In this episode, Mihir Shah, Head of Data Architecture and Engineering at Fidelity Investments, shares insights into executing a successful data strategy across all your business units, Fidelity Investments movement to the cloud, data liquidity and much more.
In this episode, Mihir Shah, Head of Data Architecture and Engineering at Fidelity Investments, shares insights into executing a successful data strategy across all your business units, Fidelity Investments movement to the cloud, data liquidity and much more.
---------
How you approach data will define what’s possible for your organization. Data engineers, data scientists, application developers, and a host of other data professionals who depend on the Snowflake Data Cloud continue to thrive thanks to a decade of technology breakthroughs. But that journey is only the beginning.
Attend Snowflake Summit 2023 in Las Vegas June 26-29 to learn how to access, build, and monetize data, tools, models, and applications in ways that were previously unimaginable. Enable seamless alignment and collaboration across these crucial functions in the Data Cloud to transform nearly every aspect of your organization.
Learn more and register at www.snowflake.com/summit
Producer: [00:00:00] Hello and welcome to the data cloud podcast. Today's episode features an interview with Mihir Shaw, head of data, architecture and engineering at fidelity investments. Mihir has spent nearly 30 years revolutionizing fidelity investments view on data. From director to vice president to CTO, Mihir has seen it all.
And luckily he shares what he learned along the way with us today. In this episode, he shares insights into executing a successful data strategy across all your business units, fidelity investments, movement to the cloud data, liquidity, and much more. So please enjoy this interview between Mahir Shaw and your host, Steve ham.
Steve Hamm: It's great to have you on the podcast
Mihir Shah: today. Thanks Steve..
Steve Hamm: Now, what is your role in the company? Tell us what you do and what your main goals are in your roles. So
Mihir Shah: I've been, I've been in this company for 27 years. I [00:01:00] have done multiple different roles in almost every single business unit. What I'm focused on today is I'm the head of data, architecture and engineering.
And we created this role about two years ago, where we said that there are certain functions. There are so important for the future that it needs an enterprise focus and enterprise status. It's not a business unit where digital strategy. So we created the obvious one was cybersecurity. That was always there, which is a horizontal across fidelity.
We created a role for head office. And then the, the one that I'm in, which is head of data. So my, in my job, it's my job to plan our entire data strategy across all our business units and actually engineered some of the core components. Okay.
Steve Hamm: So what is your, what is your data strategy across all the business units?
Mihir Shah: So I'll try to be as brief as possible. And, and I think we categorize. Uh, data infrastructure in three [00:02:00] different areas. The first is managing our master. Our strategy. That is that four core entities that drive our business, like customer, client, product employees. We want a single universal ID for all that that goes across all our systems.
If you have that. Phenomenal unifying effect on all aspects of our business. So that's the first part. The second part is our transactional and operational systems. The systems are the databases that drive our day-to-day business and the strategy there is really modern technologies that move to the cloud, but there is no and rationalization, which is very typical of most of the companies of our size.
But the key thing is that for each of our different products, you need a second. Infrastructure to drive the day-to-day operations. We're not trying to merge everything, but one, the big one that we're working on, which actually is very pertinent to this discussion is our focus on data that drives all our insights and analytics.
We [00:03:00] have about 135 different data warehouses, which are on appliances. We have data marts, we have other databases, spreadsheets, et cetera. We are going to eliminate all of that. And create a single data warehouse that cuts across all our business unit. And it's a single source of data for analytics purposes and our data science purposes across fidelity, because it's pretty ambitious and pretty.
Yeah.
Steve Hamm: Yeah. Yeah. When did you launch that new
Mihir Shah: strategy? I realize this two years ago, and I think as I mentioned, our business success depends on. The synergy between the businesses. And the question was if that's that defines a success, why is that data so siloed? Right. And that prompted this whole strategy.
