Analyzing recent incidents in the field of digital security, I highlight five key areas: from large-scale financial crimes involving cryptocurrencies to fundamentally new attack vectors using artificial intelligence. These are not just isolated cases, but clear indicators of the evolution of cyber threats.

Cryptocurrencies in the hands of transnational crime

Brazilian law enforcement struck a blow against an international drug syndicate that used cryptocurrencies to launder colossal revenues. The organization, linked to the smuggling of at least 6.5 tons of cocaine, had built a complex financial infrastructure. During a large-scale operation across four states, 44 searches were conducted and nine key figures were arrested.

The money laundering scheme included a network of shell companies, the purchase of elite real estate and, critically, the use of illegal crypto brokers to convert fiat funds into digital assets and transfer them across borders. As part of the investigation, assets worth up to 1 billion Brazilian reais (about $197 million) were frozen. This case is clear confirmation that digital currencies have become an indispensable tool for transnational organized crime, and this trend, it seems, will only intensify.

Attack on US critical infrastructure

In late July, unknown hackers carried out a coordinated attack on the operational technology of more than 30 water systems in Minnesota. The incident led to the shutdown of a water treatment plant in the city of Braham and equipment failures in other counties, forcing staff to switch to manual control.

State authorities activated cyber defense protocols, and the FBI and the US Cybersecurity Agency joined the investigation. Previously, US agencies had warned of targeted attacks by Iranian hackers on Rockwell Automation and Allen-Bradley programmable logic controllers used in water utilities and the energy sector. This incident confirms that attacks on OT systems have ceased to be a theoretical threat and have become real practice.

Ukrainian "crypto academy": social engineering in action

Ukrainian cyber police exposed a fraudulent organization operating under the guise of the "Ukrainian Financial Academy." The attackers, using social media advertising and a fake trading platform, stole $1.11 million from nearly a thousand people. The particular cynicism of the scheme lies in the fact that the victims were the most vulnerable categories of citizens — pensioners and the seriously ill. One of the victims, undergoing cancer treatment, transferred about 500,000 hryvnias set aside for therapy to the scammers.

The four organizers face up to 12 years in prison. This is a striking example of how classic "con" methods are adapting to crypto realities, exploiting not only financial illiteracy but also human desperation.

SparkKitty: a new level of mobile stealer software

Check Point specialists discovered a new virus, SparkKitty, which disguised itself as legitimate apps in the App Store and Google Play. Its main feature is not input interception, but scanning the smartphone's gallery using an OCR engine. The malware continuously analyzed photos in search of seed phrases, passwords, and QR codes, sending the recognized data to the attackers' servers.

On Android, the virus spread through the SOEX messenger (over 10,000 installations), and on iOS — through the 币coin app. This case is a critical reminder: storing wallet backups as screenshots is a grave mistake. Seed phrases should exist exclusively in offline format, on paper or hardware devices.

AI agent as autonomous weapon

The most alarming trend of the week is the use of the Chinese AI model DeepSeek by a hacker for fully autonomous attacks. Controlling the neural network via Telegram, the attacker under the nickname knaithe configured the agent to independently search for vulnerabilities, analyze exploits, and attempt breaches. The AI scanned the internet via FOFA, found fresh vulnerabilities on GitHub, and tried to apply them, demonstrating decision-making logic.

Although DeepSeek's autonomous attacks were not successful, in manual mode the hacker breached 11 Marimo objects and extracted data from three organizations through a vulnerability in NetScaler. In total, more than 460 targets were subjected to attacks. Researchers gained access to the attacker's infrastructure due to his mistake, which allowed a detailed study of the scheme. This case is just the tip of the iceberg. We are on the threshold of an era when AI will become a full-fledged tool of cyber warfare, and we need to be prepared for this right now.

My verdict: the week showed that cryptocurrencies and AI are becoming a double-edged sword. On the one hand, they open new opportunities for criminals; on the other, they give us new tools for protection. The key takeaway for investors: basic hygiene — offline storage of keys and a critical attitude toward any "guaranteed" investment offers — remains the only reliable defense.