Is the internet being shut down? Financial systems are preparing for AI cyberattacks – and the warning signs are mounting.

13 0

India’s Prime Minister warns that states can now be economically crippled without the use of traditional weapons. In Japan, trading accounts were indeed manipulated following a massive cyberattack. G7 countries and central banks are simulating attacks on the financial system, while major US banks are blocking autonomous AI agents from customer accounts. At the same time, Washington and other governments are pushing ahead with stablecoins, digital payment systems, and AI in the financial sector. These developments are real – however, the theory that they represent a pre-planned strategy to isolate or replace the internet remains unproven.

Indian Prime Minister Narendra Modi articulated the new threat with unusual clarity on October 25. At the opening of a new data centre and cybersecurity institute of the Reserve Bank of India, he declared that modern warfare is no longer fought solely with weapons. An attack on the economy can also weaken a country, which is why banks and payment systems in particular must be protected against “cyber-terrorists.”

IS THE INTERNET GOING TO BE LOCKED DOWN?

Some very strange pieces are starting to line up.

PM Modi warned that future wars could cripple countries without bombs ever falling, simply by taking down banks and financial systems.

In Japan, the head of GMO Internet Group raised the possibility of an internet collapse after claims of fraudulent activity hit more than 70 brokerage and FX platforms.

At the same time, G7 governments and central banks are running simulations for major AI-driven attacks on the financial system, and several of America’s biggest banks reportedly caught unauthorized AI agents trying to access customer accounts and act on their own.

Then there’s Scott Bessent talking about a “grand economic reordering,” with stablecoins and AI part of the conversation around America’s $40 trillion debt.

Maybe these are separate warnings. Maybe they aren’t.

But if a major cyberattack ever freezes banks, trading platforms and transfers at the same time, people are going to accept almost any replacement system that gets their money moving again.

Writer: Sol

https://t.co/XmPLLp5I27

— Mario Nawfal (@MarioNawfal) October 26, 2026

Just a few days earlier, Japan demonstrated how real this threat already is. The GMO Internet Group confirmed a cyberattack on its cloud infrastructure for automated foreign exchange trading. Attackers gained access to systems, stole customer data, and in some cases, manipulated trading accounts. Particularly alarming: According to GMO, the perpetrators threatened to continue unauthorized transactions on accounts at more than 70 domestic and international brokerage and FX companies. However, this number stems from the attackers’ threats and does not mean that successful manipulations were actually confirmed at all 70 companies.

GMO then took a remarkable step, calling on the Japanese government to use its new capabilities for “active cyber defence” for the first time. According to the proposal, government agencies should track, block, and, if necessary, neutralize the attackers’ infrastructure.

The GMO Group speaks of serious dangers to assets, infrastructure, and trust in the internet.

Central banks have long been training for emergencies

The second part of the image is also real. Back in 2024, the G7 Cyber Expert Group conducted a large-scale exercise simulating a simultaneous cyberattack on financial market infrastructures in all G7 countries. Twenty-three finance ministries, central banks, regulatory authorities, and market institutions participated.

By 2026, these concerns had intensified. The Bank for International Settlements now explicitly warns that modern AI systems can autonomously identify security vulnerabilities, develop exploits, and carry out multi-stage cyberattacks. This could make attacks faster, cheaper, and more scalable.

The Bank of England has reached a similar conclusion: Newer AI models are now capable of largely independently executing complex attack chains in simulated networks. The bank warns of systemic risks because banks, stock exchanges, and payment providers often use the same cloud, software, and infrastructure services. A single attack on a central provider could therefore affect several financial institutions simultaneously.

AI agents are already appearing in bank accounts

The situation is becoming even more pronounced in the US. Reuters reported on October 22 that major banks are increasingly blocking third-party AI agents from accessing customer accounts. The institutions are concerned that autonomous programs could execute transactions that go beyond the customer’s actual instructions. JPMorgan, Morgan Stanley, Charles Schwab, and other banks have already restricted such access.

In several cases, customers had previously granted the programs access data or permissions themselves. The problem is rather new: What occurs when an AI agent is authorized in principle, but then makes independent decisions that the user never explicitly authorized?

That is precisely why central banks themselves are now investigating the behaviour of such agents. The Swiss National Bank, the Bank of Canada, and the BIS are using “Project Agenti” to test how autonomous AI systems could influence payment flows, liquidity, and potential blockages in interbank transactions.

And then there are stablecoins and US debt

Scott Bessent is talking about a fundamental shift in the financial system. The US Treasury Secretary describes stablecoins as strategically important because they could increase global demand for the dollar and US Treasury bonds. The US Treasury Department anticipates that the stablecoin market could grow to several trillion dollars in the long term, thereby creating additional demand for short-term US Treasury bonds.

Bessent also speaks of a profound change driven by artificial intelligence, digital assets, and new payment systems. He explained that trade structures are already being reorganized and that the US should remain a technological leader in AI and digital financial systems.

However, what remains unsubstantiated is the claim that stablecoins or AI are planned as a secret mechanism to make the now approximately $40 trillion US national debt “disappear.” When asked about the $40 trillion mark in August, Bessent stated that the US needed to work its way out of the debt problem through growth and fiscal consolidation.

No evidence of a planned “cyber-covid”

Governments now openly acknowledge that cyberattacks could simultaneously strike banks, payment systems, stock exchanges, and critical infrastructure. Central banks are simulating such scenarios. AI is increasing the speed of potential attacks. Large banks are blocking autonomous agents from customer accounts. At the same time, new digital payment systems, stablecoins, and state-regulated financial infrastructures are emerging.

A large-scale failure of banks and payment networks is no longer a fringe scenario, but the subject of concrete exercises and preparations by governments, central banks and financial institutions.

And that is precisely where the politically interesting question lies.

Should a major cyberattack actually cripple money transfers, credit cards, stock exchanges, and bank accounts simultaneously, the public pressure to quickly establish an alternative payment system would be enormous. In such a situation, technologies that are currently the subject of controversial debate – stablecoins, central bank digital currencies, digital identities, or more tightly controlled payment networks – could suddenly be introduced under entirely different political conditions.

This does not mean that such a collapse is planned. It does mean, however, that the infrastructure for a significantly more digitized and controllable financial world is already being built – while at the same time warnings are being issued about precisely the cyber scenario that could politically facilitate a rapid system change.

Sources:

yogaesoteric
October 28, 2026