Yeah,
Steve Hamm: that's really cool. Now, a lot of the, the listeners to the podcast, In addition to winning, to like technology strategy, technology advice, they also want leadership and management advice. A lot of people are on the ladder to ever higher positions in their companies. [00:04:00] So if you could describe a little bit, tell us about your career as a business executive and it leader.
What have been the most important management and leadership lessons that you've learned, whether at fidelity or before. And how are you applying them at fidelity?
Mihir Shah: You have Steve. I mean, that's a pretty broad question. And so let me go through a two or three different aspects of it. The first, the first thing.
I think talent matters, right? Everything is about talent and people acquiring talent is a burden, but matching the talent to the right role is absolutely critical. One of the tenants I believe is that everyone is talented at something, right. And if somebody is not doing well in that particular role, they're probably not in the right role.
So one of the key aspects of a leader is to really find the strands and the weaknesses of your talent, and then make sure. The jobs that you assign them to amplify the, or play to their strengths and hide their weaknesses, [00:05:00] spending time on acquiring talent and making sure they have the right content is really, really critical for success.
The second thing I would say is, and this is throughout my career, the teamwork is that teamwork is a strategic differentiator. If you work as a team, And not as an individual, you will be more successful than, than anyone else. So whether you're working. On a team of 10 members squad, or you're working across two very large independent organizations on a common goal.
The success really depends on the outcome. Really depends. The quality of the African depends on how well you work as a team. And I'll give you an example. I mean, I was a CTO of asset management with. Can I draw. And I, one of my teams was an architecture team and I had brilliant architects on the team, but until I came, they were all working independently.
So the only thing I did was the same people got them to work together as a team on every problem. I encourage [00:06:00] them to seek out the best expert on the team and get their inputs. What happened then was just literally magic. Right? I R we made better decisions. We make better designs. We reduce our risk of making a mistake or a bad decision.
So it was a really live example of exact same team before. They were working as individuals after they were working as a team and just the, the, the outcomes are totally different. Right. So those are the two on the people's side. The other one big ones, especially on this particular project that we're working on.
I think one of the other big differentiator. Is that in large organizations like us, organizational boundaries and inability to work effectively across org structure is one of the biggest impediments. So if you can crack the code of how you work across the organization and you will be able to create bigger and better strategies and actually be able to execute on things that people would say is impossible today.
Yeah, [00:07:00] that's
Steve Hamm: probably enough insight. That's that's a really rich, thank you so much. Hey, I see that with the enterprise strategy, you've got cybersecurity. You got somebody charge your car, then you have you in charge of data, but obviously data and cloud overlap tremendously. So give us a little bit of a history here.
When and why did fidelity begin moving its applications and data to the cloud?
Mihir Shah: So I would say that we probably started this year. Seven years ago, eight years ago, even longer. I would presume we had pilots going on and, and, uh, multiple different strategies with different cloud vendors. But what has happened is in the last three years, it is really accelerated.
And because there is a focus on migration to the cloud, and also there's a focus leadership team in place driving this across all of fidelity. So. In the three years, I think we have roughly about 40% of our applications in the cloud. Our goal is to finish the job in the next couple of years [00:08:00] and move entirely to a cloud based infrastructure.
Yeah.
Steve Hamm: Yes. And the data will move to the cloud. To this one, huge
Mihir Shah: during the storm. Of course. Yeah. So, you know, it has to, right. It has to, I think we have learned the hard way that on executing your cloud strategy, people try to move capabilities of functions to the cloud and not think about the data. What happens is when you have shared data environments, if one function was the cloud, the other one doesn't then where do you put the data?
You're cut off. So, so the best way, the best strategy to move the cloud at the enterprise scale is to move your data to the cloud first. Right? Once you're moving the data to the cloud, once you have two copies, one on prem, one on the cloud, both are kept in sync, and then you start moving one by one, the applications.
And if you do that in the reverse order, as. And opposite to what people typically do. You you'll be [00:09:00] successful. Yeah.
