The most revealing word in Himanshu Shukla's public vocabulary may be whose. Enterprise security has long been fluent in what and where: what kind of data is this, and where is the file sitting? Shukla, the co-founder and CEO of Lightbeam, keeps pulling the conversation back to the person inside the record. Whose bank number? Whose address? Whose permissions? A database sees a field. A privacy request sees a human being waiting for an answer.
That grammatical shift is the spine of Lightbeam. The San Jose company connects scattered sensitive attributes to identity, then layers in location, access, policy, and risk. The result is meant to let a security team see more than a red alert. It can see the relationships that make the alert consequential.
The idea did not arrive during a founder retreat or a speculative tour of fashionable markets. It came from work Shukla and his future co-founders, Aditya Ramesh and Priyadarshi Prasad, had already endured at Nutanix. They struggled to definitively discover sensitive information, classify it accurately, and automate the work required to secure it. During the early enforcement years of European privacy regulation, customer requests could require data to be found, managed, and sometimes deleted across a thicket of systems. The tools offered pieces. The workflow demanded a whole.
A founder hiding inside the infrastructure engineer
Shukla's route to that problem ran through the less photogenic rooms of technology. He studied at the Indian Institute of Technology Delhi from 1993 to 1998, then worked in engineering roles at Novell, Cisco Systems, Sinett Corporation, and Ericsson. By 2013, he was at Nutanix, moving from staff engineer and senior manager to technical director and eventually senior technical director for AIOps.
AIOps is the practice of using software and machine learning to understand the noisy behavior of large technical systems. It rewards a particular temperament: patient with complexity, suspicious of isolated signals, eager to automate the repetitive diagnosis. Shukla's patent record has the same fingerprints. His listed inventions cover virtualization search, predictive resource planning, workload forecasting, automated scheduling, cluster health, and network balancing. Different titles, recurring appetite: turn a sprawling system into something observable enough to act upon.
He also co-authored research on preventing resource hotspots in private clouds. That sounds far from privacy requests until one notices the shared concern. In both cases, trouble appears when important context is scattered, systems are under pressure, and humans need the machine to recommend a useful next move.
“Privacy is the new business currency.”Himanshu Shukla
Borrow the architecture, change the object
Nutanix made its name by converging infrastructure functions that had lived in separate boxes. Lightbeam borrowed the architectural instinct and changed the object. Instead of compute, storage, and virtualization, its founders put data security, privacy, access governance, and remediation on one platform. The trick was not merely arranging several tabs beneath one logo. The common layer had to preserve meaning across them.
The identity-centric loop
Lightbeam calls that shared layer a Data Identity Graph. A conventional scanner may recognize a string as a credit card number. An identity graph tries to understand whose number it is, where else that person's data exists, who can reach it, and which business rules apply. Classification becomes the beginning of a decision instead of the end of a scan.
There is a moral argument tucked inside this technical design. Lightbeam describes its mission as restoring the human dimension to data security. Shukla's version is pragmatic. Companies want to analyze customer information, move into new markets, and adopt new tools. Customers expect dignity and control. Treating privacy as a brake makes the two sides permanent enemies. Treating it as dependable infrastructure lets trust become part of how the business moves.
December code, February capital
Shukla dates Lightbeam's practical beginning precisely: December 2020, when the team began writing its first lines of code. That detail gives the company a pleasingly unceremonious birthday. No gong, only a repository and a problem.
The founding trio brought a shared Nutanix history. Shukla took the CEO role. Ramesh became chief architect and led research and development. Prasad shaped the product and remained a visible partner in the company's public work, including a 2024 session with Shukla at the FS-ISAC Americas Spring Summit. Their topic, “Taming without Maiming: AI, Have it Your Way,” caught the tension neatly. Businesses wanted the gains from AI without surrendering control of the data beneath it.
