Cloud AWS re:Invent Coverage | Jerry Chen, Greylock

Clips
News
More from Cloud AWS re:Invent Coverage

Jerry Chen

General Partner

Greylock

Three insights you might have missed from AWS re:Invent

The AWS re:Invent gathering in Las Vegas last week provided several newsworthy announcements, including the cloud giant’s entry into the world of AI f...

Greylock’s Jerry Chen unpacks the shift to AI-native cloud trends

Cloud computing trends are entering a pivotal era as artificial intelligence reshapes how businesses and developers operate. What began as a simple migration of data centers to the cloud has evolved into a dynamic ecosystem of AI-native tools that enhance creativity and efficiency.Developers now wield advanced coding platforms and large language models, enabling rapid innovation, while enterprises integrate AI-driven workflows to unlock new possibilities and transform competitive strategies. This synergy between cloud computing and AI is setting the stage for a transformative leap in technology, according to Jerry Chen (pictured), general partner at Greylock Partners LLC

play_circle_outline Cube's 12th year journey and retrospective

play_circle_outline Unveiling the Evolution and Impact of AI in Enterprise Solutions: The Rise of AI Native Applications

play_circle_outline Unbundling trend in Cloud market

play_circle_outline Potential for building the next Amazon outside of Amazon

play_circle_outline Perception of reinvent event as peak show for Cloud and AI tech

Info
Transcript

Jerry Chen, Greylock

Jerry Chen

General Partner Greylock

The Cube's coverage at Las Vegas has been ongoing for four days now. John Furrier and Jerry Chen discuss the evolution of the Cube over the past 12 years. They highlight the importance of AWS in the Cloud mobile transition and the emergence of AI native applications. They discuss the impact of large language models and ephemeral apps on the development process. They touch on the entrepreneurial opportunities for young developers in the AI space.

They also talk about the trend of former employees of big tech companies starting their own enterprises, f... Read more

explore Keep Exploring

What has the journey of the Cube been like over the past 12 years of coverage at Las Vegas? add

What is AI native enterprise and what are AI native companies striving for in terms of their business models and technologies? add

What are the two ways to make money in tech according to the discussion mentioned in the text? add

What is the potential impact of AI on bundling and unbundling of services in the tech industry? add

What are people's perceptions of the current state of the reinvention of a certain industry or company, particularly in relation to the impact of new leadership, AI technology, and recent developments? add

bolt Powered by CUBE AI

Jerry Chen, Greylock

search

>> Welcome back everyone to the Cube's coverage here at Las Vegas. I'm John Furrier, host of the Cube. We are here for four days of wall-to-wall coverage. It's our 12th year of the Cube. We've seen the journey. It's like a documentary that's playing out right in front of us. We're documenting the whole thing, the Cube's generative process for all the videos. I'm John Furrier with my good friend Jerry Chen, who was here with the Cube on the inaugural broadcast 12 years ago. Just left VMware, just newly minted venture capitals at Greylock at that time has been part of our journey. Good friend of the Cube. Jerry, great to see you-

Jerry Chen

>> John- ...

>> here on the 12th Cube.

Jerry Chen

>> We should do a montage. We age over the 12 years from how it started, how it's going.

>> Yeah, definitely got grey hair for sure. People like, "Hey, you have grey hair now." But it was fun. I was selling the story earlier. This is a reinvent where they're doing a lot of throwback. "Oh yeah. Back in the original day of Amazon...." Have you noticed there's a lot of throwback stories?

Jerry Chen

>> Yeah, yeah.

>> Remember the way it used to be? I was telling the story about how when the Cube started, we had the stream on the whole time.

Jerry Chen

>> Correct.

>> We had no guest list. "Hey, there's James Hamilton, come on up." He's like, "What?" And then sits down and I exposes the entire back end and he was never on the Cube again. They had a PR department at that point. You were there, you came on and it was great. A great, that's actually on YouTube right now. People want to check it out, but I think the stuff that we talked about was very relevant today.

Jerry Chen

>> Sure.

