WEBVTT - Intel CEO Pat Gelsinger Talks Earnings, AI, Elections

0:00:02.440 --> 0:00:08.280
<v Speaker 1>Bloomberg Audio Studios, Podcasts, radio news. We are joined by

0:00:08.320 --> 0:00:11.959
<v Speaker 1>the Intel CEO, Pat Gelsinger, who's come straight from the

0:00:12.000 --> 0:00:15.680
<v Speaker 1>company's call with analysts. Kindly you are joining us live

0:00:15.680 --> 0:00:18.799
<v Speaker 1>on Bloomberg Television and Radio, and Pat, thank you for

0:00:18.880 --> 0:00:21.320
<v Speaker 1>your time in the quarter gone. We were just talking

0:00:21.320 --> 0:00:24.560
<v Speaker 1>about how you made painful decisions and those showed up

0:00:24.560 --> 0:00:28.120
<v Speaker 1>as impairment charges and were reflected on the bottom line.

0:00:28.520 --> 0:00:30.920
<v Speaker 1>To start, could I ask if that's it now that

0:00:31.000 --> 0:00:33.440
<v Speaker 1>those actions are taken and you now have line of

0:00:33.520 --> 0:00:35.479
<v Speaker 1>sight on what you want to do, or if there

0:00:35.479 --> 0:00:37.680
<v Speaker 1>will be more sort of restructure to come.

0:00:39.120 --> 0:00:42.400
<v Speaker 2>Yeah, we worked very hard this quarter to get this

0:00:42.560 --> 0:00:47.640
<v Speaker 2>done and the people actions the restructuring chargers you'll largely

0:00:47.680 --> 0:00:51.240
<v Speaker 2>finished this quarter, So it was a challenging quarter that way,

0:00:51.560 --> 0:00:54.880
<v Speaker 2>but to also deliver better than expect the results if

0:00:54.880 --> 0:00:58.319
<v Speaker 2>we eliminate those one time chargers and to take our

0:00:58.360 --> 0:01:00.960
<v Speaker 2>guidance up for Q four. I'm very proud of our

0:01:01.000 --> 0:01:03.080
<v Speaker 2>team being able to do both of those this quarter.

0:01:03.200 --> 0:01:06.600
<v Speaker 2>Some of the most difficult restructuring charges maybe in the

0:01:06.640 --> 0:01:10.360
<v Speaker 2>history of Intel since our memory microprocessor decision almost forty

0:01:10.400 --> 0:01:13.880
<v Speaker 2>years ago. Getting that done and still beating results and

0:01:13.920 --> 0:01:17.520
<v Speaker 2>taking the guide up. This was a very significant quarter

0:01:17.600 --> 0:01:23.400
<v Speaker 2>for our team to execute, very proud of their discipline results, pat.

0:01:23.200 --> 0:01:26.000
<v Speaker 1>The investor base of cheering you. I'm going on the

0:01:26.040 --> 0:01:28.800
<v Speaker 1>reaction of your stock in after hours, but a lot

0:01:28.800 --> 0:01:30.800
<v Speaker 1>of that does seem to be a focus on what

0:01:30.880 --> 0:01:33.280
<v Speaker 1>you were guiding to in the final three months of

0:01:33.319 --> 0:01:36.319
<v Speaker 1>this year, away from sort of the financials and numbers,

0:01:36.360 --> 0:01:39.160
<v Speaker 1>Could you talk a little bit about the product business

0:01:39.200 --> 0:01:42.319
<v Speaker 1>and the foundry business that gave you some confidence in

0:01:42.319 --> 0:01:43.000
<v Speaker 1>that guide.

0:01:43.959 --> 0:01:46.240
<v Speaker 2>Yeah, And if we take it apart a little bit,

0:01:46.319 --> 0:01:50.720
<v Speaker 2>the client business, We've launched our lunar Lake product, the

0:01:50.760 --> 0:01:55.880
<v Speaker 2>second generation of our AIPC and the AIPC category very solid.

