CyberKord Raises €2.7M as Telecom and Security Teams Bet on AI Operations

CyberKord was founded in Milan in 2026 by Alessandro Maggio, Dragos Tanasa and Cosimo Michelagnoli. It says its platform uses AI and a "Digital Twin" of a company's IT infrastructure to analyse issues, simulate fixes and apply corrective actions. The round was announced on 3 November. 360 Capital led it, with Plus Partners, Addendum Capital and Vento participating, according to Tech.eu.

The pitch rests on a gap the company describes in plain terms. Security tools can flag a potential threat within minutes. Resolving it takes far longer because people have to step in. CyberKord wants to automate that second half.

Why the money is moving now

That gap is not unique to Italy. The Hacker News ran a long piece on 29 October. It described how AI is taking over alert investigation, evidence collection, correlation across tools and response actions inside security operations centres. It calls the emerging human role the "Threat Operator": an AI-native professional who sets priorities and decides how much autonomy the AI gets. That person also judges whether the operation is actually improving security. The shift it describes is from human-in-the-loop to human-on-the-loop. Staff supervise AI-driven operations rather than running every alert by hand.

Vendors are already selling into that idea. CyberKord is small, at €2.7 million, but larger budgets show the direction of travel. The question for buyers is whether automation shortens resolution time or simply moves the bottleneck.

The network side of the same problem

Telecom operators face a version of this inside their own networks. Data Center Knowledge reported on 30 October on edge AI inference. That means running workloads closer to users rather than only in hyperscale data centres. Vendors pitch it as a way for telcos to move closer to the centre of AI infrastructure. The article quotes David Duggal, CEO of EnterpriseWeb, saying power and thermal concerns make AI at a tower or cell site a "Catch-22 situation". It points to metro sites such as central offices as the clearest near-term opportunity, where power, fibre and proximity are more practical.

Nvidia's edge work, the article notes, involves AT&T, Charter Communications, Cisco, Nokia, SoftBank, Telefónica and T-Mobile, among others. The GPU infrastructure attached to AI-RAN is also expected to help operators automate their own network operations. That is the same promise CyberKord makes at company level, applied to a much bigger estate.

"Power and thermal concerns make AI on a tower or cell site a Catch-22 situation," Duggal told Data Center Knowledge.

Operators have a second reason to care about automation: their results. Virgin Media O2 said on 3 November that third-quarter revenue fell 5.7% to £2.43 billion, according to TelecomLead. Capex was £575 million and 5G+ coverage reached 86%. The same outlet reported the company will cut thousands of jobs and target £800 million in savings by 2029. Cost pressure of that kind makes any promise of fewer manual hours worth a hearing, and worth scepticism.

Where the money is going elsewhere

The largest capital movements are not in security tooling but in AI infrastructure and energy. CNBC reported on 2 November that OpenAI signed agreements with Abu Dhabi-based entities spanning energy, government and investment. These include a deal with ADNOC and G42 to deploy OpenAI frontier technology across ADNOC's operations. They also include a five-year agreement between Inception42 and OpenAI covering government, enterprise and education. CNBC also reported, on the Saturday before, that OpenAI is seeking up to $20 billion in a new round co-led by UAE investors.

Sam Altman, OpenAI's CEO, said in a statement carried by CNBC: "Abu Dhabi and the UAE are setting a bold, global standard for AI integration across both the public and private sectors to make sure the benefits of AI reach everyone."

That is a different market from CyberKord's. It still sets the price of the compute and models that smaller security vendors build on. When the foundation layer is being financed at that scale, a €2.7 million pre-seed is a rounding error. The startup's case has to rest on workflow and data rather than model access.

Regulators and spectrum

On the policy side, the Federal Communications Commission voted on 29 October, Light Reading reported. It removed a decades-old restriction that prevented drones from operating in 50 megahertz of low-band spectrum in the 800 MHz band. The order covers command, control, communication and payload operations for unmanned aircraft systems. The FCC adopted rules for 800 MHz cellular service in 1981 and prohibited airborne use in 1991 to limit interference with terrestrial networks. It now says modern networks can prevent that interference.

Spectrum decisions of this kind change what operators can sell, and therefore what they can afford to automate. They also add to the list of connected devices whose traffic has to be monitored.

Hardware and the cost of AI investment

Unitree Robotics, the Hangzhou-based robot maker, reported on 28 October a 69.93% year-on-year rise in third-quarter revenue to 665 million yuan (US$99.7 million), according to the South China Morning Post. Net profit excluding non-recurring items fell 17.42% to 105 million yuan. Research and development spending jumped 249.58% to 127.04 million yuan, about 19% of revenue. For the first three quarters, net profit attributable to shareholders was 387 million yuan, up 270.54%, the filing with the Shanghai Stock Exchange said.

The pattern is familiar from telecom results: revenue can grow while the cost of building AI capability eats into profit. Xiaomi's planned battery investment shows a similar willingness to spend for strategic position. CnEVPost reported on 30 October that Xiaomi plans to invest 1.7 billion yuan ($253 million) for a 5.34% stake in Sunwoda Electric Vehicle Battery. The move, through subsidiary Hanxing Venture Capital, makes Xiaomi its third-largest shareholder. Sunwoda will lose control of the unit and stop consolidating it. It estimates the accounting treatment will lift 2026 net profit by 3.2 billion to 4.5 billion yuan, subject to audit.

What the funding does and does not prove

CyberKord's round is a signal about where investors think manual work can be removed. It is not evidence that it has been. The company has disclosed no customer numbers, no revenue and no measured reduction in resolution time. Tech.eu's report describes the product's intended function, not deployed results.

Buyers should also note the shape of the market it is entering. Large platform vendors are pursuing security operations automation. So are AI labs selling directly into enterprises and governments, as the Abu Dhabi agreements show. A pre-seed company has to prove it can integrate with the tools operators already run. It also has to show that its Digital Twin reflects a real estate rather than a diagram of one.

For now, the clearest fact is the cheque: €2.7 million, led by 360 Capital, announced on 3 November. Whether it buys a shorter path from alert to fix is the part that has not been demonstrated.