Challenges
- Lumilens designs and manufactures silicon photonics that connect GPUs inside the world's largest AI data centers, making source repositories, design documents, and manufacturing data a high-value target for IP theft and nation-state activity.
- A small engineering organization spread across the US, India, and APAC meant identity and access events crossed multiple time zones with no 24/7 monitoring, and noisy native cloud signals left coverage gaps.
- False positives from cloud and identity sources consumed hours of engineering time per incident, pulling principal engineers away from photonics R&D to chase routine login anomalies and benign API calls.
- Building an in-house SOC would have required hiring three to four security engineers across regions and standing up a SIEM, an investment that would still leave Lumilens with delayed coverage and brittle detection across a fast-growing global footprint.
Solutions
- Exaforce MDR continuously monitors source code, project, design assets, and cloud control planes, surfacing entity connected Threat Findings backed by the Exaforce real-time knowledge graph.
- Exaforce correlates multiple identity platforms, workspace, collaboration, and cloud workload telemetry into one graph in real time, closing the blind spots that come with siloed native alerting across disconnected tools.
- Exabot Triage and Exabot Investigate apply Multi-Model AI to every raw alert, cutting false positive load by over 90% and producing case-ready Threat Findings in minutes.
- The Exaforce data platform and MDR analyst team delivered an instant 24/7 SOC across every region, without a SIEM to deploy, no rules to tune, and no security headcount to hire.
Securing the infrastructure behind AI
Lumilens builds next-generation photonic interconnects that let AI hyperscalers scale GPU clusters past the limits of copper and legacy optical transceivers. Its founding team came out of Cisco, Juniper, Meta, Lumentum, and Coherent to design in-house silicon photonics, automated manufacturing, and full-stack systems that move data at the speed of light. With proprietary silicon designs and a fast-growing roadmap of customer commitments, the company's source repositories, photonic chip designs, and manufacturing data are exactly the kind of IP that determined adversaries pursue. Before Exaforce, Lumilens relied on default cloud signals from cloud detection tools paired with informal review by engineers. Coverage was incomplete, entity context was disconnected from cloud activity, and there was no realistic path to 24/7 monitoring without a major hiring push.
Lumilens selected Exaforce MDR for the combined value of platform access and a dedicated analyst team. Within weeks, Exaforce was ingesting telemetry from their cloud services, workspace platform, identity platform, repositories, and more, stitching identities, devices, and resources into a single real-time knowledge graph. Exabot Detect began producing high-fidelity Threat Findings that linked authentication events to code activity and cloud control plane changes. Exabot Triage closed the long tail of routine false positives, and Exabot Investigate built case context that previously took engineers hours to assemble. MDR analysts handled every escalation, leaving the Lumilens team to focus on photonics and customer delivery rather than alert review.
The change in operational speed showed up almost immediately. With Exaforce Automation Agents, Lumilens configured a recurring insider risk report that runs automatically across the knowledge graph, pulling access events touching silicon photonics repositories and cross-referencing identity and data movement into a structured summary. The first report surfaced a pattern of an employee progressively pulled larger data batches over two weeks, with no individual event severe enough to trigger an alert. The MDR team reviewed the full behavioral arc, confirmed the risk, locked the account, and followed up with the Lumilens team.
Across the first quarter of operation, blended MTTI for P0 and P1 findings shrunk to roughly 6 minutes from over 3 hours. Of 190 Threat Findings reviewed by the MDR team, every one was verified as accurate, and only 2 were real threats requiring escalation.
"I used to spend time I didn't have triaging alerts that turned out to be nothing," said Saurabh Bansal, Sr. Director, Technology & Infrastructure at Lumilens. "With Exaforce, which we think of as our force for good that exposes risk, executes security seamlessly, and acts decisively, Lumilens can keep up its pace of innovation without worrying about security. That work is already done by the time anything reaches me, and what reaches me is real. And when something does need investigation, the context is already there. I'm not hunting across five tools to piece together what happened. I'm reading a complete picture and making a call."
As Lumilens has scaled engineering and manufacturing across the US, India, and APAC, Exaforce has scaled with it. The same real-time knowledge graph that protects source code in GitHub now correlates manufacturing system access, contractor identity activity, and cloud workload behavior across every region. Engineering leadership is getting back over five days of focused time every 30 days, and findings reach the right person with full context instead of raw signals that demand a debugging session.
"The impact has extended well beyond security. Our engineering leaders are no longer spending valuable time investigating alerts that lead nowhere. They receive actionable information when something truly requires attention and can make decisions quickly because the necessary context is already available,” said Jay Robinson, Talent Senior Director at Lumilens, “As Lumilens continues to scale globally, Exaforce has helped us build confidence that our systems, intellectual property, and business are protected without requiring us to build a large security organization of our own. At a startup, every hire matters and every engineering hour counts. Exaforce has allowed us to strengthen our security posture while keeping our teams focused on what they do best, building technology that will shape the future of AI infrastructure.”
With Exaforce, Lumilens turned a fragmented set of cloud and identity signals into a single 24/7 SOC, freeing engineering time, closing IP risk, and giving the team confidence to scale photonics innovation as fast as the AI infrastructure market demands.