Steve Hamm: Yeah. So did you learn this the hard way at fidelity? Or did you map out this strategy and have pretty smooth sailing?
Mihir Shah: No, I think it was pretty intuitive to us data folks that this was the way to go, but there were groups that fidelity and dealer piloting very, very early days where they were trying to move functions first.
And it was a. Function centric strategy versus a data first strategy, and they didn't make substantial progress. It was too slow. And then obviously realize the fact that we need to sequence this the right way and move the data first. Yeah.
Steve Hamm: Uh, data lake seems to be a central piece of your enterprise data strategy.
Now on the podcast, we haven't talked much about data lakes yet, and it would be really great. I think if you could explain it, give us a little quick primmer on data lakes, what role it plays, uh, why it's so important to your.
Mihir Shah: So, so Steve, you're going to be a little and the [00:10:00] audience is going to be a little disappointed because I'm not going to actually go into a definition of data lake.
And there's a reason why I think, I think data lake data, warehouse, lake house, I think the inventing terms to describe things and the term means different things to different people. So what we have done is step back and not use any of these terms because they are. And go back to the first principles, right?
So at the end of the day, if you want to instantiate a data platform for all our reports, insights, KPIs, metrics, AIML modeling, et cetera, what do you need to do? Okay. You need to create a source to consumption data pipeline, right? Which means you identify. All your sources. And then you figure out whether how to get extract data from your sources.
Bring it in a raw format, whether it's through bulk event or change data capture, you want to bring it in the raw format, but that's not [00:11:00] usable by everyone. Maybe usable by some data scientists. So the next step is take that off our Mac, convert that into normalize base tables. And then finally convert that into curated views for end users who are using BI tools or any kind of tools they want to come up with.
So when we lay out the architecture. You will see that it completely makes sense. And we can overlay all the different terms, which are thrown around by the industry. So, so it is a kind of a silly debate, whether a warehouse is better or lake houses better. And I think a lot of this is essentially inventing new terms for just describing fundamental concepts that have already exist.
Right. So, so yes, we have a data lake. We have a data warehouse and we have a lake house,
Steve Hamm: I guess they overlap with each other. Hey, so my sense is the data warehouse that's primarily for structured data. Data lake that's primarily [00:12:00] semi-structured and unstructured. Correct. Do I at least have that? Correct?
Mihir Shah: Well, that's, that's how you have defined your data. Many people define their data lake as where they bring in the raw data and dump it and just store it for historical purposes because that's the primary source. But yes, I think your definition also works in many cases where structured data goes into a warehouse.
Our emails and PDF files, wise, transcripts images, video, et cetera, goes into a data lake, right? Yeah. Yeah. Gotcha.
Steve Hamm: Well, we talked before recently, you talked about this concept of data liquidity. Now this is a term that was coined in a recent article in the Sloan management review, and it sounds really intriguing.
Tell us about.
Mihir Shah: Sure. I think people who have worked with data, I will always use the cliche term, that data as an enterprise asset, it's an asset just like plant and machinery, [00:13:00] like real estate cash, et cetera. So the question was, if data is an asset, how liquid is it? Which means how quickly can you monetize that asset or use that asset in component?
The composite combined it with other things. How quickly can you actually use that as it in a different form? So if take real estate, for example, yes, you have a huge asset in your house, but if it promote, if you want to use that asset to buy a car, you can't really do that very quickly because it's not a liquid asset.
So the same concept, the MIT research has applied to data and say, The data as an asset. Yes. We all agree. But depending on how you architect your data strategy, your data may not be that liquid or may not be usable very quickly. And that's, that's what it meant. So this is the whole goal is to make sure the data assets that you have, you have a good architecture, good governance behind that asset, so you can [00:14:00] use it and repurpose it for any use case that may come in very quickly and monetize it.
So that's the concept of liquidity. And I think we spent a lot of time with the researchers and with our strategy and they actually featured us in their seminal paper, which are published in SMR. Okay. Very cool. Right. Cool.
