Latest / Influencer Marketing with Fexingo: Creators, Sponsorships, and Modern Endorsement Deals / How Influencers Are Buying Their Own Data Centers
Transcript
- Lucas: So Luna, I’ve been tracking a trend that honestly sounds like something out of a sci-fi novel. Influencers are starting to buy their own data centers. Luna: Wait — like, actual physical servers in a building? Not just cloud storage? Lucas: Exactly. We’re talking rack-mounted servers, redundant power supplies, the whole enterprise IT package. A few top-tier creators have quietly started investing in private server infrastructure to store their audience data and run their own recommendation engines. Luna: Okay, I need to understand the ‘why’ here. What’s wrong with AWS or Google Cloud? Or just using the platform’s own analytics? Lucas: Great question. The catalyst was the wave of API changes over the past couple of years. TikTok, Instagram, even YouTube have been tightening access to their data. They’ve restricted what third-party tools can pull, and they’ve made it harder for creators to export their follower data. Luna: Right — I remember when they killed off the full Twitter API. That was a mess for researchers and brands. Lucas: Exactly the same dynamic. So for a creator whose entire business depends on knowing their audience — demographics, watch time patterns, engagement by time of day — suddenly relying on the platform’s own dashboards becomes a risk. What if they change the algorithm tomorrow? What if they start charging for data you used to get for free? Luna: So the solution is… build your own data center? That seems extreme. Lucas: It is extreme, but for the top one percent of creators, the math actually pencils out. Let me give you a concrete example. A creator with, say, ten million followers across platforms generates a massive amount of data. Every view, every like, every comment, every click on a link. If you’re running a merchandise line or a subscription product, you’ve got transaction data too. Luna: That’s terabytes of data per year, easily. Lucas: Right. And if you want to run your own recommendation engine — say, to suggest which of your old videos a new subscriber should watch, or to personalize email campaigns — you need low-latency access to that data. Cloud costs can spiral fast when you’re moving that much data around. At a certain scale, buying your own hardware is actually cheaper on a per-terabyte basis. Luna: So what’s the upfront cost? We’re talking about a serious capital expense. Lucas: I spoke with a consultant who helps creators set this up. He said a basic on-premise setup — a few servers, networking gear, a dedicated power and cooling system — runs between two hundred thousand and five hundred thousand dollars. That’s for a colocation facility, not building your own building. But some of the biggest names are spending over a million. Luna: That’s a lot of sponsored posts to pay for that. But I guess if you’re MrBeast, that’s a rounding error. Lucas: And MrBeast is actually a great example. He’s talked about how his team built custom software to manage his video production pipeline. It wouldn’t surprise me if he has his own servers. But the more interesting case is Emma Chamberlain. She’s been very public about wanting to own her audience relationship directly. She launched her own coffee company, her own podcast network — and reportedly, she’s invested in private infrastructure to keep her customer data off third-party platforms. Luna: So the motivation isn’t just cost savings. It’s about data sovereignty and control over the relationship with fans. Lucas: Exactly. And there’s a second layer: algorithm independence. If you’re a creator who relies on TikTok’s For You page for discovery, you’re at the mercy of their algorithm. But if you can build your own recommendation engine using your own data, you can create a separate discovery layer — maybe a custom app or a website that recommends your content based on what you know about the user, not what TikTok decides. Luna: That sounds like a huge technical challenge. Most creators aren’t exactly known for their IT departments. Lucas: They’re not building it themselves. They’re hiring. There’s a whole cottage industry now of ‘creator infrastructure’ startups that help influencers set up their own servers, manage security, and build custom recommendation engines. Some are even open-sourcing their code so smaller creators can use it. Luna: Open-source recommendation engines for influencers? That’s wild. What’s the most popular one right now? Lucas: There’s one called ‘CreatorOS’ that’s gaining traction. It’s basically a Kubernetes-based platform that lets you deploy a recommendation engine using your own data. It integrates with YouTube and Instagram