The idea has been sitting in a research paper since 2004: attribute-based encryption (ABE) is a cryptographic technique invented by Brent Waters, now the head of NTT Research’s cryptography lab, that allows data to be encrypted in layers so granular that different users see exactly what they’re authorized to see and nothing more.
In practice, think of a hospital system where a patient’s record contains everything from billing details and surgical notes to psychiatric history. A billing clerk needs the insurance information; a surgeon needs clinical notes; and a psychiatrist needs the mental health history.
Under conventional encryption, access in binary – you either have the key, or you don’t, meaning that organizations routinely over-share data internally just to make workflows function.
ABE flips that logic. Each piece of data is encrypted against a specific set of attributes – job role, clearance level, department, time of day – and a user’s decryption key only works if their attributes match the policy attached to the data: nobody sees what they don’t need to.
No administrator manually managing permissions, no single key that, if stolen, unlocks everything.
For two decades, ABE was a technical achievement without the market. Then came data lakes, multi-cloud environments, AI agents handling sensitive information at scale – and suddenly, the world caught up to the research.
Twenty years in the making
“When they invented this technology, they did not envision these kinds of things going on,” said Fang Wu, vice president of business development at NTT Research and leader of the newly-announced Scale Academy SaltGrain efforts.
“All these use cases were actually identified recently, because there were many emerging situations – data lakes, IoT, cloud – that simply did not exist when ABE was created,” he told Startup Beat while at the annual Upgrade 2026 conference in San Jose, California – which convened hundreds of attendees under the theme Research to Reality.
SaltGrain is ABE’s commercial incarnation – repackaged, productized, and now being pushed to market through Scale Academy, NTT Research’s newly formalized incubation arm. It is the first test case for what NTT Research president and CEO Kazu Gomi is calling NTT Research 2.0: a deliberate pivot from pure science toward commercialization, built on the premise that decades of fundamental research have produced more market-ready technology than the company has ever tried to sell.
Bennett Indart, who runs Scale Academy, frames it as a natural evolution rather than a strategic reversal. “You start with ideas, you postulate, you test them through the scientific method, you reiterate,” he said. “And then at some point, things pop out that have commercial applicability and a market.”
The harder question, he acknowledges, is why it took so long. Big companies, he’s observed over nearly two decades inside NTT and other large organizations, are very good at generating great ideas and very bad at getting them to market. “I’ve seen it fail a lot. Great ideas get to a certain point and they kind of stall out.”
Built to avoid the usual traps
Scale Academy is designed to be structurally different from the kind of in-house innovation programs that typically produce that outcome. It is deliberately isolated from NTT’s operating companies, runs its own budget and timeline, and is staffed – pointedly – by non-researchers.
“That was done on purpose,” Wu said. “We actually needed to have the skills of the business side, the development side.”
The unit operates on a Silicon Valley-style stage-gate model: four phases – discover, incubate, commercialize, scale – with a governing board that includes Gomi and other NTT stakeholders reviewing progress at each threshold. Miss the metrics, lose the funding.
“We will make that hard decision. No more funding for this, let’s go to the next one,” Indart said.
For SaltGrain specifically, the commercial thesis turns on timing. ABE was theoretically sound years ago, but practically immature; the cryptographic community spent years stress-testing it, and it has not been refuted to date. “Brent Waters has done something like five or six ‘test of time’ awards for it,” Indart noted.
What changed recently is two-fold: AI has made the classification problem that ABE solves dramatically easier to automate, and the threat landscape AI itself has created has made the underlying need more acute.
“Along comes this explosion called AI, and it became a perfect storm for us to say: this is the right time to make that big splash,” Indart stressed.
Going direct
The go-to-market strategy is direct first – a deliberate choice to keep the product close enough to learn from early customers before handing it to large channel partners. Healthcare, insurance, financial services, and manufacturing are the initial target verticals, chosen for their regulatory exposure and the severity of breach costs.
“Everyone actually sees the value in the first meeting,” Wu said. “And they actually tell us use cases we did not really anticipate.”
Longer-term, the roadmap points toward functional encryption – a successor technology that would allow a third-party to perform computations on data without ever seeing its contents. It’s further out, mathematically complex, and computationally intensive, but on the roadmap.
So is a post-quantum cryptography paper being published this week. The pipeline, Indart suggests, is deeper than SaltGrain alone.
Success, in his definition, looks like a spinout: SaltGrain exits the incubator as a standalone company, backed by outside capital – venture or private equity – with NTT retaining an equity stake and a board seat.
“It will prove to NTT itself that it can do something like this, and hopefully we can do more of it, because there are a whole bunch of patents out there with high potential to make a big change in the market,” Indart concluded.
Whether Scale Academy becomes a repeatable model or a one-time experiment depends almost entirely on what SaltGrain does next. The research was always there. Now, for the first time, someone is trying to sell it.
Featured image: Courtesy of Salomé Beyer Vélez
