WEBVTT - Jonathan Ross and Rob Thomas Talk Groq and IBM Partnership

0:00:02.520 --> 0:00:07.080
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news.

0:00:07.520 --> 0:00:10.840
<v Speaker 2>IBM and Grock are announcing a strategic partnership to give

0:00:10.880 --> 0:00:15.760
<v Speaker 2>clients ultra high speed, low latency AI capabilities via Grock's

0:00:15.840 --> 0:00:18.720
<v Speaker 2>inference technology. For more and how this partnership is going

0:00:18.760 --> 0:00:22.279
<v Speaker 2>to provide greater access to the full potential of enterprise AI,

0:00:22.400 --> 0:00:25.160
<v Speaker 2>We're joined by Rob Thomas, Senior Vice president of Software

0:00:25.280 --> 0:00:28.440
<v Speaker 2>and Chief Commercial Officer at IBM, and Jonathan Ross, CEO

0:00:28.560 --> 0:00:31.160
<v Speaker 2>and founder of GROC and Jonathan, I want to start

0:00:31.160 --> 0:00:32.920
<v Speaker 2>with you. You know, the way that I look at this

0:00:33.680 --> 0:00:36.600
<v Speaker 2>is it's a very interesting go to market channel for you,

0:00:36.960 --> 0:00:39.320
<v Speaker 2>a sales channel. Think about all of the clients that

0:00:39.400 --> 0:00:42.760
<v Speaker 2>IBM has and how you've tried to grow the company.

0:00:42.800 --> 0:00:46.159
<v Speaker 2>Explain how people will access LPUS through this or through

0:00:46.200 --> 0:00:47.120
<v Speaker 2>the cloud matrix.

0:00:48.159 --> 0:00:51.600
<v Speaker 1>Absolutely, it's an extraordinary opportunity for both of us. IBM

0:00:51.720 --> 0:00:55.080
<v Speaker 1>is going to have their sellers sell Grock SKEW and

0:00:55.120 --> 0:00:58.400
<v Speaker 1>so now you'll be able to directly access our speed.

0:00:58.920 --> 0:01:01.959
<v Speaker 1>The advantage which is that we offer you could think

0:01:01.960 --> 0:01:05.080
<v Speaker 1>of it a little bit like offering broadband in the

0:01:05.120 --> 0:01:07.640
<v Speaker 1>era where dial up wasn't fully rolled out and people

0:01:07.680 --> 0:01:10.160
<v Speaker 1>were still trying to connect to the internet. Our lpus

0:01:10.200 --> 0:01:13.840
<v Speaker 1>are just significantly faster, but we also keep the cost down.

0:01:13.959 --> 0:01:17.360
<v Speaker 1>Just imagine if you were to offer broadband and you

0:01:17.480 --> 0:01:21.840
<v Speaker 1>charged more per bit of data that was sent over

0:01:21.840 --> 0:01:25.160
<v Speaker 1>the line, it would be on economical Broadband increases a demand.

0:01:25.840 --> 0:01:29.639
<v Speaker 1>With agentic use cases, it's particularly important to reduce the speed.

0:01:29.840 --> 0:01:31.880
<v Speaker 1>You don't want to ask a question, wait ten minutes

0:01:31.959 --> 0:01:34.600
<v Speaker 1>later and come back. You'd rather get the answer under

0:01:34.600 --> 0:01:36.240
<v Speaker 1>a minute.

0:01:36.760 --> 0:01:41.319
<v Speaker 2>Rob Under this arrangement with Jonathan, does IBM make any

0:01:41.319 --> 0:01:45.800
<v Speaker 2>sort of financial investment into GROC or is there some

0:01:45.880 --> 0:01:49.240
<v Speaker 2>kind of sales or revenue split. Explain the economics of

0:01:49.280 --> 0:01:50.560
<v Speaker 2>this deal for you guys.

0:01:50.880 --> 0:01:51.560
<v Speaker 3>Big pictures.

0:01:51.560 --> 0:01:54.880
<v Speaker 4>We have a lot of momentum in AI with Watson X,

0:01:55.000 --> 0:01:58.200
<v Speaker 4>as we said on our earnings last quarter seven and

0:01:58.200 --> 0:02:01.080
<v Speaker 4>a half billion dollars as a book of business, and

0:02:01.280 --> 0:02:04.560
<v Speaker 4>we're trying to solve the client problem of how do.

0:02:04.680 --> 0:02:06.400
<v Speaker 3>They deploy AI faster.