Steve Hamm: Now I want to focus a little bit on snowflake just for a couple of minutes here.
So when and why did fidelity first connect with snowflake and how did you initially use. The technology. And if you could give us an example or two of just some of the more important applications you're using it for,
Mihir Shah: right. Sure. So all the chief architects at fidelity, we keep in touch with all the startup companies in different startup people system.
So we take trips to Silicon valley because Seattle's DCI even, even been to Berlin and Stockholm with the various focused agenda meetings. Emerging startups. Snowflake was introduced to us on one of these trips many, [00:15:00] many years ago when it was a very small company that just started up, they had a vision, we liked the pitch.
And so we kept an eye on snowflake. It wasn't right for us at that time. But as, as snowflake develop its product roadmap. And then plus as we were talking. Process more and more data. We at one point realized that a particular use case snowflake would be a right platform because we needed massive amounts of compute power.
I can't get into the details of the use case, but essentially it was in our calculating risks for all our, our bond portfolios. We do a trillion calculations a night. And it was a growing platform. So, so we first started using snowflake over there. And then afterwards it's a very focused one particular application.
And then after that, as we were embarking on our overall. Analytics data platform strategy. Two years ago, we looked at Hadoop was something that was already prevalent in the [00:16:00] company. So we went down the path of can we use Hadoop? And we decided, no, we definitely cannot use Hadoop. Doesn't have all the database matters.
Features out of the box and we would have to build all of them. Right. So we needed a, really a database management system and snowflake was already in house. We did a POC with snowflake and went with it. So how
Steve Hamm: many years have you been with like
Mihir Shah: so far for the specific use case? Probably upwards of four or five years.
I think this, this strategy for two years, right. Where very generally we using it for, for this massive use cases. Right. Yeah.
Steve Hamm: Yeah. Can you get into any more detail with any of the new uses you're putting into something that pro here's the problem? Here's the solution. Here's what we're benefiting from.
Yeah,
Mihir Shah: sure. So I think our goal in coming back to this analytics data platform, I mentioned that we had about 130 plus data [00:17:00] warehouses, data marts, et cetera. So one of the problems we have is that in the old model where we have multiple data warehouse, The most common data sets need to be copied over and over again from the core systems to this target places, Navid databases, where they are consumed.
We did not want that. We wanted a single copy of data to be instantiated and multiple use cases. We'll be working on that single set of data. Right? So we didn't want any duplicates. There's always a single copy of data in our warehouse and everybody uses it for different use cases. So the snowflakes data sharing it really makes that possible and makes it very easy.
So for example, you bring in a brokerage transactions into snowflake. And the rest can use it for calculating risk. Finance can use it for calculating our total assets and then the marketing people can use the same data set for calculating customer value or [00:18:00] customer behavior. Even though there are different constituencies, which need the data to be seen in completely different ways, we can operate off the same base copy, which makes our overall ecosystem really simple.
We can help. Paul bunch of feeds, ETL jobs, FTP files, et cetera. And that was definitely one problem that we were able to solve with this technology. That's a great example.
Steve Hamm: Now you mentioned briefly sharing data. The data cloud really enables. All sorts of new things that were hard to do before acquiring data from third parties, sharing data either internally or externally, also monetizing your own data and selling it out in the marketplace.
Talk to us about that. I guess you'd call this a marketplace strategy. What are you doing?
Mihir Shah: Yeah. So I think the first, the first thing, the first goal is to integrate and create a marketplace between our business units. So it's an internal marketplace, right. And that's [00:19:00] what we are doing right now, which is essentially saying there's a single copy of data and multiple business units can, can use it.
So with the data sharing with the marketplace, we are enabling. And internal marketplace and enabling data producers to work with data consumers between our different business units and divisions. So that's the first, first step. The second step that we have just started is, as you mentioned, acquiring data from outside, we started actually with COVID data, which was published by Johns Hopkins.
