Intel · Security Libraries for Data Center
We Built Better Security. Nobody Cared.

- 3 ISVs and 2 cloud providers at launch
- Foundation for NIST's federal zero-trust reference architecture
- Time to first attestation cut by more than half
- Pilot to production conversions tripled
The Security Features No One Used
Intel had built advanced security into Xeon processors.
Superior technology. Better than competitors.
Zero customer adoption.
Leadership asked me to turn dormant silicon capabilities into revenue. I committed to 3 ISV adoptions and 1 cloud provider by 2020.
But first, I needed to answer a different question.
The Question That Reframed Everything
Why would customers with working security solutions rip them out to adopt ours?
I ran discovery sessions with three groups: ISVs wanting to build security products, cloud providers seeking new revenue streams, enterprises needing workload assurance across hybrid environments.
A pattern emerged.
The blocker wasn’t our technology. It was integration friction.
Our hardware was a black box. Customers would have to build integration layers themselves to make it work in their environments.
We weren’t selling a chip feature. We needed middleware that made Intel security capabilities accessible within workflows customers already used.
The Cultural Bet
Should we build proprietary software or go open source?
We went open source.
Here’s why: Open source eliminated the trust problem. Enterprises could audit code for regulated environments. ISVs could build solutions without licensing negotiations. Every integration made the platform more valuable to the next adopter.
This was the Intel way. Intel is one of the biggest open source contributors globally.
I positioned it through the lens leadership understood: We don’t make money on middleware. We make money on silicon sales. Open source removes adoption barriers.
Each middleware adoption meant selling more processors.
What We Built
Intel Security Libraries for Data Center. Apache-licensed.
The key architectural decision: integrate natively with Kubernetes and OpenStack. The platforms customers already used.
We weren’t building standalone software. We were removing friction from existing workflows.
The Validation
We launched with 3 ISVs (HyTrust, AMI, Lanner), 2 cloud providers (VMware, IBM Cloud).
NIST selected our approach as the foundation for federal zero-trust reference architectures.
But we had a problem. Strong GitHub engagement. Positive feedback. Lots of downloads.
Terrible deployment completion rates.
Most teams downloaded the software but never got it running in their environments.
The Experience Problem
I talked directly to teams who tried but failed to deploy.
Where did they get stuck? What made them give up?
Three barriers: Our setup assumed specialized security expertise most teams didn’t have. Integration required significant custom work. When something went wrong, diagnosing the problem was nearly impossible.
We completely rethought the approach.
Pre-built containers. One-command installation scripts. Native integrations with platforms teams already used. Clear feedback when things went wrong.
Time to first successful security attestation: dropped by more than half.
Pilot projects converting to production deployments: tripled.
Partners began embedding our platform directly into their products.
What It Taught Me
Zero-to-one products fail more often from distribution problems than technology problems.
We didn’t win because we had better security technology. We already had that.
We won because we solved the distribution problem by removing integration friction.
Open source was the unlock. Ecosystem leverage was the scale mechanism.
We created a new category from zero by understanding that the product wasn’t the middleware.
It was the path to adoption.