36.640 --> 0:36:38.500
<v Speaker 1>Matthew, I want to go to you on probably the

0:36:38.540 --> 0:36:41.319
<v Speaker 1>scarier bit, which was reading the blogs and the reporting.

0:36:41.940 --> 0:36:46.080
<v Speaker 1>It seems like hackers tried to exploit these flaws before

0:36:46.120 --> 0:36:49.150
<v Speaker 1>they were made public, right? Which raises the question whether

0:36:49.270 --> 0:36:51.890
<v Speaker 1>AI was what found them first. Do you see what

0:36:51.930 --> 0:36:52.110
<v Speaker 1>I mean?

0:36:53.130 --> 0:36:53.650
<v Speaker 6>Absolutely.

0:36:53.710 --> 0:36:55.350
<v Speaker 10>And I think what we have to be clear about

0:36:55.430 --> 0:36:58.330
<v Speaker 10>here is that AI is the real hero of this story.

0:36:58.730 --> 0:37:02.260
<v Speaker 10>These flaws have existed for quite some time. They're extremely serious.

0:37:02.719 --> 0:37:07.960
<v Speaker 10>And PHP, the underlying software that contained the flaw, is

0:37:08.100 --> 0:37:11.880
<v Speaker 10>present and running on something like 70% of the world's websites.

0:37:11.920 --> 0:37:14.410
<v Speaker 6>So this was very significant. And it was a very

0:37:14.450 --> 0:37:15.149
<v Speaker 6>significant flaw.

0:37:15.210 --> 0:37:20.259
<v Speaker 10>It's hard to know whether hackers, nation states otherwise had

0:37:20.520 --> 0:37:24.440
<v Speaker 10>access to this flop that had been present for years.

0:37:25.020 --> 0:37:27.420
<v Speaker 10>And I think that what OpenAI did with their tools

0:37:27.460 --> 0:37:31.570
<v Speaker 10>and what the researchers did was extremely responsible. They found

0:37:31.590 --> 0:37:33.670
<v Speaker 10>a latent bug that was there.

0:37:33.830 --> 0:37:37.290
<v Speaker 6>They found something that may have been exploited for many years.

0:37:37.330 --> 0:37:41.370
<v Speaker 10>They responsibly disclosed it to the industry and companies like Cloudflare.

0:37:41.820 --> 0:37:43.839
<v Speaker 10>And we worked tirelessly behind the scenes to make sure

0:37:43.880 --> 0:37:47.140
<v Speaker 10>that our customers were protected and and that the internet

0:37:47.180 --> 0:37:50.549
<v Speaker 10>was protected before it got out. But it's clear, it

0:37:50.590 --> 0:37:53.890
<v Speaker 10>wasn't that AI created the bug, AI found the bug

0:37:53.930 --> 0:37:56.330
<v Speaker 10>that was there, and it's incredibly important that we've now

0:37:56.410 --> 0:38:00.189
<v Speaker 10>found it, fixed it, and the software is significantly more

0:38:00.250 --> 0:38:02.710
<v Speaker 10>secure because of the work that OpenAI did.

0:38:03.910 --> 0:38:06.200
<v Speaker 1>Greg, to the best of your knowledge, are there any

0:38:06.280 --> 0:38:11.719
<v Speaker 1>examples of hackers using OpenAI's models to try and exploit

0:38:11.800 --> 0:38:13.520
<v Speaker 1>this specific PHP flaw?

0:38:14.790 --> 0:38:19.710
<v Speaker 3>So we believe that attackers are going to start using

0:38:19.989 --> 0:38:23.069
<v Speaker 3>models broadly to try to exploit these flaws. We have

0:38:23.110 --> 0:38:26.310
<v Speaker 3>monitoring on our side watching for it. And I think

0:38:26.350 --> 0:38:28.989
<v Speaker 3>we're starting to see attempts, or we have seen attempts

0:38:29.010 --> 0:38:32.089
<v Speaker 3>in the past, but not at massive scale yet. And

0:38:32.110 --> 0:38:35.350
<v Speaker 3>I believe that we're going to see just.

