00:04
all right and now I would like to welcome to the stage our Vice President of America's Partner Sales Hope Galley how's everyone doing woo come on woo
00:23
yeah it's afternoon let's go as Mike mentioned I am the Vice President of channels for Pure Storage and I'm excited to be here today to not only just represent our channel
00:35
but to represent our ecosystem partners and align with Mark Bridges my peer who leads the ecosystem team globally we've heard a lot about innovation and technology and where we're going as a business
00:48
but as Charlie mentioned in his opener there's that elephant in the room that word AI that we all know so well and then you heard Rob talk about what we're doing with Nvidia
00:59
what we've done with them for the last eight years what we're doing with them now so I'm excited to say I have the opportunity to spend time here in a few minutes
01:09
with Jennifer Saint John Foster Jennifer is the global leader of financial services at Nvidia and she is going to talk to us about what her team is doing around AI how are they meeting the needs of the customers
01:24
and how we are partnering together so before I bring Jennifer out let's show a short clip wow we're fancy we got a video and everything I love that video you would expect a great video
02:00
and graphics coming from Nvidia obviously it's a heritage of our business I know I know it's like where it all began and just so everyone knows
02:07
this is a full full circle moment for me Jennifer and I have known each other since we were 12 so that would be 2,002 as you guys can imagine right out of high school right right right out of grade school
02:20
yeah and um Jennifer also when she came I Learned this today to interview to work at Nvidia her uh interview was in the pure office yeah
02:30
here in New York City area yeah so full circle moment like best friends yeah we don't even have an office here in New York so they brought me to the pure office to interview
02:39
rightfully so so so in all seriousness um you've been there seven years now you've had multiple different roles tell us a little bit about your current role and
02:50
and how your team is engaging with your customer today sure yeah as as hope said and thank you for the warm welcome um I lead the Nvidia banking sales team here
03:01
based in New York City I've been at Nvidia for 7 years and in my entirety at Nvidia I have been working with financial services customers here in New York um we're a very lean organization
03:12
I was actually with a large global bank last week who was still asking like how do we engage how do we work directly with Nvidia and so it could be a little bit confusing at times because we go to market through our ecosystem partners
03:25
Pure Storage being a huge partner but what I love today earlier as folks were coming into the room I was I had the pleasure of walking around all these tables
03:33
and I was like hope your partners are our partners like there were so many familiar faces and so um we're a very lean sales organization so my team covering banks for North America
03:43
we only work with 10 customers we only have 10 10 named accounts and the reason being is we work with the builders of AI so you're gonna have customers that build their AI's
03:52
and then you're gonna have customers that buy their AI's and so we're vertically very uh industry verticalized specific out in the field so for example
04:00
we don't even cover insurance companies in the field because insurance companies are a bit of laggers in terms of new technologies but they will likely buy their AIs the same with retail we do have a small retail team
04:11
but retail is likely gonna buy their AIs so we work with the builders and in the field uh we have a a small field sales force um but we have account managers and solution architects and what I love is our solution architects
04:24
are true data scientists we hire true data scientists who are true practitioners they're the ones out there creating these examples training large language models and putting them into inference um
04:35
so it's very very exciting actually when I joined Nvidia in 2018 we were only 33,000 employees and today we're only I think 36,000 employees and that won't change our CEO Jensen is not gonna go hire a sales force of
04:51
you know 50 people to cover bank of America we have two so we'll we'll never be that type of company which is why our ecosystem partners are so important to us
05:00
and so thank you for having me here here today I love it and I love that my son is getting ready to graduate as a data scientist that's what I heard in all that oh very cool
05:09
that's very exciting to know that he has a future and we'll get off my payroll that'll be great um you you talk to our financials every day you've lived in it for 7 years
05:21
what would you say are the top concerns or top problems that they are trying to solve right now yeah and I apologize for my examples being very financial services and banking
05:32
domain specific but that's kind of all I've done here at Nvidia because if you think like why we're so vertical specific an AI
05:38
for medical imaging is very different from algorithms and financial services outside at Nvidia outside of our cloud partners who you're very familiar with outside of consumer internet
05:50
the Elon companies the metas and even kind of the second tier of cloud partners like Core Weave Financial Services is our No. 1 industry and the reason being is that financial services
06:02
every part of the business is built on algorithms and AI is just algorithms and so before it was AI it was HPC accelerated computing for pricing equities Monte Carlo calculating risk and now there's true AI um
06:17
having been here for 7 years we we we say internally at Nvidia that chat GPT was our iPhone moment it really brought a very scientific
06:27
mathematical applied AI research opportunity and commercialized it it brought it unlocked AI and brought it to business owners and and executives for them to actually
06:39
understand AI and be able to vision the potential within the business but I'll tell you so when I first started at Nvidia in 2018 one of my customers was Fidelity
