Crime•

Anchorage Co-Founder Expects a 'Slaughterhouse' Year for Crypto Security as AI Runs Attacks and North Korea Takes $7.5 Billion

Anchorage Digital's Diogo Mónica and Solana Foundation security chief Michael Coates warned in London that autonomous AI attacks will compound North Korea's thefts, which exceed $7.5 billion this year.

Written By :MINRK
Anchorage Co-Founder Expects a 'Slaughterhouse' Year for Crypto Security as AI Runs Attacks and North Korea Takes $7.5 Billion

Photo by Boitumelo on Unsplash

Mentioned in This Article

SOL
SOL
Solana
$69.18
2.59%

1. Security Chiefs Warn of AI Attacks and a Record North Korea Crypto Hack Total

Diogo Mónica, co-founder of crypto bank Anchorage Digital, told Solana's Breakpoint conference in London, in remarks made public on November 18, 2026, that the next 12 months will be "a slaughterhouse" for crypto security as artificial intelligence begins running attacks on its own. Michael Coates, chief information security officer of the Solana Foundation, said North Korea has stolen more than $7.5 billion this year.

The two executives delivered a combined message to an audience of developers and project teams: a new category of threat is arriving on top of one that is already costing the industry billions of dollars. Automated tools are making it cheaper and easier to break into systems, while the most capable state-backed thieves in the sector continue to succeed with methods that are well understood and still not reliably blocked.

Mónica described the coming year as "absolutely brutal." His assessment rested less on any single incident than on the pace at which AI systems have changed in recent months, and on what that change means for the number of people able to launch a credible attack.

2. How is AI changing crypto hacking?

AI systems have moved from spotting weaknesses to carrying out complete attacks without human direction, according to Mónica, and that shift has happened within the past three to six months. Where models were previously useful for finding vulnerabilities, he said, they can now build full exploits and run entire attacks autonomously.

He made the point on a panel chaired by Gordon Corera, co-host of the podcast The Rest Is Classified. The practical effect, Mónica argued, is that expertise is no longer a barrier. A person with no security knowledge can now ask a model to hack a website.

Coates reached the same conclusion from the defender's side. He said AI has sharply reduced both the cost and the skill required to hack. One consequence concerns projects that have never been attacked simply because nobody had bothered to look at them. Those projects, he said, will now be hit opportunistically.

That observation alters the risk calculation for smaller teams. Until now, a project with modest funds under management could reasonably assume that skilled attackers would spend their limited time on larger prizes. If scanning for flaws and writing the code to exploit them can be delegated to software, the attention of an attacker stops being scarce, and obscurity stops offering protection.

3. North Korea Remains the Largest Threat

Automation is the emerging danger, but Coates used a separate presentation at the conference to name the biggest existing one. He identified North Korea as the greatest threat facing crypto companies and set out the scale of what has been taken.

By his figures, more than $13 billion has been stolen over the past decade, and more than $7.5 billion of that total has been taken this year alone. About $6 billion of this year's sum came from one breach: the hack of Binance last month, which the FBI attributed to North Korea.

Those numbers mean that a single incident accounts for most of the year's losses, and that this year accounts for more than half of everything Coates attributed to the country over ten years. He said the stolen funds finance North Korea's programs, a point reflected in his description of the thefts as part of the country's economy.

The US government has continued to respond with financial measures. The Treasury Department recently sanctioned 11 people over a scheme in which North Korean information technology workers obtained jobs at foreign companies.

4. How did the Binance hack happen?

The Binance breach was a chain of familiar tactics, in Coates's account, not a novel technique. It began when a developer downloaded poisoned software after following a sponsored search result. In a second step, an engineer installed software on a work device during a job interview.

Together, those two lapses gave the attackers access to the company's network and, from there, a path to its signing services. The intruders did not steal private keys directly. They used their reach into the signing systems to authorize malicious transactions.

The distinction matters for how companies think about protecting funds. Much of the industry's security effort concentrates on keeping private keys from being copied or extracted. In the sequence Coates described, the keys never had to leave their protected environment. The attackers gained control of the systems that are permitted to request signatures, and those systems did the rest.

Each link in the chain is an ordinary event in the working life of a technology company: a search for a tool, a download, an interview with a prospective employer. That ordinariness is what makes the pattern hard to stop, since none of the individual actions looks like an attack to the person performing it.

