Baidu Wants Its XiRang Platform To Be The Enabler Of The Metaverse

A thousand people have a thousand different ideas about the metaverse. But most technologists agree that the metaverse is the next chapter of the internet. Such consensus stops short, however, as there are many different ways to think about these “chapters.”

One camp focuses on the method of interactions. Chapter one is the internet of reading text and viewing pictures. Chapter two is the mobile internet of consuming videos and using various apps. Chapter three is the metaverse, a real time, three-dimensional, fully immersive internet experience.

Another camp is centered on the idea of value distribution, and is based on blockchain technology. Chapters one and two are both centralized internet of information, while chapter three, the so-called Web 3.0, is the decentralized internet of value.

Chapter one is read-only web portals plus personal websites and therefore half centralized. Chapter two is read and write (think of blogs and social media), and mostly centralized by major tech platforms. Web 3.0 is read, write and own, where data is owned by the users, facilitated by the magic of the blockchain.

For Chinese search engine giant Baidu Inc., all these different ideas are welcome and accommodated. Unlike other tech leaders such as Meta and ByteDance, which are pushing for their own visions of the metaverse, which I wrote about previously, Baidu is tech path agnostic.

“We can provide technological capabilities in API (Application Programming Interface) and SDK (Software Development Kit) that are needed to build a vibrant metaverse, so that everybody can move ahead and not have to spend energy to do these basic tasks,” Ma Jie, vice president at Baidu who also heads up the firm’s metaverse project called XiRang, told me in a recent interview. “Our attitude is very open.”

Ma clarified a fundamental confusion about Baidu’s XiRang project. Last December, the Beijing-headquartered company held its AI Developer Conference inside a futuristic virtual world created by Baidu called Creator City on the XiRang platform.

Many media reports equated Creator City with XiRang, which has been mistakenly described as a mobile app. In fact, XiRang is an invisible web of technological capabilities that Baidu is developing to support the development of the metaverse. Creator City is just a showcase that Baidu used the XiRang platform to build to demonstrate what XiRang is capable of doing.

Think of XiRang as a SaaS (Software as a Service) provider, but for the metaverse. Developers and content creators can license or buy these capabilities to assist their metaverse projects, whether it’s a metaverse game company that wants to create a new metaverse game, or a metaverse social networking app hoping to develop its product.

XiRang to the metaverse is like Baidu Brain to artificial intelligence. Baidu Brain, the firm’s open AI platform, offers hundreds of core AI capabilities and hundreds of thousands of models for developers. XiRang, similarly, wants to be the enabler to metaverse builders.

How does Baidu plan to make money with this rather laid-back approach? The company is essentially doing the hard and perhaps underappreciated work. The SaaS track in China’s tech market, moreover, is notoriously challenging for a number of idiosyncratic reasons.

Ma’s answer is that reasonable profits are good enough. After all, many of the technical challenges of the metaverse from rendering delay, cloud rendering, virtual reality headset problems like weight and dizziness, limitations to the number of avatars accommodated in one virtual setting, and many many more, require a long time to solve.

Being early, comprehensive (in technological capabilities), and patient could help Baidu. Not being greedy is another plus. But whether Baidu’s metaverse strategy will work out in the end is anyone’s guess. At the very least, the Chinese search engine is unique in its metaverse strategy.

Below is an edited Q&A of our conversation.

Nina Xiang: How did Baidu formulate its metaverse strategy?

Ma Jie: Baidu’s virtual reality team has been working on VR since 2016. At the beginning of the Covid pandemic, we felt it would be helpful to use our VR technology to potentially facilitate large-scale virtual events. We began the XiRang project in 2020, and back then, it was not called the metaverse.

I personally feel that the metaverse could be a hopeful candidate of Web 3.0. It really doesn’t matter what names we use to describe it. We can see the evolution of computing interactions, and there are clearly great opportunities in the next innovation of user interaction and immersive experiences.

What is XiRang trying to do?