0:01:55.960 --> 0:02:00.120
<v Speaker 2>We'll be launching the commercial version of that into a

0:02:00.160 --> 0:02:03.480
<v Speaker 2>good market, we believe next year for a Windows refresh,

0:02:03.560 --> 0:02:06.960
<v Speaker 2>a corporate refresh cycle, and we start to see more

0:02:07.000 --> 0:02:11.080
<v Speaker 2>strength in the AIPC category overall, with strong set of

0:02:11.200 --> 0:02:15.200
<v Speaker 2>use cases and ISVs bringing new software into the marketplace.

0:02:15.280 --> 0:02:18.440
<v Speaker 2>So we feel good about the client business. Data center

0:02:18.520 --> 0:02:21.560
<v Speaker 2>quarter on quarter a little bit better than people forecast

0:02:21.639 --> 0:02:25.120
<v Speaker 2>for that business, and a stronger product line with our

0:02:25.240 --> 0:02:29.240
<v Speaker 2>Xeon six products now starting to ramp, you know, and

0:02:29.280 --> 0:02:32.840
<v Speaker 2>finally on the Intel foundry. Quarter and quarter growth in

0:02:32.840 --> 0:02:37.000
<v Speaker 2>that area and a strong pipeline of external customers. We

0:02:37.040 --> 0:02:41.359
<v Speaker 2>announced Amazon and two other AT and A customers, some

0:02:41.480 --> 0:02:45.080
<v Speaker 2>additional advanced packaging customers. So we're starting to see that

0:02:45.200 --> 0:02:49.200
<v Speaker 2>business come to life. And that's a long term business

0:02:49.200 --> 0:02:52.520
<v Speaker 2>because when people move designs, it takes several years for

0:02:52.680 --> 0:02:55.400
<v Speaker 2>major designs to move into our foundry. So we've been

0:02:55.440 --> 0:02:57.720
<v Speaker 2>at it now a couple of years, and we're starting

0:02:57.760 --> 0:03:01.239
<v Speaker 2>to see the results of those hard work and getting

0:03:01.240 --> 0:03:04.080
<v Speaker 2>back to competitive process technology with AT and A. So

0:03:04.440 --> 0:03:07.160
<v Speaker 2>a very good quarter of operational results.

0:03:08.120 --> 0:03:10.120
<v Speaker 1>Pat We have a question from our audience, and that

0:03:10.240 --> 0:03:13.600
<v Speaker 1>question comes from Clem DeLong, who's the CEO and founder

0:03:13.600 --> 0:03:16.240
<v Speaker 1>of Hugging Face, and he makes the point that on

0:03:16.280 --> 0:03:20.680
<v Speaker 1>his platform there are many more smaller models now available

0:03:21.040 --> 0:03:23.079
<v Speaker 1>and it's leading him to have a question for you

0:03:23.480 --> 0:03:27.640
<v Speaker 1>about Intel's place in the on device trend we're seeing

0:03:27.639 --> 0:03:30.240
<v Speaker 1>with AI and how you're going to enable more on

0:03:30.320 --> 0:03:31.440
<v Speaker 1>device AI use.

0:03:32.760 --> 0:03:35.720
<v Speaker 2>Well, thank you, Clem, A great question, and a Hugging

0:03:35.760 --> 0:03:38.200
<v Speaker 2>Face a great company, and you know, as you think

0:03:38.240 --> 0:03:40.680
<v Speaker 2>about these models, you know we do see it moving

0:03:40.720 --> 0:03:44.320
<v Speaker 2>from the training phase of you know, trillions of parameters

0:03:44.480 --> 0:03:47.680
<v Speaker 2>to the inferencing phase, and exactly like the questioner says,

0:03:47.760 --> 0:03:50.760
<v Speaker 2>in that inferencing phase, more of that will happen on

0:03:50.920 --> 0:03:54.200
<v Speaker 2>prem for data center customers, and many of those smaller

0:03:54.240 --> 0:03:58.320
<v Speaker 2>models will be retrained with enterprise's own data.

0:03:58.440 --> 0:04:00.760
<v Speaker 3>A lot of those will just run on Xeon. We

0:04:00.800 --> 0:04:01.720
<v Speaker 3>don't need a.