In February 2024, Lightbeam announced a $17.8 million Series A led by Vertex Ventures US, with Dropbox Ventures, 8VC, and Village Global participating. The company reported that its customer base had grown by roughly 300 percent over the previous year. Shukla's announcement read less like a victory lap than a long thank-you note. Employees, co-founders, investors, advisers, and early supporters appeared by name. He called the staff “Beamers” and closed with a compact founder's refrain: “We are just getting started.”
The names sketch a network around the company. Sandeep Bhadra and Chase Roberts represented the round's lead investor, Vertex Ventures US. Dropbox Ventures joined the bet. Bhaskar Ghosh and Jack Moshkovich appeared from 8VC, while Ben Casnocha represented Village Global. Shukla also thanked a circle of operators and founders that included Nutanix co-founder Dheeraj Pandey, Amit Sinha, Rishi Bhargava, and Venktesh Shukla. Their backgrounds span infrastructure, identity, security, company-building, and venture investing, which is a remarkably accurate map of the territory Lightbeam has chosen.
The gratitude is also a clue to Shukla's operating style. His public posts favor the collective noun over the solitary founder mythology. When Lightbeam expanded its product portfolio and added sales leader Mike Gesnaldo in early 2025, Shukla framed the hire as a partnership with an existing team. When capital arrived, he returned to employees and early believers. This does not reveal what every Tuesday meeting feels like, and it need not. It does show the version of leadership he chooses to put on the record: technical ambition surrounded by named collaborators.
Four markers in a longer build
The bars show sequence, not financial scale. Products rarely respect the drama of a hockey-stick chart.
The risk arrives before the prompt
Generative AI changed the urgency, but not Shukla's underlying thesis. A chatbot, copilot, or autonomous agent can retrieve, summarize, and recombine enterprise information at a speed no employee can match. If its permissions are broad, the system can make quiet oversharing suddenly industrious. The failure starts before anyone types a prompt. It begins with forgotten files, loose access, shadow copies, and the inability to explain whose information is exposed.
“AI security starts with data security.”Himanshu Shukla, 2026
In August 2026, Lightbeam introduced AI Guardrails for Enterprise Data. The capabilities map AI tools and agents to the sensitive information they can reach, apply policies, identify excessive access, and record decisions for later review. The system can block, redact, warn, allow, or audit, depending on context. Later that month, the company announced its inclusion in a Forrester landscape covering data security platforms.
The useful part of Shukla's message is its refusal to make the model magical. He has argued that the central danger is the sensitive data AI can access, use, and expose. Models matter, but the enterprise still owns the permissions, records, contracts, code, and customer details beneath them. The new interface does not repeal the old housekeeping. It accelerates the consequences of avoiding it.
Watch Shukla unpack the data-security problemTechstrong TV interview with Alan ShimelThe discipline of one more relationship
Enterprise software often responds to complexity by making another dashboard. Shukla's career suggests a different reflex: find the relationship the dashboard is missing. In cloud infrastructure, it might be the connection between forecast demand and a scheduling decision. In data security, it is the line between a fragment and the person it describes. In AI governance, it is the line between an agent, its permissions, and every sensitive thing those permissions quietly unlock.
This is why his transition from engineer to founder feels less like a reinvention than a widening aperture. The objects changed. The system habit remained. Observe the whole, preserve context, automate the dull parts, and give the human operator a decision rather than a pile of alarms.
There is ambition in Lightbeam's “privacy-first world,” but Shukla's public persona stays close to the builder's bench. His social biography uses that single word: builder. His major funding announcement reached backward to the first lines of code and outward to the people who helped. Even the company's grandest claim resolves into a modest question that any employee can understand: before this data moves, do we know whom it represents?
The enterprise will keep accumulating nouns: copilots, agents, clouds, regulations, repositories. Shukla's bet is on the connective tissue. If the software can teach each scattered byte to introduce itself, security stops guarding anonymous objects and starts protecting recognizable relationships. For a field enamored of zero trust, remembering the person is a rather elegant place to begin.