0:38:35.370 --> 0:38:35.870
<v Speaker 6>A ramp up.

0:38:35.890 --> 0:38:39.140
<v Speaker 3>I think that Cloudflare has reported seeing this kind of

0:38:39.219 --> 0:38:42.660
<v Speaker 3>very steep ramp of attempted exploitation. So I think that

0:38:42.700 --> 0:38:45.839
<v Speaker 3>we should expect attackers to try, but as Matthew said,

0:38:46.200 --> 0:38:48.379
<v Speaker 3>that we have put in a lot of effort here

0:38:48.660 --> 0:38:51.120
<v Speaker 3>to do responsible disclosure, to prepare the internet and to

0:38:51.160 --> 0:38:52.260
<v Speaker 3>make this software more secure.

0:38:53.400 --> 0:38:56.120
<v Speaker 1>Matthew, same question to you, sort of knowledge or awareness

0:38:56.140 --> 0:39:00.140
<v Speaker 1>of any specific incidents where open AI models have been

0:39:00.300 --> 0:39:02.080
<v Speaker 1>used or an attempt to be made by hackers to

0:39:02.160 --> 0:39:04.080
<v Speaker 1>use them. But also kind of give me the backstory

0:39:04.100 --> 0:39:07.180
<v Speaker 1>of when Cloudflare came into this, when you got involved.

0:39:08.100 --> 0:39:12.189
<v Speaker 10>So first of all, once the exploit is known, AI

0:39:12.210 --> 0:39:14.350
<v Speaker 10>doesn't have to be used at all. You can actually,

0:39:14.450 --> 0:39:16.860
<v Speaker 10>in this case, with a couple of requests to a

0:39:16.860 --> 0:39:20.080
<v Speaker 10>web server, take over that web server. So once we

0:39:20.160 --> 0:39:24.520
<v Speaker 10>know what the problem is, then that can be exploited

0:39:24.590 --> 0:39:27.150
<v Speaker 10>without using AI at all. And hackers right now are

0:39:27.190 --> 0:39:31.650
<v Speaker 10>exploiting this flaw at just incredible rates. Literally, we're seeing

0:39:31.810 --> 0:39:38.480
<v Speaker 10>millions of attempted attacks using this flaw now across all

0:39:38.500 --> 0:39:42.100
<v Speaker 10>of our network every single minute of every single day.

0:39:42.200 --> 0:39:44.439
<v Speaker 6>So the flaw is now known.

0:39:44.989 --> 0:39:47.850
<v Speaker 10>And if you are running PHP on your servers, you

0:39:47.969 --> 0:39:50.990
<v Speaker 10>need to update PHP to the latest version.

0:39:51.070 --> 0:39:53.130
<v Speaker 6>It's free. You can get it updated immediately.

0:39:53.150 --> 0:39:55.370
<v Speaker 10>And that's the only thing that's going to keep you

0:39:55.450 --> 0:40:01.009
<v Speaker 10>safe from this. What I think, again, was incredibly responsible

0:40:01.190 --> 0:40:03.290
<v Speaker 10>of the OpenAI team. And I think most of the

0:40:03.330 --> 0:40:06.469
<v Speaker 10>frontier labs are working with companies like Cloudflare in this way.

0:40:06.989 --> 0:40:10.020
<v Speaker 10>is that when they found this flaw a couple months ago,

0:40:10.040 --> 0:40:13.439
<v Speaker 10>they reached out to us and they said, we found this.

0:40:13.600 --> 0:40:17.239
<v Speaker 10>They disclosed it to us in a confidential way. We

0:40:17.280 --> 0:40:19.650
<v Speaker 10>were able to work with our team to make sure

0:40:19.670 --> 0:40:22.830
<v Speaker 10>that we had put protections in place so that when

0:40:22.850 --> 0:40:26.070
<v Speaker 10>this was actually released, as much of the internet was

0:40:26.290 --> 0:40:28.840
<v Speaker 10>protected as possible. And they didn't just work with Cloudflare,

0:40:28.880 --> 0:40:32.610
<v Speaker 10>they worked with WordPress, with the PHP community, anyone that

0:40:32.650 --> 0:40:36.810
<v Speaker 10>could actually make as much effort to protect this. And again,

0:40:36.830 --> 0:40:39.360
<v Speaker 10>I think it's really important to say here, AI is

0:40:39.400 --> 0:40:41.779
<v Speaker 10>the hero here. This is a latent flaw that has

0:40:41.800 --> 0:40:44.680
<v Speaker 10>been in the code for a long time. It may

0:40:44.700 --> 0:40:47.740
<v Speaker 10>have been being exposed by and used by nation states

0:40:47.780 --> 0:40:50.320
<v Speaker 10>or other hackers to do all kinds of bad things.

