In our last issue, I wrote about smart cities and the growing debate over who owns the enormous amount of information they collect.
This debate might help explain why a strange technology race is taking shape today.
Companies, governments and researchers are scrambling to build tools capable of protecting sensitive information in a world where AI can infer information we never intentionally shared in the first place.
This new privacy market is set to become a massive business.
And could also become one of the most important technology battlegrounds of the next decade.
It’s Complicated…
More than a decade ago, Target figured out a teenage girl was pregnant before her father did.
The retailer’s analytics systems noticed changes in her shopping habits and began sending coupons for baby products to the family’s home.
Her father reportedly stormed into a Target store demanding answers, only to later discover the prediction had been correct.
At the time, that story seemed shocking. But today it almost feels primitive.
Because Target made that prediction using relatively basic data analysis compared with the much more advanced AI systems being built today.
The entire online privacy economy was built around the simple idea of trying to protect information you knowingly handed over. Companies collected your data, but you decided whether to share it. And cybersecurity firms tried to keep it from leaking or getting stolen.
But AI is changing the definition of privacy.
Back in 2013, researchers from Cambridge and Microsoft demonstrated that Facebook Likes could accurately predict highly personal traits including political views, personality characteristics and sexual orientation.
That study used data from just 58,000 volunteers.
Now imagine what modern AI systems can do with much larger pools of information coming from smartphones, wearables, connected cars, smart homes and intelligent infrastructure.
AI can even use Wi-Fi to identify us.
That’s why the old privacy model won’t work in the AI era.
Since AI can increasingly connect patterns humans could never notice on their own, a hospital may soon no longer need your medical records to infer certain health risks. Or a retailer may not need your purchase history to predict your behavior.
In other words, AI is slowly changing privacy from a data problem into an inference problem.
And it’s creating a massive new market almost overnight.
Estimates vary, but the market for so-called “privacy-enhancing technologies” was about $4 billion in 2025. But it could grow to more than $28 billion by 2034.
Image: scoop.market.us
That represents 7X growth in less than a decade. Yet, most investors still aren’t paying attention to this sector.
Back in January 2024, I recommended Palantir (Nasdaq: PLTR) as my No. 1 stock pick for the year. The company helps governments and businesses make sense of enormous amounts of data, making it an important piece of AI infrastructure.
We eventually sold half the position in our Strategic Fortunes model portfolio for a gain of 994% and the remaining shares for a gain of 780%.
Today, another infrastructure story is starting to emerge. But this one isn’t about helping AI become more powerful.
It’s about helping us control what it learns.
And some of the technologies being developed today are truly fascinating.
One is called “zero-knowledge proof.”
It allows someone to prove something is true without revealing the underlying information itself.
For example, you could prove you’re old enough to buy alcohol without exposing your birthdate. Or prove you live in a certain city without handing over your full address.
Another emerging technology is called homomorphic encryption.
It allows computers to run calculations on encrypted information without decrypting it first.
That might sound absurd, but it could become extremely important in a future where hospitals, banks and governments want AI systems to identify patterns without exposing the sensitive data underneath.
There’s also growing interest in synthetic data.
Instead of training AI on real personal information, companies create artificial datasets that statistically behave like the real thing without being connected to actual people.
That means a hospital could train AI systems without exposing patient histories. Or a city could model traffic patterns without storing every driver’s movements.
Naturally, Big Tech is moving aggressively into this area.
Last year, Apple introduced something called Private Cloud Compute, designed to let Apple Intelligence process more complicated AI requests while limiting how much personal information is exposed to the cloud.
And Google and Nvidia are both pushing heavily into “confidential computing,” which is designed to protect sensitive information even while AI systems are actively processing it.
In fact, governments and corporations are becoming increasingly nervous about what these systems can learn.
The European Union’s AI Act now bans certain AI systems designed to infer highly sensitive personal characteristics, including some forms of biometric categorization and emotion recognition.
And Samsung temporarily banned employees from using ChatGPT internally after reports that workers uploaded sensitive source code and confidential meeting notes into the system.
That tells us something important.
Here’s My Take
The AI boom is creating a strange new problem.
The smarter AI becomes, the harder it is to control what it can learn.
That helps explain why companies, governments and regulators are all racing to build new privacy technologies.
I’ve spent the last few years focusing on the infrastructure needed to make AI possible. Now I’m also keeping an eye on the infrastructure needed to keep it under control.
Because the next great AI infrastructure market may not be about helping machines learn more.
It may be about helping them learn less.
Regards,
Ian KingChief Strategist, Banyan Hill Publishing
Editor’s Note: We’d love to hear from you!
If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to [email protected].
Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!