0:04:01.600 --> 0:04:05.800
<v Speaker 2>Special accelerator, or it'll be Zeon plus an accelerator.

0:04:06.160 --> 0:04:08.440
<v Speaker 3>But also then the edge use.

0:04:08.320 --> 0:04:10.840
<v Speaker 2>Cases and AIPC use cases.

0:04:11.360 --> 0:04:11.440
<v Speaker 3>Ed.

0:04:11.560 --> 0:04:14.160
<v Speaker 2>I call it the three laws of the edge, right,

0:04:14.200 --> 0:04:16.760
<v Speaker 2>the laws of physics, right, the speed of light if

0:04:16.760 --> 0:04:18.599
<v Speaker 2>I have to go to the cloud, I had time.

0:04:18.920 --> 0:04:21.960
<v Speaker 2>Your second law is the law of economics, right, it's

0:04:22.000 --> 0:04:25.279
<v Speaker 2>cheaper to run on my device. And finally, the laws

0:04:25.320 --> 0:04:28.599
<v Speaker 2>of the land my data on my device in my

0:04:28.720 --> 0:04:31.320
<v Speaker 2>data center. So we see it moving from a cloud

0:04:31.400 --> 0:04:34.880
<v Speaker 2>training world to an edge and on prem world for

0:04:34.960 --> 0:04:36.040
<v Speaker 2>inferencing for the future.

0:04:37.279 --> 0:04:39.960
<v Speaker 1>Pat I therefore have to ask you about GOUDI. Your

0:04:40.000 --> 0:04:42.920
<v Speaker 1>AI accelerator did not get the traction that you'd hoped,

0:04:42.960 --> 0:04:45.880
<v Speaker 1>and you won't be making the outline target of five

0:04:45.960 --> 0:04:48.720
<v Speaker 1>hundred million dollars in sales for this fiscal year or

0:04:48.760 --> 0:04:52.720
<v Speaker 1>calendar year. Why is it a use case? Issue that

0:04:52.800 --> 0:04:56.080
<v Speaker 1>the data center operators aren't looking at it or something else.

0:04:57.240 --> 0:04:59.960
<v Speaker 2>Yeah, two factors that we pointed out. One is is

0:05:00.279 --> 0:05:02.719
<v Speaker 2>now we have Gouty three coming out, so there's more

0:05:02.880 --> 0:05:07.080
<v Speaker 2>enthusiasm for the next generation product that we're just starting

0:05:07.080 --> 0:05:10.320
<v Speaker 2>to ramp now. The second one is the software and

0:05:10.440 --> 0:05:11.680
<v Speaker 2>the software use cases.

0:05:11.760 --> 0:05:13.960
<v Speaker 3>The AI world has moved very.

0:05:13.800 --> 0:05:18.359
<v Speaker 2>Rapidly and innovating very quickly, and the robustness and the

0:05:18.400 --> 0:05:21.839
<v Speaker 2>completeness of the Gouty software a stack hasn't been as great.

0:05:22.279 --> 0:05:26.640
<v Speaker 2>And it's an accelerator, not a full reprogrammable GPU, so

0:05:26.680 --> 0:05:30.000
<v Speaker 2>that makes some of the software porting and optimizations a

0:05:30.000 --> 0:05:33.480
<v Speaker 2>bit more difficult. Obviously, as we get more of that

0:05:33.600 --> 0:05:37.880
<v Speaker 2>running on Goudy, we solve the software problems. But secondly

0:05:37.920 --> 0:05:41.480
<v Speaker 2>then it's also to move to a complete GPU capability,

0:05:41.600 --> 0:05:44.920
<v Speaker 2>which we'll be doing out in time, and that development

0:05:44.960 --> 0:05:47.080
<v Speaker 2>effort is also well underway, so we feel we have

0:05:47.120 --> 0:05:49.840
<v Speaker 2>a solid position. The world is coming to more on

0:05:49.960 --> 0:05:53.800
<v Speaker 2>prem and enterprise use cases. Like the last question ask

0:05:54.160 --> 0:05:58.080
<v Speaker 2>and Gouty three good early indications from the marketplace and

0:05:58.120 --> 0:06:00.000
<v Speaker 2>its success as we ramp it in twenty five.

