Cloud storage company Bitcasa has come a long way since its former CEO Tony Gauda had to explain to skeptical judges at the TechCrunch Disrupt conference in 2011 how it was able to encrypt data while being able to scale its storage efficiently.

Many remember Bitcasa by its go-to-market strategy of $10 per month of infinite storage, which “was almost entirely suicidal,” Bitcasa CEO Brian Taptich said in an interview with Talkin’ Cloud. But the company learned a lot about the product, namely “that there was massive demand for it” and that it worked at a global scale, he said.

At that time the company was also being approached by the developer universe – “two person dev shops all the way up to companies like Samsung and Huawei who were participating in this war for customer data ownership but didn’t have the arms to do that.”

The company spent 2014 redeveloping the platform and relaunched last year, announcing a licensing agreement with SanDisk for the development of a SanDisk-branded cloud storage solution.

Fast-forward to 2016 and Bitcasa has won a patent for its deduplication and smart caching techniques that help it deliver secure and efficient cloud storage.

The patent covers the process through which data is encrypted with Bitcasa, which Taptich said is unique because it happens at the block level, not the file level.

“The way that Bitcasa works, and this speaks specifically to the patent, is…on the client-side we’ll take this file and we’ll chop it into blocks. At no point in this process do we know what the actual file is,” Taptich said.

According to Taptich, on the client side, a user uploads a file; let’s say it’s a HD copy of The Martian, which is then chopped into blocks. The blocks are encrypted on the client-side and Bitcasa creates a hash against the encrypted data.

“Once we create a hash on that one block, we then check that hash against our existing database to see if that block already exists in our database,” he said. “Since you’re the first person who is uploading it, it doesn’t. Then we will assign that block its own marker based on metadata which is then also encrypted with a key that you own and then we’ll upload that block.”

For the second person uploading The Martian, they’ll go through the same process but because Bitcasa already has the encrypted block, instead of uploading the block itself will create metadata that attributes the same block that was uploaded by the first person. Taptich said this “is way more efficient” because it’s not uploading every single block of data, while still protecting privacy and security of the user.

Client-side encryption, where the customer owns the encryption key, means that “not only do I not what data you have in my system, I don’t know which data is yours,” Taptich said.

“The sleight of hand that Bitcasa has done is we’ve said, ‘look, at the end of the day this data is end-user data. It’s not our data, it’s not Google’s data, it’s not the NSA’s data – it’s the user’s data.’ So we’re going to build systems that don’t sacrifice privacy at the expense of utility and price.”