0:02:06.920 --> 0:02:10.280
<v Speaker 4>So this partnership is all about what Jonathan said, which

0:02:10.320 --> 0:02:13.799
<v Speaker 4>is five x performance at twenty percent of the cost.

0:02:14.200 --> 0:02:17.200
<v Speaker 4>We've seen it with Watson X running on GROC and

0:02:17.240 --> 0:02:20.200
<v Speaker 4>so we will be distributing GROC as part of our

0:02:20.240 --> 0:02:23.720
<v Speaker 4>go to market and there's a revenue share as part

0:02:23.760 --> 0:02:26.720
<v Speaker 4>of that. We are really excited because we've seen clients

0:02:26.880 --> 0:02:30.440
<v Speaker 4>already getting an impact to how they're deploying AI because

0:02:30.440 --> 0:02:32.280
<v Speaker 4>of the integration of our technology together.

0:02:33.720 --> 0:02:35.640
<v Speaker 5>Let's talk about that, Rob a little bit more. Because

0:02:35.760 --> 0:02:37.480
<v Speaker 5>you're the man who's in charge of the software business,

0:02:37.480 --> 0:02:40.600
<v Speaker 5>you're also really responsible for the world revenue and profitability

0:02:40.680 --> 0:02:43.440
<v Speaker 5>of your company. So help us understand why grow was

0:02:43.440 --> 0:02:46.320
<v Speaker 5>the obvious choice. How is it helping your clients get

0:02:46.360 --> 0:02:48.040
<v Speaker 5>outs as faster? On the inference side of.

0:02:48.040 --> 0:02:52.480
<v Speaker 4>Things, we looked at every possibility in the market, and

0:02:53.160 --> 0:02:57.120
<v Speaker 4>the clients are looking for significant performance, so some of

0:02:57.160 --> 0:03:00.000
<v Speaker 4>that changes how your call center operates or how you're

0:03:00.080 --> 0:03:04.080
<v Speaker 4>supply chain runs, and then you combine that with a

0:03:04.120 --> 0:03:05.200
<v Speaker 4>fraction of the cost.

0:03:05.600 --> 0:03:07.120
<v Speaker 3>Suddenly the economics makes sense.

0:03:07.400 --> 0:03:10.280
<v Speaker 4>AI does have a cost problem, and we think this

0:03:10.320 --> 0:03:12.959
<v Speaker 4>breaks through that. And IBM we've said we're going to

0:03:13.040 --> 0:03:15.880
<v Speaker 4>drive four and a half billion of productivity by the

0:03:15.960 --> 0:03:19.120
<v Speaker 4>end of this year. That's another example of AI truly

0:03:19.200 --> 0:03:21.480
<v Speaker 4>having an impact. And the number one question I get

0:03:21.480 --> 0:03:23.680
<v Speaker 4>from clients now is how are you doing that at

0:03:23.720 --> 0:03:26.840
<v Speaker 4>IBM and can you help us do that? And we

0:03:26.880 --> 0:03:29.680
<v Speaker 4>think the combination of IBM and GROC can make this

0:03:29.760 --> 0:03:31.440
<v Speaker 4>a reality for any company.

0:03:32.880 --> 0:03:36.080
<v Speaker 5>Well, let's dig into that a little bit now with you, Jonathan,

0:03:36.120 --> 0:03:39.000
<v Speaker 5>because the integration with what's the next orchestraate? What does

0:03:39.000 --> 0:03:41.480
<v Speaker 5>that look like on your side? How does that happen

0:03:41.520 --> 0:03:42.400
<v Speaker 5>and happen seamlessly?

0:03:43.680 --> 0:03:48.840
<v Speaker 1>So the wantsonex API is available for anyone to use today,

0:03:48.880 --> 0:03:52.360
<v Speaker 1>It'll be invisible to most users. It'll simply work. We

0:03:52.400 --> 0:03:55.200
<v Speaker 1>have a compatible API and this is something we've been

0:03:55.200 --> 0:03:58.680
<v Speaker 1>working on. We will also work on some lower level

0:03:58.720 --> 0:04:03.280
<v Speaker 1>integrations with VLM, which is a technology that IBM is

0:04:03.360 --> 0:04:06.680
<v Speaker 1>very deeply involved in. But it should just be transparent.

0:04:06.720 --> 0:04:08.840
<v Speaker 1>You should just get more speed. Just imagine one day

0:04:08.880 --> 0:04:11.560
<v Speaker 1>you come home, you had dialogue and now you have

0:04:11.640 --> 0:04:13.400
<v Speaker 1>broadband in a cost less.