0:06:00.960 --> 0:06:03.400
<v Speaker 1>If you're just joining us on Bloomberg television and radio.

0:06:03.600 --> 0:06:08.279
<v Speaker 1>We're live with the Intel CEO Pat Gelsinger. Pat Bloomberg's

0:06:08.320 --> 0:06:12.680
<v Speaker 1>reported that arm and Qualcom, two of your peers, are

0:06:12.680 --> 0:06:17.320
<v Speaker 1>interested in buying some or part of Intel. Can you

0:06:17.360 --> 0:06:20.400
<v Speaker 1>confirm whether or not you received a formal offer from

0:06:20.440 --> 0:06:23.840
<v Speaker 1>them or anyone else, and what Intel's consideration of them

0:06:23.880 --> 0:06:24.200
<v Speaker 1>would be.

0:06:25.640 --> 0:06:29.919
<v Speaker 2>Yeah, And obviously there's been rumors and churn in the industry,

0:06:30.160 --> 0:06:33.000
<v Speaker 2>and I think all of that just indicates how important

0:06:33.200 --> 0:06:38.320
<v Speaker 2>Intel is to the Intel into the industry structure and semiconductors,

0:06:38.320 --> 0:06:40.520
<v Speaker 2>and how important those supply chains are.

0:06:41.080 --> 0:06:41.640
<v Speaker 3>For the world.

0:06:42.080 --> 0:06:45.640
<v Speaker 2>We don't view any of these rumors as particularly emeritus.

0:06:45.680 --> 0:06:50.320
<v Speaker 2>We laid out a strategy, we reaffirmed that with our

0:06:50.360 --> 0:06:53.920
<v Speaker 2>board of directors, and obviously, as we've taken these steps

0:06:54.360 --> 0:06:57.520
<v Speaker 2>this quarter and restructuring and getting our costs where they

0:06:57.560 --> 0:07:00.480
<v Speaker 2>need to be, we have the capacity to go on

0:07:00.520 --> 0:07:05.320
<v Speaker 2>the journey that we've laid out to produce sustainable profitability

0:07:05.560 --> 0:07:09.279
<v Speaker 2>and result in shareholder return. So with that, I'm committed,

0:07:09.320 --> 0:07:11.600
<v Speaker 2>the team's committed, We have the board support for the

0:07:11.600 --> 0:07:14.640
<v Speaker 2>strategy that we've laid out, and Q three shows that's

0:07:14.680 --> 0:07:15.720
<v Speaker 2>exactly what we're going to do.

0:07:16.760 --> 0:07:20.400
<v Speaker 1>I've been trying to help our audience understand Intel's historic

0:07:20.440 --> 0:07:23.560
<v Speaker 1>and current business right the products, but also foundry. And

0:07:23.600 --> 0:07:26.000
<v Speaker 1>what you said was that twenty twenty six is when

0:07:26.040 --> 0:07:29.800
<v Speaker 1>those third party customers start to show up. But could

0:07:29.840 --> 0:07:32.800
<v Speaker 1>you just outline the path to that what you want

0:07:32.840 --> 0:07:35.880
<v Speaker 1>to hold yourself to for next year in developing that

0:07:35.960 --> 0:07:37.880
<v Speaker 1>business as a sort of standalone unit.

0:07:39.080 --> 0:07:42.120
<v Speaker 2>Yeah, and it's a great question ed, and clearly we've

0:07:42.200 --> 0:07:45.320
<v Speaker 2>started to see that already emerge. And even this quarter

0:07:45.400 --> 0:07:49.880
<v Speaker 2>we saw quarter and quarter growth of the external foundry business,

0:07:50.240 --> 0:07:52.520
<v Speaker 2>largely driven by some of those early.

0:07:52.240 --> 0:07:53.560
<v Speaker 3>Packaging design wins.

0:07:53.560 --> 0:07:56.840
<v Speaker 2>So these advanced packages are sort of a faster on ramp,

0:07:57.120 --> 0:08:01.240
<v Speaker 2>but ultimately it's becoming a wayfer foundry for the industry.