And what people used to do is actually go to the site seven, 7:00 PM every night, download a CSV for. And then upload that into the data warehouses. What we asked snowflake to do was create a schema using the same. And make it available as a share. So everyone in the industry could use it. So snowflake went ahead and worked with the third party.
Got that done. All we did was got permission to use it. [00:20:00] So that particular schema for COVID data just pops up in our data warehouse without us doing anything. It was a star schema. Yeah. A lot of people using that. So, so that was the first one. And then since then we worked with our vendors facts that definitely Morningstar.
We said, Hey, we are getting hundreds and thousands of FTP files every day. And it will be great for you and great for us if we can get that data in form of a share data share. And so we don't have. Manage the infrastructure. We don't have to manage the latency. So a lot of vendors are now providing data in, in, in, uh, in directly uploaded into snowflake and available to us as a shared.
Yeah, no, that's a great
Steve Hamm: example. Now you mentioned security and things like that. And privacy, I know that you've had a lot of experience with data governance and related topics. How are you protecting end customer's data?
Mihir Shah: Yeah. So. I think our company has been extremely conservative when it comes [00:21:00] to using, and actually even whether it's protecting or even using internally our end customer's data.
So there are significant number of governance steps that are involved in making sure that before anybody can get access to or can the building of particular data science model, they actually have to go through a number of. And get permission to do that. Some of the steps are very mechanical in the sense that does this person have access to PII data or not.
So those policies where we are digitizing those policies and building it into the access layer of our, of our platform, right? So you don't have to go to the manual steps. We know who you are, and you're, you're about to access something. And the query that you fire we'll actually, we'll be monitoring.
Based on the policies that you're in, in your entitlement. So, so a lot of stuff we have built into our data access layer. The the most common policies around data access is already built in. [00:22:00] The second thing we have done is we have hidden all the PII data, right? So by default, it's not even available on our data warehouse.
You have to go through a separate step to access via data. So if a data scientist, you don't need to know my name, you don't need to know my social security number. That data is actually useless, right. So why even expose that data. So just eliminate it completely hide it. And if you do need it for.
Specific reason then you have to stay. There is go through the steps with our risk and compliance and legal team. And then we'll enable the access. There are some building controls because it's a shared environment, but there is only one copy. The good thing in this architecture. The owner of the data always have, has control of the data.
So even if it's not BI data of customer data, any data sets, there's always an owner and any consumer who wants to use that data first has to take permission from the owner. It's a quick electronic consent. [00:23:00] It's still a step in wall where the owner of the data has to give permission. So we are building a whole bunch of controls into the environment, and then there are standard data governance functions that are data ops functions, where in parallel to building this particular platform, each business unit, we have stood up data governance teams on the business side.
Which take care of defining the data, cataloging all the data, defining policies, the whole decision making process, data quality issues. So we have very specific functions in these business units that take care of some of the governance functions.
Steve Hamm: Okay. Very good. Very good. I'm feeling pretty secure right now.
Want to talk a little bit about the future of technology looking ahead over the coming year or so, what are the biggest trends that you see in managing and analyzing data?
Mihir Shah: So the couple of, couple of things I'll point out one. When you talk about enterprise data [00:24:00] architecture in the past, we are, our thinking was limited to.
The boundaries of your particular company. So if my job was enterprise data architecture, I would think about what's the data attraction for fidelity. And I would stop at the boundaries of fidelity. I think enterprise data architecture now means you need to cover your entire EcoSys. Not just fidelity your vendors, your partners, your institutional clients, and all the other third parties that you deal with.
And you had to build your architecture in context of all the external interfaces that you have and all the relationships that you have, not just within the four walls of the company. I think that's the big change that cloud has brought around and things like marketplaces and data sharing. Those are all.
Things can, that can be leverage leverage when we talk about the entire ecosystem and not just about a single company. So that's the one big change. I think the second big one, I think is there's going to be [00:25:00] many, many, many different words called data marketplaces, private or public with the ready to use schemers.