0:40:50.660 --> 0:40:53.120
<v Speaker 6>But now that it's been exposed, we've actually fixed it.

0:40:53.120 --> 0:40:55.750
<v Speaker 10>And what I'm optimistic about is, yeah, there may be

0:40:55.790 --> 0:40:59.279
<v Speaker 10>times when bad guys use AI to find things. But

0:40:59.440 --> 0:41:02.100
<v Speaker 10>right now, we've got the good guys who have more resources,

0:41:02.520 --> 0:41:05.040
<v Speaker 10>have access to great tools like the ones provided by

0:41:05.140 --> 0:41:08.219
<v Speaker 10>OpenAI and the ones provided by companies like Anthropic. And

0:41:08.239 --> 0:41:10.400
<v Speaker 10>we're going through code and we're going to find a

0:41:10.400 --> 0:41:12.160
<v Speaker 10>bunch of flaws. We're going to find a bunch of bugs.

0:41:12.489 --> 0:41:15.090
<v Speaker 10>But I actually think the Internet is getting more secure

0:41:15.510 --> 0:41:17.850
<v Speaker 10>because of the work that companies like OpenAI are doing.

0:41:19.290 --> 0:41:22.250
<v Speaker 1>Greg, you wrote quite an offensive post on X, sort

0:41:22.270 --> 0:41:25.530
<v Speaker 1>of recalling your own history with PHP as a programming language.

0:41:25.870 --> 0:41:28.050
<v Speaker 1>One of the ideas that you talked about was suggesting

0:41:28.090 --> 0:41:31.450
<v Speaker 1>that the governments or I think you call them authorized defenders,

0:41:31.510 --> 0:41:35.090
<v Speaker 1>but they scan the internet for vulnerable systems and in

0:41:35.180 --> 0:41:39.040
<v Speaker 1>some cases just remotely patch them. And I think that

0:41:39.080 --> 0:41:41.779
<v Speaker 1>raises a lot of questions about permission, like who decides

0:41:42.120 --> 0:41:44.100
<v Speaker 1>that that should be the case and where the line

0:41:44.120 --> 0:41:44.660
<v Speaker 1>gets drawn.

0:41:46.840 --> 0:41:48.259
<v Speaker 6>So I think that we are.

0:41:49.239 --> 0:41:51.680
<v Speaker 3>You should look at PHP as one example of the

0:41:51.739 --> 0:41:54.180
<v Speaker 3>kinds of flaws that AI is now able to find.

0:41:54.780 --> 0:41:56.660
<v Speaker 3>And we're in this world where there's what we call

0:41:56.680 --> 0:42:00.080
<v Speaker 3>a defender's window, where frontier models are ahead of what's

0:42:00.120 --> 0:42:02.460
<v Speaker 3>broadly available. And so that means that the labs, those

0:42:02.480 --> 0:42:07.210
<v Speaker 3>who have access to these cutting edge capabilities, can find vulnerabilities,

0:42:07.230 --> 0:42:09.850
<v Speaker 3>can peer into the future before anyone else does. And

0:42:09.870 --> 0:42:11.670
<v Speaker 3>so that means if you are a company and you

0:42:11.730 --> 0:42:14.250
<v Speaker 3>secure yourself using the frontier models, you will be more

0:42:14.290 --> 0:42:16.970
<v Speaker 3>secure and able to stay ahead of what is broadly

0:42:17.010 --> 0:42:18.989
<v Speaker 3>available and what attackers can do. And so I think

