There’s a line drawn in the sand between Web2 and Web3. But it doesn’t need to be that way, according to James Tromans, head of Web3 at the world’s third-largest online data storage facility Google Cloud. Tromans argues that the two are far from mutually exclusive. 

Critics of Big Tech — so-called Web2 companies such as Facebook, Amazon and Troman’s own parent firm Google, which like other Google services including YouTube and Gmail, is part of multinational software conglomerate Alphabet Inc. — argue that the monopoly effect they exert over user-generated content, online social interaction and once-shared internet architecture is exploitative.

In contrast, proponents of the decentralized philosophy underpinning Web3 — a new phase of the internet built around decentralized blockchain technologies, the metaverse, and non-fungible tokens (NFTs) — aim to return ownership over digital production to internet users.

Prior to entering Web3 with Google, Tromans held trading and engineering leadership positions at banking giant Citi. He then built on his academic roots in artificial intelligence — he holds a Ph.D. in computational neuroscience of vision — as a specialist in data engineering and machine learning at Google before taking over at the helm of Web3 operations at Google Cloud in May 2022.

Despite what some might see as a disparity between Google, a Big Tech giant that reported US$76.4 billion in Q2 revenue this year, and decentralized finance, Tromans said Google Cloud aims to provide infrastructure for Web3 developers to build. In this exclusive interview at Token 2049 in Singapore, he spoke to Forkast’s Will Fee about Google Cloud’s blockchain activities, decentralization, and where he sees AI use heading next.

The interview has been edited for clarity and length.

Will Fee: What does the existence of a dedicated Web3 department at Google Cloud mean for the wider Web3 industry?

James Tromans: At Google Cloud, we have a very large startup and early-stage company ecosystem. But beyond that, we also have a very significant enterprise user base. Our move into Web3 is a journey that goes back to 2018. What you’re seeing now is the result of many years of hard work to build our connectivity. That allows us to bridge the gap between the startup ecosystem, large enterprises and financial services organizations such as Chicago Mercantile Exchange, and HSBC. So in terms of Web3 in general, we’re able to play a little bit of a matchmaking exercise, plus a lot more than that as well.

Fee: Beyond that matchmaking exercise, what areas does the department focus on?

Tromans: Unlike some of our competitors, we have a dedicated product and engineering strategy in this space. That includes first-party engineering capabilities from Google, specifically Google Cloud, that we’ve brought to market. That has helped raise our integrity across the whole ecosystem globally. We also work hand in glove with our go-to-market teams and our business development teams in North America, Europe and of course the wider Asia region. 

That helps us have those technical deep conversations with not only startups in this space, but also the well-established players, showing that we speak their language and understand their pain points. Our effort in this space is really to attract these types of businesses to our platform. We have two main sets of products that we’ve been building — which is our blockchain node engine and our data analytics capabilities.

Fee: What first prompted Google and Google Cloud to enter into blockchain development?

Tromans: Any first-mover activity is usually driven by just a few people. Those that see the future maybe before it’s quite happened. That was particularly true in this case. There were a number of people, but two Googlers [Google developers] in particular, who in the early days saw the opportunity to take blockchain data off-chain and index it in a more accessible database. That way it would be more familiar to traditional — let’s say Web2 — software engineers. They’ve been doing that since 2018 with Bitcoin, Dogecoin, Ethereum, and a whole host of different blockchains as a community-led offering. 

By remaining very much in touch with the blockchain community, we were able to onboard those data sets and make them more accessible to the wider group. It wasn’t just Google, which goes to show we’re an open-source company at heart. That’s still ongoing as part of our public data set offering. For context, that data set offering also includes things like Covid data and other data sets. Those are for the common good, things we think are important for people to have access to. Our crypto data sets sit alongside them. 

In addition to that, more recently, we’ve built first party engineered capabilities that are in the same vein. They’re data pipelines built and engineered by Googlers. Therefore they have service-level operations (SLO) and agreements (SLA) that enterprises may want to build on top of. That’s a slightly different value proposition to the community-led offering but no more or less important. It’s all part of one holistic strategy designed to make Web3 more accessible to the average developer.

Fee: Does that Web3 strategy run throughout Google, or is there a divide there with the Google Cloud side?

