Who’s Powering AI Infrastructure? It’s Not Just the Hyperscalers

Just a few years after migrating to Google Cloud Platform, SiteGround’s websites are up to five times faster, according to Nikolay Todorov, the web hosting provider‘s CEO.

There’s a reason GCP is one of the biggest names in the game: It has an extensive cloud network with more than 202 points of presence and is supported by 2 million miles of fiber across the globe.

“This is the same network that powers services like Google Search and YouTube, so SiteGround clients are tapping into technology used by billions any second,” Todorov wrote in the official post.

Since the switch, SiteGround has also strengthened security, blocking more than 90% of malicious traffic.

“In the AI era, data security and reliability are paramount, as threats evolve and website users and restrictions become even more demanding,” Todorov wrote.

While impressive, SiteGround’s upgrades are just one part of a larger trend many industry folks have been seeing.

We are getting a firsthand glimpse of what the future of Web 4.0 actually looks like: infrastructure that doesn’t just deliver data, but also autonomously learns, adapts, and protects itself as threats evolve.

As Todorov said: “While currently these technologies serve complex enterprise use cases, they reflect the sophistication already at play beneath SiteGround’s infrastructure.”

Smaller hosting providers today are often overshadowed by industry giants like GoDaddy, WordPress, and even SiteGround.

But that doesn’t mean they can’t compete. Among the common threads we’re seeing is a focus on user-centered experiences that are more personalized by AI.

So, instead of building costly AI infrastructure from scratch, smaller hosts should take advantage of existing managed AI and cloud services from trusted providers like GCP, AWS, or Azure.

It’s a cost-conscious way to deliver AI-powered features, like auto-scaling, personalization, AI security bots, and chatbots, without the upfront costs that most startups want to avoid.

Years ago, being “built for scale” meant handling things like traffic spikes on Black Friday. Today, as most of us know, scale means more than just supporting more users and data.

It’s that and massive AI workloads.

Average bandwidth speeds went from something like 10 Mbps in the late 2000s to 100 Mbps in 2025. Some even predict that 10 Gbps (equivalent to 10,000 Mbps) will become the standard by 2030.

Whether that’s true, it one reason why the industry has been seeing an uptick in more user-focused delivery.

Take a look at a few examples:

• WordPress AI plugins: According to a study, 40 AI-powered WordPress plugins collectively averaged more than 26 million visits per month. And it’s suspected that 75% of WordPress users will use at least one AI plugin by the end of 2025.
• eCommerce personalization: AI-powered product recommendations can lead to a 10-30% increase in sales. Nearly 45% of online shoppers are also more likely to buy something on websites that offer a customized experience.
• Chatbot integration: Around 40% of businesses use chatbots for customer support. It’s for good reason, too: Those bots respond three times faster than human agents and have been shown to improve resolution times by 90%.

While hyperscalers are excited to unveil the latest GPUs and next-gen systems, hosting providers are bridging the gap between those innovations and the everyday users who actually rely on them.

Think about the last thing you ordered on DoorDash. If it shows up cold, spilled, or squashed, does it really matter how great the food is? Not really. What matters is that you’re still hungry.

The same can be said for the next era of web hosting, when the tech is only as valuable as the experience it can deliver.