0:42:19.010 --> 0:42:22.490
<v Speaker 3>that defender's window is a very important concept for every company,

0:42:22.570 --> 0:42:25.549
<v Speaker 3>for every defender to really think through. But it's also

0:42:25.590 --> 0:42:27.750
<v Speaker 3>very clear we're going to be finding lots and lots

0:42:27.790 --> 0:42:31.270
<v Speaker 3>of flaws that are there in the software, that have

0:42:31.290 --> 0:42:33.140
<v Speaker 3>been there for many years, that are tech debt, that

0:42:33.160 --> 0:42:36.300
<v Speaker 3>have piled up as people have built software on top

0:42:36.580 --> 0:42:42.080
<v Speaker 3>of just many, many years' worth of accumulated tech debt.

0:42:42.600 --> 0:42:45.600
<v Speaker 3>And so to me, the question is, how does the

0:42:45.660 --> 0:42:48.830
<v Speaker 3>internet stay secure? And there is precedence. With, for example,

0:42:49.210 --> 0:42:52.650
<v Speaker 3>as botnets became prevalent, having the FBI get court orders

0:42:52.710 --> 0:42:57.430
<v Speaker 3>to clear out malware from botnets because that's something that's important.

0:42:57.450 --> 0:42:58.640
<v Speaker 3>It's not just the site owner.

0:42:58.580 --> 0:43:00.480
<v Speaker 6>Who is affected if they get attacked.

0:43:00.540 --> 0:43:03.980
<v Speaker 3>It's really the whole internet becomes less robust and secure

0:43:04.020 --> 0:43:06.819
<v Speaker 3>if there isn't someone who's trying to make sure that

0:43:07.239 --> 0:43:12.550
<v Speaker 3>patches are applied or that if there is malicious activity,

0:43:13.030 --> 0:43:15.970
<v Speaker 3>that there is some sort of way of notifying site

0:43:16.090 --> 0:43:19.930
<v Speaker 3>owners and making sure that that improvements are made. And

0:43:19.950 --> 0:43:22.730
<v Speaker 3>so I think that the core question we have, I

0:43:22.750 --> 0:43:24.700
<v Speaker 3>don't think that we have the answers yet, but I

0:43:24.719 --> 0:43:26.759
<v Speaker 3>think that we're really asking the question is, as the

0:43:26.840 --> 0:43:31.520
<v Speaker 3>scale of vulnerabilities found increases, and it's harder and harder

0:43:31.860 --> 0:43:34.380
<v Speaker 3>for anyone to make sure that they're taking all of

0:43:34.400 --> 0:43:38.930
<v Speaker 3>the patches, how does this whole system need to change?

0:43:39.290 --> 0:43:41.109
<v Speaker 3>And so if you look at PHP, we had two

0:43:41.170 --> 0:43:43.850
<v Speaker 3>months worth of very manual process. We were on the

0:43:43.870 --> 0:43:46.649
<v Speaker 3>phone with many of our partners, government to really try

0:43:46.690 --> 0:43:49.850
<v Speaker 3>to land as well. We see other vulnerabilities coming. We've

0:43:50.070 --> 0:43:52.090
<v Speaker 3>been working on several of them as well. And I

0:43:52.130 --> 0:43:56.460
<v Speaker 3>think that the system of making sure that we stay ahead,

0:43:56.820 --> 0:44:00.180
<v Speaker 3>that defenders are advantaged, and that there is some sort

0:44:00.200 --> 0:44:03.170
<v Speaker 3>of process, some sort of societal process for keeping the

0:44:03.210 --> 0:44:06.310
<v Speaker 3>internet secure, that's the question that we think that the

0:44:06.350 --> 0:44:07.130
<v Speaker 3>people should be asking.

0:44:08.230 --> 0:44:10.930
<v Speaker 1>Greg, before I let you go, I've been using DOTS

0:44:11.110 --> 0:44:14.870
<v Speaker 1>a lot. Alongside Muse, Grokbot, I kind of put this

0:44:14.890 --> 0:44:17.790
<v Speaker 1>question to Sam the other day, trying to work out

0:44:17.850 --> 0:44:20.110
<v Speaker 1>how to make an agent useful to me. Is there

0:44:20.140 --> 0:44:22.719
<v Speaker 1>any updated data you can give me on how DOTS

0:44:22.800 --> 0:44:26.360
<v Speaker 1>has gone from an adoption perspective and a user perspective?