So we talked about the COVID data. The us government has a treasure trove of data. It's all accessible. It's not secret as a public, as a taxpayer. You can actually access that data is available to public, but it's not leveraged, right? Because it's very difficult to consume. There's a huge opportunity for taking these datasets and making them available to the public in an easy, simpler way.
If you look at healthcare, you have now a healthcare industry standard data model, which is called fire FHI. So that data consumption and sharing becomes easier across healthcare entities, right? Provider payer, et cetera. They all now are able to share data to these standard industry standard data models.
There's no reason why we cannot extend that into multiple different vertical industries to [00:26:00] enable marketplaces. The other thing I think is all these years, data engineering was almost like a subset of software engineering. It's almost emerging as a separate discipline. And a separate specialty. So it has a lot of impact on how you acquire talent, your college curriculums.
It there's going to be more and more specialization around data engineering versus software engineering, because at the end of the day, software is software, but data is not software. Data is data. You use software to manage. But data management is a very different specialization.
Steve Hamm: So it's understanding the nature of that because of data.
And also a lot of math rather than a
Mihir Shah: lot of coding, right? There is a lot of mapping, lyrical and general principles of managing data. Data sets are governed by. Ashley algebra and some maps, even, even how we organized it as managed by mats. So, so there's definitely a need for specialization in, as we go into the future.
I see a [00:27:00] lot of it happening right now. There's a lot of articles being published around data engineers and data engineering specialty. And I see a lot of ads for data engineers, right. So yes, some of them, I hope that works for you. Yeah. Yeah, no, that's great.
Steve Hamm: Fascinating, modern. We
Mihir Shah: live it.
Steve Hamm: It's just what the
Mihir Shah: future holds.
Steve Hamm: I've been asking you to put on your visionary cap now and look further in the future five years or more, some of your answers to the first question, actually I think do go out several years, but look out five years or more. How do you think data will help transform businesses since
Mihir Shah: this.
Oh, I'm not a futurist, but they may take a guess. I think there's this first big area. Think about all the different things that we do with gender data and we have problems. So healthcare is definitely one of them. If you look at healthcare, I think it's going to get completely transformed. [00:28:00] Both the patient.
And the actual practice of medicine will get transformed by data. I know that multiple companies have tried before and we need to continue to try it because it's a healthcare can be truly transformative data and it's going to happen. So that's one area. I think, I think the other one is pretty obvious.
A lot of mundane jobs will be automated through to automation using AI ML models. And if you're seeing that trend happening very, very, very, very rapidly in, in middle offices and back offices, where instead of making binary decisions, if then else kind of positions software is able to make decisions based on probability and statistics, which when you're.
For the trade in, right? So I have a trade coming in. Should I send it to London? Should I send it to New York? What is the venue, which is the trading partner. I want to choose for this particular trip. All these things are getting automated to [00:29:00] statistical methods than actually simple rule-based methods.
And then the other thing I just mentioned, I think there is, uh, there is a huge treasure trove of data in. Government organizations, whether it's us government, whether it's India or take anywhere in the world. So in order to understand our environment, better poverty and societal issues, we need better than better models to predict the outcomes of government and social programs.
We need to be able to. Spend our money, public money, more wisely. And I think as we get a handle on all the data that is available out there and better methods to capture data, I think we'll be able to use our public money much more effectively. Uh, mental program, the social programs,
Steve Hamm: you're talking about massively complicated simulations with so many different factors in them.
To
Mihir Shah: some extent, Steve, I think I'm going back a couple of years, but if, if there is a grant [00:30:00] given by United nations, a massive grants and billions of dollars for. Country. They also have inspectors who are measuring the outcomes and they are, they're all manual and they're all models. I would say that those are ripe for a very data centric approach and which will give us much better results and help us deploy that those assets much more effectively.
