A speculative technology piece written to explain an advanced cryptographic concept (FHE) in a high-impact business context. It bridges technical constraints, market applications, and the advertiser privacy arms race — making a dense academic topic legible to a mixed business-technical audience.

Role: Sole author
Audience: SaaS founders, adtech execs, fintech stakeholders
Use Case: Thought leadership + narrative framing

Scientists and mathematicians researching select areas are prized commodities in the tech industry, and it was no surprise to insiders that a mathematician working on fully homomorphic encryption was poached from Microsoft by Meta in 2021. To the outside world, however, HE is not only an esoteric concept, but has only recently been proven viable in a variety of arenas.

Basics of Homomorphic Encryption

HE is a complex system that revolves around a simple concept: While the user is unable to access the underlying data they are able to perform operations on it. When decrypted, the results would be the same as if they had been performing those operations on the original data. The level of complexity of the operations is dependent on the algorithm used and the compute power available.

Put another way, imagine a doctor’s office that has folders of patients’ records. Homomorphic encryption allows someone to add up the number of diagnostics tests performed or the number of procedures performed over the entire set of folders. While doing so, however, they are unable to access specific patient records or link personal identifying information with that data.

The Final Frontier for Advertisers

In an era when consumers are as privacy-conscious as they have ever been, the ability to analyze large swaths of data looking for insights has become constrained. Apple’s announcement recently that it would block tracking via cookies and Google’s attempt to move on from using the same in analytics caused a significant amount of consternation.

Homomorphic encryption would erase that issue for Google, Meta and other hyperadvertisers that have access to literally billions of records. Analysis could be made based on the encrypted data of groups of users to create audiences and enable targeting that would be similar to that the companies once were able to provide advertisers.

Stumbling Blocks Remain

The first proof-of-concept for a homomorphic encryption standard took many thousands of times longer than performing the same data analysis using plaintext variations of the sets. New algorithms and standards have reduced the wait to a fraction of that time, but simply put, the compute requirement is massive and mathematicians designing and advising on the encryption protocols face two problems:

• The most valuable encryption protocols are those that allow analysis to be performed using multiple operations, i.e. division, subtraction and multiplication along with addition. However, fully homomorphic encryption protocols are even more compute-intensive than partial ones.
• Trying to create algorithms that enable faster analysis runs the risk of making it easier for others to access the data by weakening the encryption standard itself.

Still, the opportunity for hyperadvertisers is massive, and the knock-on effects for other companies is immense. In industries with strict privacy regulations like banking, healthcare and others, the ability to perform this same analysis, or more importantly, outsource it to data analytics firms without sacrificing user privacy, could be a game changer. The only question is how long it will take for the algorithms to catch up with the idea.