0:44:28.900 --> 0:44:31.720
<v Speaker 3>It's really hockey-sticking. I mean, it's really exciting to see

0:44:32.420 --> 0:44:35.560
<v Speaker 3>all the usages. And actually, I think that there's so

0:44:35.620 --> 0:44:39.350
<v Speaker 3>many surprising things across work and also your personal life.

0:44:39.489 --> 0:44:42.810
<v Speaker 3>And even, for example, for protecting people, one of my

0:44:43.090 --> 0:44:48.670
<v Speaker 3>friends told me that her dot noticed a phishing email

0:44:49.090 --> 0:44:52.140
<v Speaker 3>in her inbox said, hey, do not open this email.

0:44:52.160 --> 0:44:54.800
<v Speaker 3>It's a phishing attempt. And her admin also has access

0:44:54.840 --> 0:44:57.700
<v Speaker 3>to her email. And it also pinged her admin saying,

0:44:58.000 --> 0:44:59.799
<v Speaker 3>do not open this email. And so I think that

0:44:59.840 --> 0:45:02.580
<v Speaker 3>we're finding this proactive form factor of an AI that's

0:45:02.620 --> 0:45:05.100
<v Speaker 3>looking out for you in your personal life, in your

0:45:05.140 --> 0:45:09.950
<v Speaker 3>work life, to be something that just helps people save money,

0:45:10.390 --> 0:45:13.870
<v Speaker 3>is able to increase people's productivity. It's something you could

0:45:13.890 --> 0:45:16.009
<v Speaker 3>not get a person to do, to be looking out

0:45:16.050 --> 0:45:18.569
<v Speaker 3>for you constantly. I think that there's just so much

0:45:18.610 --> 0:45:22.170
<v Speaker 3>potential that's being unlocked. I think that we can make

0:45:22.190 --> 0:45:27.330
<v Speaker 3>people more productive, make work more joyful, and also help

0:45:27.350 --> 0:45:29.410
<v Speaker 3>people just live more fulfilling lives.

0:45:31.110 --> 0:45:35.200
<v Speaker 1>Greg Brockman of OpenAI, Matthew Prince, CEO of Cloudflare, Cloudflare

0:45:35.300 --> 0:45:39.080
<v Speaker 1>key to protecting a lot of what's happening in internet infrastructure.

0:45:39.120 --> 0:45:41.480
<v Speaker 1>Thank you both very much. That does it for this

0:45:41.540 --> 0:45:44.740
<v Speaker 1>edition of Bloomberg Tech. Look, it's been an hour dominated

0:45:44.800 --> 0:45:47.880
<v Speaker 1>by earnings stories. In the case of Meta and Tesla

0:45:47.900 --> 0:45:51.520
<v Speaker 1>and IBM, AI spending is a big theme. But in

0:45:51.540 --> 0:45:53.860
<v Speaker 1>the market, I'd say the kind of prevailing sentiment is

0:45:54.020 --> 0:45:58.260
<v Speaker 1>AI concern or AI risk that's out there in equity markets. Also,

0:45:58.280 --> 0:46:02.280
<v Speaker 1>a fascinating discussion of a company keeping its founder CEO

0:46:02.320 --> 0:46:06.290
<v Speaker 1>in place amid a big political story here in California.

0:46:06.370 --> 0:46:08.670
<v Speaker 1>Recap all of that on the podcast. You know where

0:46:08.710 --> 0:46:12.670
<v Speaker 1>to find it online on iHeart, Spotify, Apple, and of course,

0:46:12.930 --> 0:46:15.770
<v Speaker 1>on all the Bloomberg platforms. Thanks for being with us.

0:46:15.850 --> 0:46:17.340
<v Speaker 1>We're almost at the end of the week here in

0:46:17.380 --> 0:46:20.210
<v Speaker 1>San Francisco. This is Bloomberg Tech.