The Silicon Age Collapse: Systemic risks to complex digital civilization
The first part of an ongoing series exploring threats that could trigger catastrophic cascading failures across the modern world.
Just over three millennia ago, civilization in the Mediterranean suffered a systems collapse so severe it took centuries to recover. Societies that depended on complex trade networks and fragile supply chains were brought to their knees as a perfect storm of war, famine, climate change, and natural disasters shattered many of the world’s great powers.
We now call this catastrophe the Bronze Age Collapse. It marked the end of an era named after the metal ancient societies relied upon to manufacture tools, armor, and weapons. As the dust settled, cities were left in ashes. Literacy disappeared across much of the region and took centuries to recover. Human progress slowed as the peoples of Europe, Asia, and northern Africa picked up the pieces.
Today, we depend on a vastly wider range of resources than the societies that sprang up around the Bronze Age Mediterranean. Our global society depends on vast supply chains, chaotically intertwined financial systems, and hard-to-maintain critical infrastructure that frequently runs on old, insecure software. In a globalized world, the complexity of the interconnections between industries is almost impossible to map - let alone rebuild after a major disruption.
In the worst-case scenario, a severe disruption to global systems could cause starvation on a biblical scale. It could rapidly degrade our ability to safely operate hazardous facilities such as nuclear power stations, risking environmental disasters. On a smaller scale, an organization hit by a risk swarm or swept up in a dependency cascade could watch its future wiped out in days, or even hours.
In the 21st century, vulnerable global systems and accelerating technological change mean that catastrophic failure is not a distant concern but a very real risk that warrants serious attention. The good news is that, unlike the people of the Late Bronze Age, we can study these vulnerabilities before disaster strikes.
In a new series focused on systemic risk, Machine will speak to the people working to understand and address threats to the infrastructure which powers digital civilization. This topic will become a core editorial focus for Machine.
We begin with this anchor piece: a scene-setting exploration of some of the most significant systemic risks facing humanity. This article will be updated as we learn more, new threats emerge, and fresh research reshapes our understanding.
Our systemic risk series starts with a civilizational frame, but the issues raised also have organizational relevance, making this topic salient to decision-makers at every level, from enterprises and startups to local councils and central government.
This subject matter may appear doomy. But the purpose of our series is not pessimism, but resilience.
The people of the Bronze Age did not know what was coming. We do. Which means there is still time to make sure a Silicon Age Collapse never comes to pass.
Dependency cascades
Digital civilization is not built from isolated systems, but from chains of dependence in which everything relies on something else. For example, to stay operational, a supermarket needs access to a vast array of systems ranging from payment networks to refrigeration technology - not to mention roads, fuel, and electricity.
It also depends upon banks, which in turn rely on cloud infrastructure, identity systems, payment rails, and shared interconnections with other financial institutions. When one node fails, the damage quickly spreads laterally into seemingly unrelated sectors.
This is what we call a dependency cascade: a failure that propagates through interconnected systems. A cyberattack on a cloud provider could disrupt logistics operations, interrupting food supplies to supermarkets. Feedback loops then amplify the damage, creating further shocks.
READ MORE: The evolution of resilience: Driving AI innovation without increasing risk
Dependency cascades are difficult to prevent because no single organization owns the whole chain. When each participant optimizes for its own efficiency, the system as a whole becomes faster, cheaper, and more brittle.
We recently saw the impact of a dependency cascade on a relatively small scale after a "thermal event" at an AWS data center triggered failures that caused the crypto exchange Coinbase to halt trading. In simple terms, a server overheating halted the operation of a key financial market.
That is systemic risk in miniature: a failure in one place producing consequences far beyond its point of origin. Combined with the concentration risks explored in the next section, dependency cascades can quickly become catastrophic.
Concentration risks
Our world depends on a frighteningly small number of companies, facilities, and platforms.
The internet appears decentralized. In reality, much of it runs through a handful of cloud providers operating on semiconductors manufactured by an extraordinarily small number of advanced fabrication plants. Some of these facilities sit in geopolitically sensitive regions such as Taiwan, where a single company produces an estimated 90% of the world's most advanced semiconductor chips.