That's great information. There's a lot more to hold there. Some people think really need to dig deep and get to know the real you in the rain. A lot of close and personal.
Steve Hamm: We're just about done here. And we'd like to finish on a more personal note kind of lighter. And I understand that you were one of the early online gamers and you're still at it even though.
You got to kind of middle-aged I guess I wouldn't be honest
Mihir Shah: 50.
Steve Hamm: That's not an industry. That's not an insult now. And you're looking forward to being immersed in the metaverse. Tell us about that. [00:31:00] Why do you love gaming so much? And what do you see in the metaverse.
Mihir Shah: Interesting in gaming because of the technology.
I was a technology. I was a software engineer. And when the first multiplayer games came out, which you could play against other players on the office land, for example, I was fascinated by the technology itself. How do the same call the frame so quickly? How did they get the clicks back to back and forth so quickly?
So that's what started it. And I think the gaming industry still fascinates me. If you look at any domain of software engineering where it's, high-speed transaction processing, whether it's modeling or simulation, they are way ahead. And they actually are on the cutting edge. So I've continued to play games.
It's a big stress reliever for me. So if I have a stressful day, I'll go to my gaming machine and shoot a few bad guys. That's what I do now. It's mentioned my age. I do have one request to the gaming industry that you need to have separate servers for 50 [00:32:00] plus people because I ended up competing with the 14 to 22 crowds sometimes.
And you can't keep up. Yeah, they got
Steve Hamm: great reflexes.
Mihir Shah: Yes, absolutely.
Steve Hamm: Now a lot of people are talking about the metaverse these days, and there's probably as much confusion about that as, as there is about data lakes. Right? So what's, what's your view of the better verse? What, what is it and what do you see in it that you think will be most interesting to
Mihir Shah: you?
Well, I think if you put your getting head on it, it's a combination of how you enhance that gaming experience. To some extent, then create a virtual world is more holistic than the two dimensional gaming that we do today. I mean, just giving one example, which is gaming. It's essentially a virtual world where you can only play games, but then many people live their lives on Facebook.
And this becomes another way of social interaction, finding out more information in [00:33:00] a positive way. And so we'll see how it emerges, but I saw the real estate prices on the metaphors are really. We'll see if it's
Steve Hamm: another bubble as if we need another bubble in this whole,
Mihir Shah: probably another bubble, but I see how that emerges.
But to me, I think more than the metaphors, the technology is fascinating. It's a combination of. The physical, the VR lenses and in the physical world and bits and bytes on a computer house, how quickly we can simulate stuff. So,
Steve Hamm: yeah, you know, it's, it's really interesting. I just ordered a, a new MacBook pro.
This is an max and this is like a super cool. 10 years ago, but I realized I do it for a film editing, but it's feel better. They get it to gameplay. I mean, it's like both, it's two ends of the universe, but that's what you need all that horse power for. And I think, you know, it's driving the chip business to.
Mihir Shah: Very interesting. [00:34:00] Right. And I think if you want to bet on the metaphors, you probably want to bet on some of the chip companies use the GPS. Yeah. Yeah. Interesting.
Steve Hamm: Interesting. Well, this has been a fun conversation and a really interesting conversation, really, a lot of depth to it. When you talked about the, your enterprise data strategy and the role that cloud is playing in that, and that it's not just your enterprise, but you see this.
Your whole ecosystem being, being part of it. That's I think that's the most expansive vision that I've actually heard described by anybody on the podcast or anywhere else. So I think I really applaud you guys for, for going so big. And I think people be. Just to see how it goes and, and what kind of new capability gives you and also your customers and each of those businesses.
So well done.
Mihir Shah: Thank you, Stephen. Thank you for hosting. This podcast
Producer: joined the world of data collaboration at snowflake summit. This June in Las Vegas at [00:35:00] snowflake summit, you can learn from hundreds of technical data and business experts about what's possible in the data cloud. Learn more and register for snowflake summit at www.snowflake.com/summit.