>> Your thesis when you started coming out of VMware at Greylock was you made one comment I want to get your reaction to, I'm looking for the next Amazon, when I asked what's your investor? You were like, in a curious way, like "I'm looking for the next big thing." Amazon still is that next big thing, but we were talking about how Cloud scale you were on the early days of seeing that 20 mile stair of what Amazon could turn into and you've been on many other times. Looking back now, first of all, reaction to that and then looking back now.

>>

>> What's been the key moments for you for AWS on this industry as Cloud went mainstream?

Jerry Chen

>> I mean I would say the past 12, 13 years you've been living this Cloud mobile transition, right? The two platforms that came out in 2008, 2007, the iPhone came out and the AWS launched in '08 or something like that, I guess 13 years ago. And so what we got was we were on the right side of history that this move towards Cloud mobile and just rewriting your applications and you say phase one was just move the stuff you had in your data center to the Cloud, right? Phase two is, okay, let's rewrite things for in a Cloud native way. And Amazon's really pioneered what Cloud native means. Now we're seeing what AI native means. And so I think you've seen all the announcements this week and clearly we talked a bunch startups and budget developers and bunch of customers. It's what is AI native enterprise? What's AI native companies as AI or die for both customers and for startups. And so I think we're seeing Amazon try to reinvent themselves to say "We're the AI native Cloud."

>> And make it easier for developers. The core thing Matt Garmin put on his first slide was developers, "I love developers." And he said, "They're near and dear to my heart." And they said, "I got a billion dollars of credits." And I'm like, "I got to make an email to Garmin and get those credits." But that's a game statement right there. They're saying-

Jerry Chen

>> Take page out Microsoft's book, right? Developers.

>> Okay, so now if you're a developer, and I was telling the story last night at a dinner. It was a reporter who's first time at re:Invent. I go, "You got to remember in 2008 when the financial crisis hit, it was kind of Armageddon. You had the Web 2.0 post bubble and then you saw YouTube come out. They got bought by Google and that whole Web 2.0 face, but a lot of companies were kind of hanging around. People were unemployed, they couldn't get gigs. And so people were trying to figure out how to sell cereal in their apartment. Couple developers that became Airbnb. So you had Dropbox, so all this kind of web SaaS players because it was serendipitous and they were experimenting." And so there's a lot of experimentation in the narrative now experiment. And so that brings up the serendipity. So this is the entrepreneurial question for you, Jerry, is that we're in a market where it's kind of transitioning big wave, old wave, transition to a new wave, AI native. There's a lot of serendipity in meeting people, experimenting with stuff and then just huge upside. So the question is, these entrepreneurs that are under 30, say 25 years old, right out of college, what are they building on? And so what do you see? Because when they were 15, they were using Google Docs. Cloud was my dad's Cloud.

Jerry Chen

>> Not your dad's automobile. It's not your dad's Cloud.

>> My dad used the console, it doesn't have autocorrect, can't make comments. So code writing, I bring that up, it's a little bit over the top, but to make the point generationally, what do you see in the next generation?

Jerry Chen

>> I mean I think the way you're going to teach CS and code changes with ai, both the tools from GitHub Copilot to always AI engineering tools like Cursor, Cognition, et cetera, they're all great. We're in a couple called Stack Plates launched, that's thing called Bolt.new amazing product, right? Vesaro V-0. Bolt.new is an incredible way to prototype apps quickly. And I think what you're seeing is two things. One, the tools you have to make yourself a better fast developer are there. So everyone has some new superpowers. Again, we thought Cloud gave you superpowers. Developer tools like Copilot or Cursor or Bolt gives you superpowers. Number two, how you think about apps change. So you have this thing called the LLM, the large language models, video models. You have this-

>> Ephemeral apps.