06:49
and the solution architect I was working with um at the time who was based in Boston he again is a true practitioner data scientist wanted to show the power of AI to the executives
07:00
and so he created an example and he took Nvidia's Fidelity 401k plan document several different examples of our 401k plan document and he trained the original NLP Natural language processing
07:14
before it became now large Language Model so the original NLP Bert that was released from Google around 2018 we trained it on our plan document we brought it to Fidelity
07:23
we put together this very simple user interface and we asked it a question and we said what is my max contribution of my 401k and it gave us back an answer and it was the right answer
07:34
so we've been doing this for a very long time but now we've got a an executive audience so that's very exciting I love that but in terms of the challenges so um again working with lots of banks
07:45
but in enterprises in general you have to break down these silos so there's lots of silos within the enterprises there's lots of red tape and governance within banking um and within those silos are
07:56
you know data silos and so unblocking that breaking down those silos being able to um create an enterprise capability an AI factory a single supercomputer that the entire enterprise can use
08:11
a common set of software and AI tools in order to do that and be successful you gotta break down the silos and your data has to be ready so I would say
08:19
those are some of the number one challenges that our customers see today but I will say out of the top 10 banks that we're covering five of them the top five
08:27
either have invested in AI factories or going to an invest in AI factories because the economics to do that in the cloud if you look at Bloomberg GPT from I think three years ago it's a white paper
08:41
it's public to anyone so I'm not saying any confidential information um it it shows their benchmarks of using 512 GPUs in AWS to train Bloomberg GPT over two months do the math it's tens of millions of dollars
08:57
so if if enterprises are serious about AI they're bringing some of their assets and building AI factories on prem there are lots of challenges around data readiness and data silos what's happening next
09:09
and the conversations we're starting to have with our customers is capacity planning for inference at scale and I know that pure just announced uh dynamo inference for KV cash right
09:20
that's gonna be super helpful a great capability inference at scale capacity planning is gonna be a big challenge especially when you look at reasoning models Deepseek was released by High Flyer
09:30
a hedge fund in China around the January time frame um at GTC San Jose in March Jensen our CEO announced Dynamo Triton for inferencing reasoning models
09:42
that was like two and a half months later and he said reasoning models are gonna require 10 x 20 x and soon 100 x compute to be able to do inference so so what I'm hearing is the the combination of breaking down that data silos
09:59
making sure that you have access to it is so key to what you're trying to accomplish with AI solutions for your customers so when we're talking about EDC our enterprise data cloud
10:09
that's completely in alignment with the work that you're doing yeah absolutely that's exciting to hear did everyone hear that haha it's exciting to hear right
10:19
so um so Rob mentioned that we uh worked together for seven eight nine years right developing solutions together
10:28
I think EDC and the announcements that we just talked about through um Dynamo we've LED all into um another a solution together
10:37
yep um anything you want to lead the team with leads around the ecosystem and the pure partnership from your view yeah you know
10:44
it's um it's interesting when we work with enterprises like you're starting to get benchmarks out in the public from our enterprise customers
10:50
a lot of it being um AIs that have helped with productivity internally right so all of the gen AI the large language models the chat GPT functionality kind of like not the sexy AI right
11:01
it's like PDF extraction document insights right but hugely valuable think about um an investment banker who has a 500 page public uh information book
11:13
all things publicly available on Nvidia now trains a large language model to take those 500 pages summarize them down to two pages and then use true gen AI to create five slides for an investment banker
11:26
that's about to walk into anvidia and pitches on an m a MNA it's pretty cool right and so that's happening right now that that's what's coming out in production right now
11:35
looking forward in the conversations we're having is all around uh AI agents and that's super exciting and that's really
11:41
where we're gonna need our partnerships from the ecosystem to help build them for our customers we're gonna need access to to data and fast storage
11:50
because how are you gonna have an AI agent be autonomous think independently reason and make decisions as a JP Morgan Chase investment banker unless it's trained on its own data
12:03
and so it's really this whole data flywheel that Nvidia will walk you through being able to make this open source large language model specific for banking and that's a great opportunity
12:13
but in order to do that we need partners like pure we need partners like our Oems that have built uh AI factory reference architectures with us and we certainly need the partners I see in a head at WWT and CDW
12:25
to help actually bring this to market yeah well Jennifer I there's no doubt there's an exciting road ahead we're in an exciting journey with you I want to take the opportunity to say
12:34
thank you for being here I know you're gonna be here for the mixer and later on this evening so um thank you to you and your team
12:40
for the work that you're doing with us and uh thrilled to have you as a colleague a friend a ecosystem partner and look forward to the road ahead thank you having me thank you
