01 / WHO THEY ARE NOWThe fixer, not the smoke alarm
A security analyst opens her laptop on a Monday. There are 4,000 alerts waiting. She will read maybe forty. This is the quiet scandal of modern data security: the tools are very good at shouting and very bad at doing anything about it. Teleskope was built for the part everyone else skips - the fixing.
Out of an office on Park Avenue South in New York, Teleskope runs what it calls the first agentic data security platform. Translation: software that behaves like a security team rather than a dashboard. It crawls through a company's cloud, its SaaS apps, its forgotten databases, finds the sensitive data, and then - the interesting part - acts on it. Redacts. Masks. Encrypts. Deletes. No human required to approve each move.
By late 2025 the company had a $25 million Series A in the bank, customers like Ramp and The Atlantic, and a claim that, for once, is testable: 85% of the companies that pilot it end up paying. That is a high bar to clear in enterprise security, where pilots usually go to die.
02 / THE PROBLEM THEY SAWEverybody could see the leak. Nobody could plug it.
Here is the dirty secret about sensitive data: most companies do not know where theirs is. Home addresses end up in a log file. A spreadsheet of medical details lands in the wrong bucket. Customer records get copied into a test environment and forgotten. The data does not stay where you put it - it sprawls.
The previous generation of tools - think AWS Macie, BigID, Varonis - got reasonably good at the first half of the job. They could tell you that you had a problem. They were considerably less good at the second half, which is the half that matters. They flooded teams with false positives, struggled to scale, and then handed the actual cleanup back to the humans. Which is a bit like a fire alarm that rings beautifully and then asks you to put out the fire yourself.
03 / THE FOUNDERS' BETTwo engineers who had already done it once
Elizabeth Nammour and Julie Trias met on Airbnb's data security team. Nammour, a University of Pennsylvania computer science grad and first-generation American, had landed in security through an internal "reverse pitch" and was handed an unglamorous task: corral the booking platform's enormous, sensitive data sprawl. Trias - a military veteran and former Linux kernel developer - was one of the early engineers on Airbnb's site reliability team before moving to data security.
Together they built internal tooling that did the thing the market could not. Nammour wrote about it. Security teams and investors read it, then wrote back with the same complaint: we want this, and we cannot build it ourselves. That was the bet - that what worked inside one well-resourced tech company could become a product for everyone who lacked Airbnb's engineering bench.
They left in 2022 to find out. By their own admission they were green at the venture game - Trias has said they did not even know what a design partner was. They learned fast.
The name is a tell. A telescope exists to see the things too far away to spot with the naked eye. Teleskope's whole pitch is seeing the data nobody else can find. Subtle, it is not.
04 / THE PRODUCTSmall models, big context
Most AI startups reach for one enormous model and hope. Teleskope went the other way. It runs several smaller, fine-tuned models, each pointed at a specific kind of data - one for conversational text, one for structured tables, one for code and log files. Nammour's logic is refreshingly unmagical: break the problem into smaller pieces and you get both more accuracy and more speed.
The platform follows a three-beat rhythm. Understand: discover and classify sensitive data across cloud, SaaS, and on-prem, including the shadow data nobody remembers creating. Decide: read the company's own policies and translate them into risk-based action. Enforce: redact, mask, encrypt, isolate, or delete - and tighten access that was too generous. It detects more than 150 distinct types of sensitive information out of the box, and can be trained on custom ones.
For the nervous security buyer, there is a deliberate design choice: it ships air-gapped and single-tenant. Your data does not wander off to a shared cloud to be analyzed. The watchdog, sensibly, stays inside the house.
The short, fast history
05 / THE PROOFThe receipts
Claims are cheap in security software. Teleskope's case rests on a few numbers that are harder to wave away. The 85% pilot-to-paid conversion is the one that should make competitors uncomfortable - it means the product tends to survive contact with a real environment. Growth ran 600% year over year heading into the Series A. And the customer list skews toward companies that have a great deal to lose.
Three rounds, one direction
Total raised to date: roughly $32.2M. The line, as they say, goes up.
M13 led the Series A and put Managing Partner Karl Alomar on the board - the operator who helped scale DigitalOcean from startup to a NYSE listing. Repeat backers Primary Venture Partners and Lerer Hippeau came along again, which is its own kind of vote.
06 / THE MISSIONGive every company the team it cannot hire
Airbnb could afford to build its own data security tooling. Almost nobody else can. Teleskope's mission is to close that gap - to hand a mid-sized company the equivalent of an always-on, embedded data security team, without the headcount or the in-house platform engineering.
That mission has a deadline attached, and the deadline is AI. Every model that gets trained, every pipeline that gets built, multiplies the places sensitive data can hide and leak. The companies racing to ship AI are, often unknowingly, racing to scatter their data wider than ever. Teleskope is positioning itself as the cleanup crew for exactly that mess - sanitizing datasets before they feed a model, governing what flows through AI workloads.
07 / WHY IT MATTERS TOMORROWThe alarm that finally answers itself
Skeptics have a fair question. Plenty of companies have promised "automated remediation" and quietly meant "a slightly nicer dashboard." Trusting software to delete or alter production data on its own is a real leap, and security teams are paid to be paranoid. Teleskope's answer is its architecture - small accountable models, air-gapped deployment, policies the customer writes - and its conversion rate, which suggests buyers are actually making the leap.
Return to that analyst and her 4,000 Monday-morning alerts. In the world Teleskope is building, most of those never reach her, because most of them were already handled. The sensitive field was redacted overnight. The over-permissive access was tightened before anyone exploited it. The forgotten copy of customer data was found and dealt with. The smoke alarm has, at last, learned to carry a fire extinguisher.
Whether the rest of the industry follows is the open question. But the bet two Airbnb engineers made in 2022 - that detection without remediation is half a product - looks, four years and $32 million later, like the right one.