It is the infrastructure of the metaverse. Twenty years ago, if you wanted to set up a website, you needed to learn how to purchase servers, how to set up things like software stacks, etc. It could take months to get these things ready.

But now, there are many ready-to-use services, content, and templates to enable one to set up a website very quickly. XiRang wants to do the same thing: to provide these basic technological capabilities to facilitate others to build the metaverse.

You mean like what Roblox is doing?

Not exactly. We want to be like the Amazon Web Services (AWS) for the metaverse. It is closer to the infrastructure layer of today’s computing ecosystem.

XiRang is an invisible platform. Creator City, a virtual world where we held our Developer Conference last year, is really just a showcase of XiRang’s abilities.

We wanted to show developers that they can use XiRang’s tools, software and other capabilities to build their own virtual worlds. Our capabilities include avatar, movements and interactions, natural languages, multimedia display, and many more. We will build other capabilities such as cloud rendering to push the industry forward.

So it’s similar to what Meta is doing? Meta is doing a lot of these things too.

Our positioning might be a bit closer to the infrastructure layer. For example, Meta is building up Horizon suites of products for different purposes. But for us, one Creator City is enough. We want other partners to come in to build their own metaverses to enrich this ecosystem. After all, building virtual worlds is not our core competence.

If XiRang wants to be the AWS for the metaverse, does it mean partners who use XiRang capabilities will need to use Baidu Cloud?

We have an open attitude. XiRang is the enabler of the metaverse, and we won’t ask people to be tied to Baidu Cloud. But we will provide all these more infrastructure type of capabilities and services, from cloud upward and including our AI capabilities too. We also want to help our partners to achieve interoperability and interconnectedness.

So it won’t be a centralized platform like Facebook?

No.

Then, how do you plan to make money?

It is okay for us to make reasonable profits in one place. One business model in the internet era is to support small and medium developers, then hope to have a share of the pie when they grow up and begin to make money. But for bigger companies with more capabilities, to-business model is perhaps more appealing.

We can provide license authorizations, technological cooperation, or do joint investments to achieve reasonable returns. We can use different methods and be flexible. But it may be too early to consider that earlier model to support small and medium developers now.

It sounds like a SaaS business model?

Yes, but it is not purely a Saas business. SaaS is a light business model. But if someone wants to have a customized set up so that they can have more control, we are happy to provide this more comprehensive solution.

It’s been a while since XiRang’s Creator City drew lots of attention last year. Any updates of the platform lately?

Last December, during our Developer Conference inside Creator City, we had 100,000 people attending the meeting at the same time.

Actually, I want to clarify this point. Most virtual worlds in the U.S. can only host under 100 avatars in one virtual setting. That is a huge gap to 100,000 people. My understanding is that many of those 100,000 people sitting in the conference center inside Creator City couldn’t interact with those around them.

Yes, what we meant was that we were able to host 100,000 people on one set of servers. You can think of it as all these 100,000 people were on the same server in the context of gaming.

When people play games nowadays, they often need to choose a server. People on different servers couldn’t interact with each other. What we did was that these 100,000 people were on the “same server” (even though they still need to be hosted on a set of servers) so that they can all interact with each other. We designed our own framework to do this.

How will China’s data privacy and security laws, as well as a stricter regulatory environment, impact the metaverse?

This is where we provide value too. We will eventually be an international operation and we need to do the data compliance work in different jurisdictions. Despite differences in laws and regulations in different countries, there are commonalities too.

We as a technology provider can help with this compliance work as well. This type of work could be a burden for many content creators. But we can learn and build on our expertise in this area as we expand in different countries.

Anything you can share about overseas expansion?

We recently jointly established a company with Meta Media to build a virtual city called YuanBang utilizing our XiRang platform. Blue Focus, which build their virtual universe based on XiRang, is also expanding overseas. We are in discussion with them too on this aspect.

We’ve also talked to tourism bureaus and economic development bureaus of several countries to see if we can bring some of their tourism sights and culture to Chinese consumers via the metaverse.