5. Fake Interviews, Deepfakes and Borrowed Identities

Recruitment is a recurring entry point in the methods both men described. Coates called fake job interviews a "telltale tactic of North Korea," used alongside poisoned software tools and deepfake video calls. He said North Korean hackers have stolen $400 million through deepfake calls alone.

The approach runs in two directions. In one, a target who believes they are being interviewed for a job is persuaded to install software, as happened in the Binance case. In the other, operatives try to get hired by the companies they intend to rob.

Mónica explained what happens when that second route is closed to them. When operatives cannot be hired directly, he said, they use people in the United States to take the jobs. Remote keyboards and mice are then attached to the equipment, allowing the attackers to operate inside the company's network from elsewhere while the employer believes it is dealing with a domestic employee.

The arrangement defeats checks that rely on the location or identity of the person who was hired, because the person who passed those checks is real. What the employer cannot see is who is actually controlling the machine.

6. How Anchorage Reduced Its Exposure

Anchorage has changed its own practices in response to the infiltration risk. The company now onboards new staff in person at its office in the United States, a step that removes the possibility of a new hire who never physically appears.

It has also restricted the hardware employees use. Every member of staff works from a locked-down Chromebook, which serves only as a means of logging into remote development machines. Mónica said the setup eliminates 90% to 95% of the company's corporate attack surface.

The reasoning follows from the attack chain Coates outlined. If an employee's device cannot run arbitrary downloaded software, then a poisoned tool from a search result or an installer supplied during a fake interview has far less opportunity to take hold. The sensitive work happens on machines the company controls centrally, and the device in the employee's hands is reduced to a window onto them.

Neither measure depends on new technology. In-person onboarding and tightly restricted laptops are administrative choices, and Mónica presented them as a direct answer to the hiring and download tactics that he and Coates had attributed to North Korean operatives.

7. A Free Framework and AI Used for Defense

For teams without the resources of a crypto bank, Coates pointed to STRIDE, a free framework that the Solana Foundation launched in the spring. It sets out 18 core controls for crypto projects.

Adoption among the largest Solana applications has been broad. Protocols representing more than 74% of the total value locked in Solana decentralized finance have gone through the framework, Coates said, including 10 of the 11 biggest. Total value locked measures the assets deposited in a protocol's smart contracts and is the standard gauge of how much money a decentralized finance application is responsible for.

The foundation is also turning the technology behind the new threat to defensive use. It employs AI to reverse-engineer programs soon after they launch on the network, with the aim of finding flaws before attackers do. Coates said that work prevented more than $5 million in potential losses over the past month.

The same capability that lets an unskilled attacker probe a project can therefore let a defender examine far more code than a human team could review by hand. The contest Coates described is partly a race over who points those tools at a new program first.

8. The Basics That Still Fail

Both executives closed on a less dramatic theme: many successful intrusions begin with elementary mistakes. Mónica singled out the practice of connecting email inboxes to AI agents, which he said invites prompt injection. In such an attack, instructions hidden in content the agent reads, such as an incoming message, cause it to act on behalf of whoever wrote them.

Coates urged teams to "protect the keys." His recommendations were specific: use hardware security keys, configure multisignature wallets properly, and have plans in place for rotating keys and for losing them. A multisignature wallet requires approval from several separate keys before funds can move, which limits the damage if any one of them is compromised, provided it is set up correctly.

He summarized the pattern he sees across incidents. Attackers are getting in through basic weaknesses and then pivoting toward the keys. The warning from London, taken as a whole, is that the coming year will bring a larger number of attackers with better tools, and that the defenses most likely to matter are ones the industry already knows how to build.

Related Articles

NEWSLETTERS

Don't miss another story.

Subscribe to the MINRK Newsletter today.

Tick the newsletters you want. No selection? You'll get MINRK Headlines.

By signing up, you will receive emails about MINRK products and you agree to our terms of use and privacy policy.

MINRK Daybook

Market analysis for crypto traders and investors.

EVERY WEEKDAY

MINRK for Advisors

Defining crypto, digital assets and the future of finance for financial advisors.

EVERY THURSDAY

MINRK Protocol

Exploring the tech behind crypto one block at a time.

WEEKLY

MINRK Long & Short

A must read for institutions. Insights, news and analysis delivered weekly.

EVERY WEDNESDAY

MINRK Headlines

The biggest crypto news and ideas of the day.

EVERY WEEKDAY

MINRK Policy

Examining the intersection of cryptocurrency and government.

WEEKLY

MINRK Research

Join thousands of readers who rely on MINRK for data-driven insights on the latest digital asset trends.

MONTHLY