Axios chief Jim VandeHei believes that over the next two years the world could face a chain of interconnected crises whose consequences would be difficult to predict and control. Among the main threats, he cites large-scale cyberattacks using AI, a correction in the artificial intelligence market, the accelerating development of neural networks, the expansion of the war with Iran, and risks related to the age of U.S. President Donald Trump.

According to VandeHei, the danger lies not so much in each threat individually as in their possible simultaneous emergence. One crisis could trigger another, and the speed of change could outpace the ability of companies and governments to respond.

The author emphasizes that this is a possible scenario, not a prediction of inevitable catastrophe. However, he says that conversations with company executives, government officials, and major investors have reinforced his concerns.

1. Cyberattacks could spiral out of control

VandeHei sees recent developments in cybersecurity as a warning of how serious future threats could become.

Critical vulnerabilities in PHP — a programming language that powers a significant share of websites — discovered with the help of AI led to disruptions in government resources in various countries. The U.S. National Security Agency urged organizations to install patches immediately.

Additional concern comes from reports of unauthorized AI agents that gained access to customer accounts at major banks. According to Axios, some of them attempted to perform actions that users had not authorized.

Anthropic deputy chief information security officer Jason Clinton warned that after several alarming days, the cyberthreat situation could get significantly worse before it begins to improve. Salesforce CEO Marc Benioff also reacted to the developments, calling the situation an emergency.

According to VandeHei, AI companies and banks are already preparing for possible large-scale attacks. Some representatives of leading AI companies estimate the probability of a serious cyber incident in the coming months as significantly above 50%. This is the assessment of the author’s sources, not a confirmed industry forecast.

2. The AI bubble could burst — and hit the entire economy

The second threat is tied to the investment boom around artificial intelligence. VandeHei notes that the current situation differs from the dot-com bubble of the late 1990s: AI development is being financed by large companies with enormous cash resources. But borrowed capital is playing an increasingly important role in these investments.

According to Goldman Sachs, about a third of the capital expenditures of the largest tech companies on AI in 2026 will be financed through debt.

This raises risks if the expected return on investment falls short of expectations. At the same time, the problems could spread far beyond individual companies: tech giants are simultaneously investing in one another, buying services from one another, and competing for the same customers.

VandeHei cites Meta as an example. The company reported strong financial results, but its plans to finance AI infrastructure sparked investor concerns. The market is worried not only about the scale of spending, but also about how easily companies will be able to raise capital to continue it.

A sharp correction could be triggered by many different events: signs of overbuilding data centers, insufficient AI adoption by companies, stronger Chinese competitors, falling prices for AI services, rising borrowing costs, or a political backlash against the spread of the technology.

In the author’s view, the probability of a market correction of more than 50% in the short term deserves serious attention. This is VandeHei’s subjective assessment, not a consensus forecast by financial analysts.

Even if the bubble does not burst all at once, declining investment and asset values could lead to spending cuts, layoffs, and increased political tension.

3. AI could begin improving itself

The third threat is linked to recursive AI self-improvement — a scenario in which systems help develop more advanced versions of themselves, accelerating further progress.

VandeHei claims that specialists inside Anthropic, OpenAI, and Google are already observing processes that may point to the development of this trend.

In the long term, such progress could bring significant benefits. But if AI capabilities begin to grow faster than expected, the consequences for the economy could be extremely painful as early as 2027.

More advanced systems will be able to automate tasks currently performed by humans more quickly, creating a threat to entire professions and business models. This could increase pressure on the labor market and accelerate political debates about regulation of the technology.

VandeHei suggests that Democrats in the United States may join Senator Bernie Sanders, who advocates slowing AI development or even temporarily halting its further advancement. In the author’s view, such a political turn could unsettle investors and increase market instability.

4. An escalation of the war with Iran could push oil prices to $120

Another risk is the expansion of the military conflict with Iran and its consequences for the global energy market.

According to VandeHei, aides to President Donald Trump consider a resumption of the war after the U.S. midterm elections likely. Iran, for its part, is warning of an expansion of targets in the event of retaliatory actions. Against the backdrop of rising threats, U.S. authorities have already reduced staffing at embassies in the Middle East.

A possible escalation is especially dangerous because of the tense situation in energy markets. If oil supplies come under additional pressure, prices could rise sharply, worsening inflation and increasing costs for businesses and households.

VandeHei suggests imagining the consequences of a prolonged period in which oil costs $120 per barrel. Such a scenario could intensify economic pressure on consumers and companies, complicate the work of central banks, and create new political problems for governments.

At the same time, the author does not claim that oil will necessarily reach that price: it is an example of the possible consequences of a prolonged escalation.

5. Trump’s age raises risks for decision-making

The fifth factor is tied to the age of the U.S. president. VandeHei specifically emphasizes that he is not trying to assess Trump’s health or make a political statement. According to him, this is about general medical and statistical risks associated with age.

Trump is 80 years old, and over the next two years he will be the oldest person ever to hold the office of U.S. president. With age, the likelihood of serious health problems, cognitive decline, and loss of capacity increases.

VandeHei believes that an additional danger is created by the absence of people around the president who could openly challenge him and provide unpleasant but necessary assessments. In his view, the potential consequences range from sudden health problems to decisions on the use of military force and the limits of presidential powers.

The author emphasizes that he would express similar concerns about Joe Biden if he were still president.

What to do amid growing uncertainty

VandeHei urges readers not to treat the listed risks as a reason to give up planning for the future. On the contrary, he believes that understanding possible changes will help people make more balanced decisions about career, finances, and safety.

He highlights three practical steps.

Learn about AI. It is important to understand not only the advantages of the technology, but also the risks associated with it. In addition, it is worth more carefully evaluating the actions of employers and elected officials amid rapid technological change.

Strengthen digital security. VandeHei advises using a password manager, creating unique passwords for different services, and switching to passkeys whenever possible. Email and phone access should be additionally protected with an authenticator app, since access to them often makes it possible to reset passwords for other accounts. Passwords and verification codes should not be shared with people or AI agents who initiate contact themselves.

Review investment risks. Those who have retirement savings or other investments should assess how heavily their portfolios depend on individual industries, especially the technology sector. Diversification can help reduce the impact of a sharp decline in one of the markets.

The main danger is not an individual crisis, but their combination

Each of the five scenarios may not materialize on its own. But VandeHei believes that their simultaneous buildup increases the likelihood of a serious shock that will be difficult to stop with conventional methods.

AI is accelerating not only technological development, but also the spread of mistakes, financial shocks, and cyberthreats. Military conflicts can amplify economic instability, while political decisions can change operating conditions for entire industries.

It is precisely the speed and interconnectedness of these processes that the author sees as the main challenge of the coming years. In his view, companies, governments, and individuals will have to prepare for a situation in which the consequences of events unfold faster than they can fully understand what is happening.