Pushing Against Digital Borders: The New Geopolitics Of The Internet And AI

Article originally published on Forbes.com. Image courtesy of Getty.

Here is an uncomfortable truth: The dream of a unified global internet is fading.

Once, the internet represented a sort of technological “shining city on a hill.” It offered the promise of lifting billions of people from poverty, widespread education and infinite new streams of global commerce.

But today, we’re entering a digital era shaped by sovereignty, security and national self-interest.

Countries are tightening privacy laws and rethinking digital dependencies. If data becomes jurisdiction-bound, enterprises—especially in regulated sectors like finance, healthcare and defense—must redesign their platforms for local compliance, edge processing and infrastructure sovereignty.

Concerns that were once merely about privacy now present national security issues. A decade ago, Europe prioritized personal data rights more than the U.S.; today, data sovereignty is geopolitically charged. Outdated legal frameworks meet modern digital complexities. Globally, we’re seeing fragmented regulations: GDPR in the EU, India’s DPDP, Brazil’s LGPD, China’s PIPL, plus emerging AI laws. It’s a patchwork, and it’s often conflicting—not just a legal problem, but a systems architecture one.

To navigate this, companies must localize workloads, regulate cross-border data transfers and control metadata. But most lack the tools and expertise to continuously adjust their infrastructure in a dynamic state of play.

Achieving Scalable Compliance

I recently spoke with an automotive executive whose company operates in Mexico, Brazil and the U.S. Their supply chain faces daily unpredictability, constant cost shifts and regulatory changes. It’s hard to plan or scale. Their experience underscores a key point: A global footprint isn’t enough if a company isn’t compliance-ready.

Many tech companies tout global reach, but few ensure end-to-end compliance. True readiness requires customized deployment models, sovereign virtual environments, geofencing, localized monitoring and legal support. For enterprise customers in regulated industries, tools to facilitate compliance aren’t “nice-to-have”; they’re nonnegotiable, and geopolitics only accelerate this trend.

What used to be primarily the domain of privacy advocates is now a matter of national security. Countries are decoupling digital infrastructure from foreign providers, increasingly wary of foreign data hosting. Hyperscalers are being viewed as strategic risks, prompting governments to build sovereign clouds. I’ve spoken recently with people in the Middle East, Europe, Australia and Canada. Sovereign cloud initiatives are everywhere.

This geopolitical shift demands bespoke infrastructure. No two countries, or customers, have identical requirements.

The takeaway? Build for borders and work to operate beyond them. I used to talk about free information flow, and I still believe it is a greater good ... but the world is changing. Leaders must rethink infrastructure blueprints to account for fragmentation and design for modularity and segmentation. Build for compliance zones, adopt regional SLAs and partner with those who track regulations and respond locally. Make compliance as scalable as your cloud. The future belongs to infrastructure that can adapt across legal systems and geographies.

Adding ‘Fuel To The Fire’

Of course, if SaaS infrastructure—software, data, cloud platforms—is complex, AI is pure chaos.

AI models rely on massive, diverse datasets, many containing PII, biometrics, regional content ... all highly regulated. And now governments are catching up. Increasingly, they require that any AI training involving citizens’ data happens locally, not blended into global datasets. This requires a fundamental shift in AI infrastructure.

Take federated learning, for example: This means training models locally on edge nodes while sharing insights, not raw data, so that the data never leaves the country. Consider also model segmentation; distinct models must be run in different regions to comply with local laws. Think about inference localization; we have to ensure that decisions and data storage stay within jurisdictional boundaries. And it must all be auditable, which is nearly impossible with current AI systems.

These things matter most in high-stakes sectors such as defense, finance and healthcare. In these ecosystems, cross-border AI missteps won’t just bring fines. They can set fire to contracts, irreparably destroy trust or even threaten national security.

There have been many articles written recently about how dangerous it is to give AI unrestricted access to sensitive data. AI gets more powerful the more it knows about you, and that’s a double-edged sword, putting everyone from users to actual governments into a potentially precarious situation.

The other day, I used ChatGPT to help summarize a transcript. I asked it to draft an email from the meeting notes ... and it gave me something unrelated. It mentioned someone’s mom, a random neighborhood and names that had nothing to do with my prompt. I asked again and got another random response. I realized something was off; likely, the wrong GPU was responding and I was getting output meant for someone else.

Obviously, that’s one user’s perspective. But imagine what’s at risk when AI systems handle Social Security numbers, defense data or financial records. It’s not just an IT problem; it’s a trust and security problem.

Managing SaaS platforms is already complex in today’s geopolitical climate; managing AI adds a whole other layer of chaos. As an enterprise leader, you have to build for that level of risk from day one.

An Accelerating Risk Environment

The reality is that, while these concerns are top of mind for many individual executives, companies tend to be inherently risk-averse. Often, this is because they’re not fully aware of just how fast that risk is evolving, particularly within the added tangle of international regulations.

To be clear, I’m not trying to sound alarmist. However, we are at a turning point. Enterprise leaders must pause and ask themselves: “Do we have the infrastructure and partners to simultaneously navigate tightening geopolitical constraints and a heightened compliance risk environment?”

Smaller companies, in particular, may not have the resources to adapt quickly to shifting rules. And at a larger scale, you’re seeing national governments scrambling to build EU- or country-specific clouds to avoid entanglements with bad actors, preserve sovereignty and protect their citizens. Europe took the lead on data privacy five years ago; the U.S. is still struggling to catch up. Time will tell whether these efforts are efficient or effective.

But one thing is clear: Everything is accelerating. And with the added curveball of AI proliferation, it’s now clear that unfettered access to data isn’t just risky but potentially catastrophic. Still, a lot of companies are only just waking up to that fact.

Enterprise leaders must ask themselves critical questions about how they are going to build digital infrastructure for the future and take steps accordingly. The landscape is changing by the minute, and the time to pivot is no later than now.