Tromans: Google product areas are run fairly independently. They’re their own businesses by design. Google is also a very engineering-led organization and we like to go where the customer demand is and what our customers are asking for. We therefore run largely independently from, say, a YouTube or Google Search. Of course, we also talk all the time to make sure that the channels of communication are open. 

But our strategy is to service cloud customers and listen to what those customers are asking us to do. We need to make sure we meet them where they need us to be. That might mean opening a new data center in a particular region or adding a new product. So that’s how we’ve been approaching this. We don’t see that as contradictory to anything that Google as a whole is doing.

Fee: What specifically is Google Cloud able to offer that is otherwise missing from blockchain services?

Tromans: First, there’s the blockchain node engine. Google has data centers in many regions, in many different countries, and we own our own undersea cables in addition to the public Internet. Those things are additive to the Web3 ecosystem. Our node engine makes it easy for an individual developer at home or an enterprise to turn on a blockchain node in one of our locations around the world and run, say, an Ethereum full node or an Ethereum archival node at the click of a button. We manage all the developer operations for that process. 

We also do site reliability engineering for that process. That basically takes all the heavy burden off that individual or that enterprise to focus on the thing that they would rather be doing. So if they’re building a web3 product, maybe running nodes is not their value differentiator. So we can do that in a cost-effective way for them.

That’s something we’ve done for Ethereum. We’re also doing it for Solana and Polygon. So we’re excited by that. We’re slowly adding more chains. We won’t add every chain, but we will look to broaden that adoption. Google Cloud has millions of developers on it that are interested in technology and building to solve business problems. Our main strategy is to make it easier for those folks to use Web3 technology.

Fee: Given your background in AI and current work in Web3, where do you see a crossover between the two fields?

Tromans: In some ways, generative AI has taken the spotlight off Web3 — in a healthy way. It has given it some space to breathe and find its feet and work on new innovations. But I think AI is going to change many areas of how we work, how we do technical innovation. Google’s been an AI-first company for probably ten years. So there’s the question of Web3 for AI. Then there’s AI for Web3. 

In the case of AI for Web3, it will be used in the same way in other areas. For example, helping write better quality code, helping bootstrap smart contracts. You have an idea, you want to get something out on paper. As a software engineer, you sometimes suffer from writer’s block. You don’t know where to start. So having something that can template for you, some code that you can build from, AI can help there. 

It can also help in future to spot security vulnerabilities and common security mistakes. We know that due to the immutable nature of smart contracts and the fact that they’re generally fairly open, they get reviewed very thoroughly and can potentially be exploited. Anything that AI can do in that regard is incredibly important. 

But it goes further than that as well. To step away from the pure blockchain space, I’m a big proponent of different forms of decentralized identity and digital identity. I can imagine a world where we have wallets holding our data with AI agents that sit in your wallet and help you buy things by selectively sharing your private information. Just the right amount to get new goods and services. 

Those AI agents could help us acquire things or negotiate on our behalf. In that way, AI and Web3 can fit very nicely together, particularly as a way to allow Web3 users to own and make use of their own data. But we’re just at the beginning and we’ll need to see where it takes us.

Fee: You mentioned you’re a proponent of decentralization. But you head the Web3 division of one of the world’s biggest Web2 organizations. How are you received by the more die hard areas of the decentralization community?

Tromans: I would suspect Google as a business is not really for diehards who are more about the politics and the philosophy of Web3 than they are the technology. Our strategy is not to try and win everyone around. You’re never going to make everybody happy. Still, I think we’re going to see a continuum of adoption. 

One of the things that Web3 affords all users is the choice to own your own private keys and be your own bank. Would you want to have some sort of multi-party computation approach where it’s you and friends, or you and a cloud provider, that together can secure your wallet? Or do you want to take your wallet, Airgap it and put it on a ledger or some other hardware provider? 

Not everyone will want to own their own private keys. Not everyone will want to be their own bank. But those things give choice. So there are different elements of decentralization. By building for decentralization, first, it means those who need and want it can have it. That doesn’t then preclude us from having elements of centralization where it makes sense at different levels of the stack. For me, the two are not mutually exclusive at all. 

When we stop talking about Web3 and Web2 as separate things and we just talk about the web again, that’ll be when we reach mass adoption. The technology that fades into the background, that’s what provides real utility for individuals and for businesses. It’s no longer just a buzzword or political talking point.