0:08:01.440 --> 0:08:05.000
<v Speaker 2>And as we said, three major design wins this quarter,

0:08:05.240 --> 0:08:08.480
<v Speaker 2>you'll one with Amazon and two other compute use cases.

0:08:08.520 --> 0:08:10.880
<v Speaker 2>So we'll be updating, you know, the market as we

0:08:10.960 --> 0:08:14.360
<v Speaker 2>win more of those customers and describing the lifetime deal

0:08:14.440 --> 0:08:17.600
<v Speaker 2>value of those to the marketplace as well. So we'll

0:08:17.600 --> 0:08:21.720
<v Speaker 2>give some of those breadcrumbs along the way. But it

0:08:21.840 --> 0:08:24.960
<v Speaker 2>takes a couple of years for people to move these designs.

0:08:25.000 --> 0:08:27.040
<v Speaker 2>A year to design it, a year to bring it

0:08:27.080 --> 0:08:31.080
<v Speaker 2>into production until it starts to ramp in the marketplace.

0:08:31.480 --> 0:08:34.559
<v Speaker 2>So key milestones for eighteen A that customers are now

0:08:34.600 --> 0:08:40.280
<v Speaker 2>beginning those design cycles, the design tools, the design libraries

0:08:40.360 --> 0:08:43.720
<v Speaker 2>are being moved over to eighteen A, but it really

0:08:43.760 --> 0:08:46.960
<v Speaker 2>isn't until twenty six that those results start to really

0:08:47.000 --> 0:08:50.280
<v Speaker 2>help Intel's products, and then twenty six twenty seven until

0:08:50.280 --> 0:08:53.240
<v Speaker 2>we start to see meaningful volumes externally.

0:08:54.440 --> 0:08:57.240
<v Speaker 1>That way just days away from an election in this country,

0:08:57.679 --> 0:09:00.920
<v Speaker 1>and I'm particularly interested in your assessment planning for what

0:09:01.040 --> 0:09:03.880
<v Speaker 1>might happen to the Chips Act programs. Of course, I

0:09:03.920 --> 0:09:06.520
<v Speaker 1>don't expect you to know what will happen in the election.

0:09:07.240 --> 0:09:10.120
<v Speaker 1>You do not have a crystal ball. But former President

0:09:10.160 --> 0:09:14.440
<v Speaker 1>Trump has kind of made disparaging remarks about President Biden's

0:09:14.480 --> 0:09:16.960
<v Speaker 1>economic agenda. How are you planning for that? Please?

0:09:18.280 --> 0:09:20.839
<v Speaker 3>Yeah, And I'd say maybe three different perspectives.

0:09:20.920 --> 0:09:23.840
<v Speaker 2>Number One, the Chips Act was a bipartisan act and

0:09:24.160 --> 0:09:27.040
<v Speaker 2>with that strong support from both sides of the aisle.

0:09:27.720 --> 0:09:30.440
<v Speaker 2>And with that, we believe the importance of that in

0:09:30.520 --> 0:09:35.320
<v Speaker 2>terms of industrial policy. The semiconductor industry is well supported

0:09:35.960 --> 0:09:41.320
<v Speaker 2>by Congress and both parties. Second, we are disappointed by

0:09:41.640 --> 0:09:45.600
<v Speaker 2>how long and how slow. The dispensing of funds has

0:09:45.640 --> 0:09:49.360
<v Speaker 2>been and well over two years at this point. I've

0:09:49.400 --> 0:09:52.880
<v Speaker 2>invested thirty billion in capital and we've seen zero dollars

0:09:52.880 --> 0:09:55.240
<v Speaker 2>of the Chips grants at this point. So we do

0:09:55.360 --> 0:09:57.960
<v Speaker 2>believe that that has been too slow, and we're somewhat

0:09:58.000 --> 0:10:01.280
<v Speaker 2>frustrated by that, and we've taken that of our financials

0:10:01.320 --> 0:10:04.679
<v Speaker 2>for this year. That said, Gil, there's also the Investment