Jerry Chen

>> Ephemeral apps. So you have this before you had the internet, which is an incredible resource. What that gave you, they gave billions of users, access to information, ease of communication. Now you have all that plus these large language models that do reasoning to help you even more. So nothing ever gets replaced. We just kind of build these layers. Again, the Cloud layer from infrastructure, networking, storage. Now we have everything we built beforehand, these new AI tools to make you faster. And then you have storage, elasticity, what Amazon pioneered, the internet, mobile phones. But now you have reasoning models, access to large language models to understand communication languages. So that puts you in a whole different creative space. It's just when mobile came out, they built like said Airbnb, Uber, you had maps and GPS and the power of your hand. Now you have all that plus these large models. So I say what does AI native mean for least like 25-year-old founders and developers or large enterprises is take everything you have but imagine have this super powerful AI resource to plug into and that just creates the new creative space.

>> Yeah. The other thing I have great insight, the other question I want to ask you is that if you look at, because we used to joke about you, Pete Suncini and Steve Herod as the VMware mafia, and PayPal has a mafia and it is a term to denote people who have worked at big great companies have gone on to do other things, usually build companies or be an investor. If you look at some of the stripes I've been interviewing that are under the age of 30 or around that age group, they all came from these big companies like Facebook Meta, Google, Amazon, Twitter, Uber-

Jerry Chen

>> Uber, yeah.

>> So the people who have built by hand large scale systems. And you know what they're working on enterprise solutions. So you have another wave of personas working on hard enterprise stuff. It used to be when we were growing up, it's like, oh yeah, consumer enterprise lags, enterprise not lag-

Jerry Chen

>> enterprise has always been fun in my book.

>> No, I'm just saying from a trend standpoint, of course we love the Enterprise, we dig it through every day, but they're actually starting companies with the Enterprise land because they walk into someone like a J. P. Morgan Chase or a Gold and say, "Oh, you got identity problems. Well, we built a system over at Uber to do that." And so they're seeing problems. I've never seen that before. I've never seen the pure entrepreneurial, but it comes from their DNA that comes from, it's the pedigree of the entrepreneurs. It's that-

>> . ...

>> where they've come from. What's your view on that? Do you see the same thing or-

Jerry Chen

>> Sure. I would really say especially from systems and infrastructure folks talking about Uber, before that, Meta, they've seen the future or Google, they've seen problems at scale. They've also seen this problem before when they have access to infinite resources, compute stores, networking, and you test village of users, they can see where the rest of the world is. So the old line, the futures here is not evenly distributed. So we'd like to say those folks come out of Uber or whatever, saw the future and they come out these credit companies like Chronosphere or Tim Perl from Uber's data team for example. I think you're seeing these folks come out and say, "Hey, I can solve these enterprise problems, the security scale of observability, now with the things less as I learned inside these big companies." And then AI, like I said, is additional resource and it makes a lot of these apps more approachable. This whole thing, we talk about the system engagement of chat. So chat lick Slack and messaging has always been around, but was never easy to scale because you had to have a human other side. But now with these large language models, chat is easy and working voice is working for the first time ever before it was kind of working. And so I would say the usability of making enterprise apps because you have these large lives models has changed. So the system engagement is still chat system of action. These agents now are actually working. And so it's kind of cool because before Lala's young developers were like, I don't want to touch enterprise. Have you seen ServiceNow or Oracles UI or Concurs? They don't want to touch that, but holy crap, I can now make all that stuff sexy again because A, design aesthetic, B, mobile, and C, can use things like chat or voice to make these things approachable. I can make software human-like with Ai.

>> And so make it better. So let's talk about your investment thesis. What are you seeing on the investment side? Obviously rounds are with AI is pretty hyped up right now. What's the market from a VC perspective as you look at trying to get the early stage action?