APIs — the ones that are still open — and it can also ingest CSV exports. The founder told me they have over two thousand creators using it, mostly in the one hundred thousand to one million follower range. Luna: So the trend is trickling down. It’s not just the mega-creators anymore. Lucas: It’s starting to. But there’s a catch. Running your own infrastructure means you’re responsible for security. If you’re storing user data — emails, viewing habits, maybe even purchase history — you have to comply with GDPR, CCPA, and all the other privacy regulations. One breach and you’re liable, not the platform. Luna: That’s a huge risk. Most creators barely have a privacy policy on their website. Lucas: Exactly. And that’s why some advisors are pushing back. They’re saying, ‘Look, unless you have a dedicated security team, you’re better off using a managed cloud service that handles compliance for you.’ But the trade-off is you’re still renting your data infrastructure, not owning it. Luna: I wonder if brands will start requiring creators to have their own data centers as a condition for big sponsorship deals. Like, ‘We’ll only work with you if you can give us direct access to your audience data, not just platform analytics.’ Lucas: That’s actually already happening in a small way. I know of a few DTC brands that have clauses in their influencer contracts requiring the creator to provide raw data exports — not just screenshots from Instagram Insights. And if the creator has their own infrastructure, they can do that easily. If they don’t, the brand might not get the granular data they want. Luna: So owning a data center becomes a competitive advantage in negotiations. Lucas: Exactly. It signals that you’re a serious business, not just someone who posts selfies. And it gives you leverage. You can say, ‘I can guarantee you clean, first-party data about my audience, with no platform black box.’ That’s incredibly valuable to a brand spending millions on a campaign. Luna: Let me play devil’s advocate. Isn’t this just a trend that will fade? Like, how many creators actually need this? Lucas: It’s definitely not for everyone. For a creator with under a million followers, the cost and complexity probably aren’t worth it. But for the top hundred or so creators globally — the ones who are essentially media companies — it makes sense. And as the tools get cheaper and more user-friendly, I expect it will become more common among mid-tier creators too. Luna: It reminds me of how publishers moved from using third-party ad networks to building their own ad servers in the early 2000s. At first it seemed crazy, then it became standard. Lucas: That’s exactly the right analogy. The creator economy is going through the same infrastructure build-out that the publishing industry went through twenty years ago. First you rely on platforms. Then you realize you need to own your own stack. And the data center is the ultimate expression of that. Luna: Before we wrap, I want to talk about something related. We’ve been having these deep conversations about marketing and infrastructure, and a lot of listeners have told us they actually use these insights in their own work. If these marketing conversations have sparked something you’ve actually used — a strategy, a tool, a way of thinking — we want to keep this show free and independent. No ads, no sponsors. If you want to support that choice, you can find us at buy me a coffee dot com slash fexingo. Lucas: Yeah, it’s a small way to keep the conversations going. And we appreciate every bit of support. Luna: So back to data centers — do you think we’ll see creators pooling resources to build shared infrastructure? Like a co-op data center for influencers? Lucas: That’s actually a really interesting idea. There’s already a startup trying to do that — they’re calling it ‘Creator Cloud.’ The pitch is that a group of creators can share the cost of a colocation facility and split the data storage. But the challenge is trust. Would you really want your audience data sitting on a server that another creator has access to? Luna: Yeah, that’s a non-starter for most. But maybe if it’s properly partitioned with strict access controls. Lucas: Possible, but it adds complexity. I think for now, the trend is toward individual ownership among top creators, with open-source tools making it accessible to more people over time. Luna: It’s fascinating to watch the creator economy mature. From sponsored posts to data center investments in just a few years. Lucas: And we’re still in the early innings. I wouldn’t be surprised if in five years, having your own server rack becomes as normal for a top creator as having a manager or a lawyer. Luna: Alright, that’s a wrap for this episode. Thanks for listening.