0:10:04.720 --> 0:10:07.320
<v Speaker 2>tax credit that we are seeing in it is much

0:10:07.440 --> 0:10:10.280
<v Speaker 2>larger than the grant portion of the CHIPSAC and that

0:10:10.400 --> 0:10:13.480
<v Speaker 2>is now law and that will be beneficial. But clearly,

0:10:13.520 --> 0:10:17.360
<v Speaker 2>I think whoever wins the election, the importance of this industry,

0:10:17.440 --> 0:10:21.240
<v Speaker 2>the importance of rebuilding this supply chain of manufacturing is

0:10:21.280 --> 0:10:25.080
<v Speaker 2>well supported. And Intel as a designer, R and D

0:10:25.640 --> 0:10:28.920
<v Speaker 2>and manufacturer is unique in the United States and in

0:10:28.960 --> 0:10:31.640
<v Speaker 2>the overall structure of the industry, and we expect the

0:10:31.679 --> 0:10:36.600
<v Speaker 2>broad support rate of the administration regardless which party wins

0:10:36.920 --> 0:10:37.680
<v Speaker 2>as a result.

0:10:37.880 --> 0:10:41.440
<v Speaker 1>Thank you, Pat. Two quarters ago your CFO told me

0:10:41.920 --> 0:10:44.920
<v Speaker 1>you were two over indexed to CPU, and we know

0:10:44.960 --> 0:10:48.080
<v Speaker 1>what happened with Goudi. You made painful decisions in the

0:10:48.120 --> 0:10:52.000
<v Speaker 1>court to gone reduce headcount by sixteen five hundred. Do

0:10:52.040 --> 0:10:54.760
<v Speaker 1>you now have the plan in place and are you

0:10:55.040 --> 0:10:56.680
<v Speaker 1>confident from here?

0:10:58.280 --> 0:11:00.400
<v Speaker 2>You know, we laid out a clean sheet pro that

0:11:00.480 --> 0:11:03.440
<v Speaker 2>we went through to really benchmark the business. We took

0:11:03.440 --> 0:11:07.360
<v Speaker 2>a conservative view of the funding plan, as we call

0:11:07.400 --> 0:11:10.559
<v Speaker 2>it internally, saying, you know, we're taking a modest growth

0:11:10.559 --> 0:11:13.800
<v Speaker 2>outlook and we're building our cost structure to that very

0:11:14.200 --> 0:11:17.240
<v Speaker 2>modest outlook. But we've also built now a lot of

0:11:17.280 --> 0:11:19.880
<v Speaker 2>optionality to scale up into the business. We have a

0:11:19.920 --> 0:11:24.200
<v Speaker 2>shell ahead where we have capacity that we can scale

0:11:24.240 --> 0:11:29.760
<v Speaker 2>into if necessary, and we've made a smaller, more efficient

0:11:29.920 --> 0:11:34.320
<v Speaker 2>Nimbre company that is able to execute with more agility

0:11:34.400 --> 0:11:37.439
<v Speaker 2>and speed as well. So we feel like we're now

0:11:37.600 --> 0:11:40.640
<v Speaker 2>well set up for the future. This was a challenging

0:11:40.720 --> 0:11:43.800
<v Speaker 2>quarter for me for the company, but we feel like

0:11:43.800 --> 0:11:46.080
<v Speaker 2>we're well set up for the future, and the good

0:11:46.120 --> 0:11:50.319
<v Speaker 2>operational results that exceeded the markets expectations and the increased

0:11:50.360 --> 0:11:52.840
<v Speaker 2>guidance for Q four, we feel like we're now in

0:11:52.840 --> 0:11:53.959
<v Speaker 2>a very solid position.

0:11:55.280 --> 0:11:58.000
<v Speaker 1>Intel CEO Pat Gelsinger, thank you for joining us here

0:11:58.040 --> 0:12:00.560
<v Speaker 1>on this special blobot technology straight from the the school.

0:12:00.840 --> 0:12:01.680
<v Speaker 1>Appreciate your time.

0:12:01.800 --> 0:12:03.000
<v Speaker 3>Thank you. Ed Like always