Jerry Chen

>> Yeah, I mean look, we're always trying to be the first mighty in seed series A. We want to back the founders right when they leave the companies or coming out of school and they have this kind crazy idea, all things, the valuations, every VC is going to about valuations all the time. So it won't touch that with the 10-foot pole because it's never going to be the right price until it is, until it's a $100 billion company. Then any price is the right price. But I think we're just seeing, it's just the enthusiasm of this new platform shift, I think, like I said '04, '05 was kind Web 2.0, then '08, '09, '10 kind of Cloud mobile, and we saw a whole bunch of folks invest in native mobile experiences like Instagram, native mobile photo sharing or native mobile databases in the Cloud. Now we're seeing these AI native applications. And so I think you're seeing that kind of enthusiasm we did around that platform shift around Cloud mobile now apply to AI. It's not really a platform shift yet because the way you consume software is still the same. We're playing around with wearables and glasses and goggles, but the way you consume software is still the same. But you see that same enthusiasm of like, "Hey, I could rewrite everything for the iPhone. I can write everything for the Cloud. I could rewrite everything for using AI now." And it's pretty interesting. It's early, but it'll be a race.

>> So I'll get your reaction to that. First of all, good insight there. So I see three areas of AI. AI producers, "I make AI", AI consumers, "I consume the AI", and data suppliers.

>> .

>> Because I mean, everyone's a tight resource. The oil is the resource, whatever. All the foundation models use the same data. I mean they're all kind of pedal as fast as they can, but differentiation will come from data. So do you agree with that? Is there an area I'm missing in terms of categories of who you are? I mean, obviously you look at the players. Are you either a provider, consumer-

Jerry Chen

>> Okay. A provider, is what in your mind?

>> Someone who's building AI technology to people. And I would be a consumer. I use OpenAI or whatever. I use some rag that's an application that's not really AI, but in terms of my mind, but data suppliers on, "Hey, I got unique data that could add to your corpus." So OpenAI would be a provider.

Jerry Chen

>> I think definitely everyone's going to be a consumer. It would either be visible or invisible, but everyone's going to be consumer, just like everyone's a consumer with the internet at some form or fashion. So I think it's going to be there, but it's be ubiquitous. Providers are going to change. You have these pure play folks like OpenAI, you have everything that's the stack of AI tools that's still settling out. We don't know. index building agents,-

>> Picks and shovels,

Jerry Chen

>> Picks and shovels, still letting build agenetic systems, knowledge systems for your consumers. And then for sure data is ubiquitous, right? Again, it's both for pre-training, it's for post-training for evaluations. What I mean by that is when you have your AI doctor, you're going to have someone both get the training for the model, but also make sure the doctors say the model's working. Yay or nay, right? Evals, right? Evaluating the output of the models. And so I would say it's beyond our PhD knowledge for you and I, since we have zero PhDs between the two of us. But there's argument that we're running out data for the pre-training, things are moving to post-training, and then things are move to of this evaluation. So high value evals, but what I mean is lawyers, doctors, accountants. So we do compliance, Sarbones, Oxley, legal. You want to have specialists saying the model's working, not working or the app's working, not working. So I think you're see a lot of the data emphasis, John, move to that domain. Is the model doing what it's supposed to for healthcare, for legal, for technology. And coding is the first, right?

>> It's like process management, getting the workflows, automate those.

Jerry Chen

>> There's expertise. So there's the bunch of horizontal stuff that will figure out coding is the one we talked about, does the code work or not? So if it's binary like that, yes, the code ran or didn't run, that's easy, when it's a little fuzzier, like, "Oh, is this AI-generated image good or bad? Is AI-generated song good or bad?" Who knows? Right? And that's a whole new set of consumer AI experiences that we're excited about. Is this AI experience interesting or not? Who's to tell? And then on these enterprise use cases like legal, healthcare, compliance, are you doing what you expect? Right? And who's going to know if an AI lawyer is good except a real lawyer? And so you're going to see a lot of, I think, emphasis on the data move towards the outputs of ai, not just the inputs.

>> So 12 years ago I asked you what are you investing in? You said the next AWS-

Jerry Chen

>> Next Amazon. Yes.

>> Well, it's here. It's called Amazon Web Services. What's your view of where they're at compared to some of the Cloud competitors? Google's got a nice Cloud they're building, they've got some good AI going on as well, and Amazon's playing the long game, get Microsoft doing their thing. What's your view of the current Cloud market?