0:04:15.160 --> 0:04:18.719
<v Speaker 2>Rob where's the demand coming from on your side, like IBM,

0:04:18.800 --> 0:04:22.200
<v Speaker 2>Granite or some other agentic workload that they want to

0:04:22.279 --> 0:04:25.679
<v Speaker 2>run using the GROC lpus. Are these public sector names?

0:04:25.720 --> 0:04:29.520
<v Speaker 2>Are they private sector SMEs? I'm trying to understand who

0:04:29.560 --> 0:04:30.480
<v Speaker 2>you're serving with it.

0:04:30.960 --> 0:04:34.080
<v Speaker 4>As often happens, I would say financial services have been

0:04:34.360 --> 0:04:37.240
<v Speaker 4>early adopters. But the thing that has changed in the

0:04:37.240 --> 0:04:40.320
<v Speaker 4>market in the last six months is everything is moving

0:04:40.440 --> 0:04:44.280
<v Speaker 4>to multimodel. We have IBM models that we open source,

0:04:44.320 --> 0:04:47.680
<v Speaker 4>which are the Granite models. We announced the partnership with Anthropic.

0:04:48.120 --> 0:04:50.880
<v Speaker 4>We have a partnership with Mistraw and Lama, just to

0:04:50.960 --> 0:04:53.960
<v Speaker 4>name a few. What is incredible about what Jonathan and

0:04:54.000 --> 0:04:57.840
<v Speaker 4>team have built is any model can run and get

0:04:58.000 --> 0:05:00.040
<v Speaker 4>instant improvement run.

0:05:00.080 --> 0:05:01.279
<v Speaker 3>The LP used from ROCK.

0:05:01.400 --> 0:05:03.839
<v Speaker 4>So I think this is a combination of a multimodel

0:05:03.880 --> 0:05:07.360
<v Speaker 4>world accelerating inference with ROCK.

0:05:07.800 --> 0:05:09.240
<v Speaker 3>I think this is a great combination.

0:05:11.040 --> 0:05:15.920
<v Speaker 2>Jonathan, does this capacityority exist or are you supply constraints still?

0:05:15.920 --> 0:05:17.800
<v Speaker 2>You've got to go out and build it either in

0:05:17.880 --> 0:05:20.039
<v Speaker 2>Saudi Finland here in the States.

0:05:21.240 --> 0:05:23.960
<v Speaker 1>So the entire world is supply constrained, and I would

0:05:24.000 --> 0:05:26.400
<v Speaker 1>actually expect that to continue for at least the next

0:05:26.440 --> 0:05:29.400
<v Speaker 1>five to ten years when it comes to AI. Our

0:05:29.440 --> 0:05:31.719
<v Speaker 1>advantage is that we have a supply chain that actually

0:05:31.839 --> 0:05:35.400
<v Speaker 1>ramps much faster, so customers will be able to come

0:05:35.400 --> 0:05:38.560
<v Speaker 1>to IBM, put in an order and we will be

0:05:38.560 --> 0:05:41.480
<v Speaker 1>able to fulfill that faster than you would be able

0:05:41.480 --> 0:05:45.400
<v Speaker 1>to with other technologies. But the supply constraints that are real,

0:05:45.640 --> 0:05:47.839
<v Speaker 1>and this is another reason to start working with IBM

0:05:47.960 --> 0:05:50.719
<v Speaker 1>sooner the sooner you get access to that capacity, the

0:05:50.760 --> 0:05:52.440
<v Speaker 1>sooner you're going to have it. I can't tell you

0:05:52.480 --> 0:05:55.600
<v Speaker 1>how many startups come to us and other companies come

0:05:55.640 --> 0:05:58.960
<v Speaker 1>to us and they are looking for capacity, because some

0:05:59.000 --> 0:06:01.760
<v Speaker 1>of them are actually growing twenty or even thirty percent

0:06:01.880 --> 0:06:05.839
<v Speaker 1>per week or per month, which is an astronomical growth rate.

0:06:06.400 --> 0:06:10.440
<v Speaker 1>But by approaching us early, we can build to your needs.

