Some days in security are defined by a single incident. June 18, 2025, was defined by a pattern. Within roughly 24 hours, an Iranian crypto giant had its hot wallets drained and torched, tens of thousands of Fortinet firewalls were confirmed compromised at scale, a Swiss banking giant saw employee data surface on the dark web via a procurement vendor, and researchers flagged fresh multi-stage malware hiding in the Python Package Index. Four different sectors, one common thread: the weakest link was never the front door.

Nobitex: when a hack is arson, not robbery

Iran’s largest cryptocurrency exchange, Nobitex, confirmed unauthorized access to part of its reporting infrastructure and hot wallets, then watched on-chain investigators trace outflows that quickly climbed from an initial $48.65 million on Tron to more than $81 million across Bitcoin, Dogecoin, and EVM-compatible chains . Blockchain security firm Elliptic put the total closer to $90 million and attributed the operation to the pro-Israel hacktivist group Gonjeshke Darande, also known as Predatory Sparrow .

The detail that matters most for defenders is what happened next. Instead of laundering the funds, the attackers routed them into vanity addresses embedded with anti-IRGC messages — strings so long that generating usable private keys for them is, in Elliptic’s words, computationally infeasible . In other words, the crypto was deliberately burned. This was not financially motivated crime; it was destructive, geopolitical signaling — and it followed a reported cyberattack on Iran’s state-owned Bank Sepah a day earlier . When nation-state and hacktivist motives converge, data destruction and reputational damage become the objective, not the payday.

FortiBleed: 75,000 firewalls, and no zero-day required

The same week, threat intelligence firm Hudson Rock disclosed what researchers dubbed “FortiBleed”: verified login credentials for roughly 75,000 Fortinet FortiGate firewalls and SSL VPN gateways, spanning 194 countries, circulating after being harvested by infostealer malware . Fortinet confirmed that its firewall and VPN gateway products had been targeted in a credential-theft campaign, noting that the activity aimed to illegally obtain user credentials rather than exploit a software vulnerability .

The uncomfortable lesson is that there was no exotic CVE to patch. The credentials were real, current, and independently verified — a reminder that the security appliance protecting your network is only as strong as the endpoint hygiene of the administrator logging into it. Edge devices remain a top-tier target precisely because compromising one grants trusted, encrypted access to everything behind it .

UBS and ChainIQ: your vendor’s breach is your breach

Also on June 18, more than 130,000 UBS employee records were reportedly published to the dark web after a ransomware attack against ChainIQ, a third-party procurement services provider. The dataset reportedly included direct-dial phone numbers for senior leadership, including CEO Sergio Ermotti, with the Hunters International group linked to the leak . The bank’s share price dipped in pre-market trading as the story broke .

This is the supply-chain problem in its purest form. UBS was not breached at its own perimeter; a business partner with legitimate access to employee data was. Every vendor with privileged data access is an extension of your attack surface, and contracts that do not mandate breach notification timelines, data minimization, and post-incident audit rights are liabilities waiting to mature.

PyPI: the long tail of open-source risk

Rounding out the day, researchers disclosed a multi-stage malware attack on the Python Package Index (PyPI), with a malicious package sitting in the public repository . Open-source registries are the connective tissue of modern software — and they remain a high-leverage, low-cost entry point for attackers who know that most teams install dependencies far faster than they review them.

Meanwhile, in AI: faster models, fewer desks, and a humanoid named AEON

The AI news cycle did not pause for the breaches. OpenAI CEO Sam Altman said GPT-5 could arrive “sometime this summer,” while Google rolled out its Gemini 2.5 Pro, Flash, and Flash-Lite models . Amazon CEO Andy Jassy told employees in a memo that generative AI will meaningfully reduce corporate headcount in the coming years . And Hexagon unveiled AEON, an industrial humanoid robot built on its measurement, AI, and autonomous-systems expertise, at its global Hexagon LIVE event .

Read together, the day’s stories point in one direction: AI is expanding the attack surface faster than most organizations are expanding their defenses. Model rollouts, AI-driven workforce changes, and robots entering physical plant operations all introduce new identities, new data flows, and new trust relationships to secure.

What security teams should do this week

June 18 will not be remembered as the day of one breach. It will be remembered as the day the industry got another clear look at how modern compromise actually works: through trust, not force. The organizations that come out ahead will be the ones that treat every partner, package, credential, and model as part of the perimeter — because they are.

Are you ready to secure your business?

Let our security experts review your current posture and identify hidden risks.

Get a Free Security Audit