Jerry Chen

>> I think we talked about this a couple said there's only two ways to make money in tech, bundling and unbundling. And you can argue that Amazon was the great bundling of everything into the Cloud, compute, storage, networking. Now AI stuff into the Cloud. You can argue then for a while we saw this unbundling and people apply the consumer like the unbundling and crisis list. Every category crisis list is a startup. But you say for a while we're unbundling all the Cloud guys to databases, snowflake, Databricks, observability, Chronosphere, Data Dog know applications. And so you say we're unbundling all the pieces of Amazon for a bit. And that was a investment thesis that we were looking at Greylock, investing into the different little categories of Amazon unbundling. So the question to your point is with AI, is that going to be unbundling? So it's OpenAI, Ethropic, these other models, a component of a bundle, AI Cloud or a separate business in itself, or these AI apps, are they going to be part of a bundle or unbundled and consumers want to buy and everything piecemeal? And so-

>> That's a good question.

Jerry Chen

>> That's a great question. And I think you see Amazon, Azure, Google bundling a bunch of services together and they say the AI stuff should be tied to your storage, to your networking. Lot of great reasons for that as well. But then you're seeing GPU Clouds like Crusoe, Lambda, guys like Together or Fireworks or a company we're involved in called Baseten that does know inference for your models and they work on all the Clouds, like Baseten's on Google, Amazon, I think Oracle, a bunch of guys. So I think you might see some unbundling services because of AI, because it's very different than what you had before. These guys don't go away, but I think there's a great opportunity to build the next Amazon, maybe outside Amazon, maybe on Amazon. Amazon-

>> Specialty Clouds could be a nice thing.

Jerry Chen

>> It's either, one of two things. One, what we call BYOC, bring your own Cloud. We're a company called WarpStream that confidently acquired BYOC or private Cloud, right? It's just Amazon's so darn big now, really on-premise, we're going to be in your VDC, like your Cloud, your keys, your bucket. That's pretty damn interesting. But then to your point, there might be some specialty Clouds out there just because the gravity has moved towards this whole new stack. And look, again, nothing ever goes away, but we're bullish that there's a bunch of new companies started in this unbundling phase. And look at Snowflake and Databricks. Two great companies that are unbundling parts of the data and they have a real chance to create gravity well around them.

>> Yeah, great, great insight. Final question before we break. What's the hallway conversations like? We're out both seeing each other at the different events, A lot of hallway action sidebars. What's the scuttlebutt? What are you hearing? What's your ear in the ground tell you?

Jerry Chen

>> I think people feel like this reinvent is probably, I don't know if the intent as large as maybe the peak levels, right? So post-COVID, we were here, the has dropped, it's been building back up, but we think both Matt Garman joining with new CEO and with AI and a lot of the moves that are moving, plus this crowd, people feel like, holy cow, it feels like it's the peak show again, right? And before it was a couple of years where who knows what was going on, it was going back, but for many years it still is. It's the center of Cloud tech and potentially AI tech now. I think people feel that way again.

>> Yeah, I would agree with that. I also add that I've heard a lot of first timers coming, first time at re:Invent. Really welcome to the Jungle, baby. Jerry, great to see you. Thanks for coming on, wrapping up our day. Appreciate you. Thanks.

Jerry Chen

>> I appreciate you, John.

>> All right.

Jerry Chen

>> It's always a pleasure to be here. And hey, congratulations to you, the Cube, the whole business you put together. It's pretty freaking amazing, man.

>> Thank you very much.

Jerry Chen

>> Congratulations.

>> Thank you, Jerry Chen, part of our journey from day one, 12 years ago here at re:Invent, 15 years with the Cube. And again, we'll continue. We're just doing our part, just documenting history, sharing the data with you. Soon we'll have a generative agent technology to cut it all up for you.

Jerry Chen

>> Next year, your agent and my agent, we'll do an interview with our virtual agents.

>> I'll have my agent, call your agent. My sub agent, multi-agent collaboration, big part of the story. Thanks for watching. See you next time.