0:06:12.080 --> 0:06:14.880
<v Speaker 5>You were just mentioning Rob about all the partnerships you

0:06:14.920 --> 0:06:18.080
<v Speaker 5>have when it comes to llms and the offerings that

0:06:18.120 --> 0:06:21.880
<v Speaker 5>you're intertwining within yours. Will you go to others to

0:06:22.080 --> 0:06:24.719
<v Speaker 5>ensure that inference is as fast as possible or is

0:06:24.720 --> 0:06:26.000
<v Speaker 5>it this exclusive with GROC.

0:06:26.680 --> 0:06:30.240
<v Speaker 4>We are open to working with anybody in the ecosystem

0:06:30.240 --> 0:06:33.720
<v Speaker 4>of AI around what we're doing specifically on the acceleration

0:06:33.800 --> 0:06:36.640
<v Speaker 4>with GROC. We want to lean into this partnership. That's

0:06:36.680 --> 0:06:39.799
<v Speaker 4>why this is the one that we've announced today, because

0:06:40.040 --> 0:06:44.359
<v Speaker 4>we have confidence working together with GROC. As Jonathan mentioned,

0:06:44.400 --> 0:06:46.960
<v Speaker 4>we're also enabling some of the lower level technologies and

0:06:47.000 --> 0:06:50.520
<v Speaker 4>open source like VLM. So this is the right place

0:06:50.560 --> 0:06:52.599
<v Speaker 4>to be when it comes to inference. But when you

0:06:52.600 --> 0:06:55.440
<v Speaker 4>think broadly about what's happened in AI, we have many

0:06:55.440 --> 0:06:59.000
<v Speaker 4>companies working with us on agents. Last week we announced

0:06:59.040 --> 0:07:02.719
<v Speaker 4>SMP Global is now running on Watsonnext Orchestrate as an example.

0:07:02.960 --> 0:07:04.599
<v Speaker 3>So we're always open to new partnerships.

0:07:06.040 --> 0:07:09.400
<v Speaker 5>And let's just talk about Jonathan the go to market

0:07:09.400 --> 0:07:13.080
<v Speaker 5>strategy here of teaming with the age old Juggernau that

0:07:13.280 --> 0:07:18.760
<v Speaker 5>is IBN, that has so many deep relationships across global enterprises.

0:07:19.320 --> 0:07:21.559
<v Speaker 5>But is that how you're going to work this going forward?

0:07:21.640 --> 0:07:24.840
<v Speaker 5>It is teaming up with companies that have those legacy relationships,

0:07:24.920 --> 0:07:26.280
<v Speaker 5>or do you still go out there and win the

0:07:26.320 --> 0:07:27.040
<v Speaker 5>business yourself.

0:07:28.160 --> 0:07:29.680
<v Speaker 1>So I would say this is a peanut butter and

0:07:29.760 --> 0:07:33.360
<v Speaker 1>jelly sort of relationship in the sense that oftentimes when

0:07:33.400 --> 0:07:37.360
<v Speaker 1>we meet with sea level executives, those sea level executives

0:07:37.440 --> 0:07:41.119
<v Speaker 1>turn to their tech teams and ask them to evaluate GROC.

0:07:41.480 --> 0:07:44.680
<v Speaker 1>And I've been in meetings where the CTO did that

0:07:45.240 --> 0:07:48.000
<v Speaker 1>and the response from the person is I already use GROC.

0:07:49.000 --> 0:07:51.560
<v Speaker 1>It's my default for everything. So we already have the

0:07:51.560 --> 0:07:54.480
<v Speaker 1>bottoms up. We have two point three million developers already

0:07:54.480 --> 0:07:57.240
<v Speaker 1>building on us. For comparison, open Aie has four million

0:07:57.880 --> 0:08:01.760
<v Speaker 1>now going to those deep relationships from IBM, and the

0:08:01.800 --> 0:08:04.560
<v Speaker 1>fact that IBM is a trusted partner who's been delivering

0:08:04.600 --> 0:08:07.760
<v Speaker 1>for decades. You put those two together and that's an

0:08:07.800 --> 0:08:09.200
<v Speaker 1>amazing go to market motion.

0:08:11.360 --> 0:08:13.840
<v Speaker 5>Wow, it's been great having you both on to talk

0:08:13.880 --> 0:08:16.000
<v Speaker 5>about the go to market strategy. Jonathan Ross, CEO Grock,

0:08:16.040 --> 0:08:18.880
<v Speaker 5>of course, Rob Thomas of Senior Vice president of Software

0:08:18.960 --> 0:08:21.240
<v Speaker 5>over at IBM. We thank you both very much.

0:08:21.240 --> 0:08:21.640
<v Speaker 3>Indeed,