Manufacturers, telecommunications networks, software platforms, banks, hospitals, governments and retailers all rely on much of the same underlying infrastructure.
This concentration delivers enormous efficiencies. Large technology providers can invest in resilience, security, and scale far beyond the reach of smaller actors. But it also creates a dangerous form of systemic risk known as common-mode failure: the simultaneous disruption of multiple systems due to a single shared cause.
If thousands of organizations depend on the same provider, they also inherit the same vulnerabilities. A cloud outage, cyberattack, or operational failure can quickly spread across sectors and national borders.
Regulators are becoming increasingly concerned about this concentration. Earlier this year, the Bank of England warned that cloud outages and distributed denial-of-service attacks could create cascading systemic failures across the financial system.
A similar pattern is emerging in artificial intelligence. Although open-source models are showing promise, the most influential systems are being developed by a relatively small group of companies.
Policymakers and industry leaders have warned that this concentration may pose risks not only to competition but also to economic resilience, national security, and financial stability.
Civilizations become fragile when too much depends on too little. The concentration of critical technologies may prove to be one of the defining systemic risks of the Silicon Age.
Infinite attack surfaces
Every layer of convenience creates another layer of vulnerability.
The attack surface of civilization is expanding faster than we can secure it. Every connected device, API, AI agent, and industrial control system creates another potential pathway for failure or exploitation.
For most of human history, critical infrastructure was isolated, local, and difficult to access. Today, power stations, ports, hospitals, water utilities, factories, vehicles, and financial systems are connected to digital networks. Systems that were once physically separated are now linked together.
READ MORE: Bank of England warns of risks lurking in “opaque and hidden corners” of the financial system
AI is accelerating this trend. The World Economic Forum’s Global Cybersecurity Outlook 2026 found that 87% of respondents identified AI-related vulnerabilities as the fastest-growing cyber risks facing organizations.
This trend extends far beyond traditional security and deep into the real world. Researchers recently identified nearly 70,000 internet-exposed operational technology devices that control critical industrial processes worldwide, many of which run outdated software with known vulnerabilities. As factories and industrial facilities become increasingly automated, that number is likely to grow.
Meanwhile, the boundary between digital and physical infrastructure is disappearing. According to Forescout’s threat research, attacks targeting operational technology protocols surged by 84% during 2025, while exploitation of connected IoT devices continued to rise.
Organizations - and civilizations - become less resilient when too much is exposed.
Unmanageable complexity
The complexity of the world today is both its greatest strength and a source of profound vulnerability. As Joseph Tainter, the historian of collapse, observed: "More complex societies are more costly to maintain than simpler ones."
Complexity is often mistaken for progress. In reality, it is a trade-off. Every new layer of technology, regulation, infrastructure, and specialization can make a system more capable. It can also make it harder to understand, maintain, and repair.
No individual understands the global financial system. There is no single engineer who fully understands the software stack powering a hyperscale cloud provider. And no government comprehends the full web of dependencies connecting global supply chains, communications networks, energy systems and critical infrastructure.
When Machine spoke to Dynatrace Vice President Bob Wambach, we asked whether organizations are now operating systems beyond human understanding or control. His answer was blunt: “This has already happened.”
Enterprises contain thousands of interacting services and dependencies that no individual can fully grasp. As AI becomes increasingly involved in software development, infrastructure management, and business operations, that challenge is likely to grow. Organizations will find themselves relying on systems that are partly designed, optimized, and operated by machines.
The Bronze Age Collapse demonstrated how interconnected systems can fail in ways that contemporaries neither anticipated nor understood. The Silicon Age may face a similar challenge. Not because our systems are weak, but because they have become so intricate that nobody can confidently predict how they will behave under extreme stress.
The end of manual fallback
For centuries, complex systems retained a simple safety mechanism: humans.
When machines failed, people stepped in. Human judgment acted as a buffer between technical failure and systemic collapse.
That safety barrier is disappearing.
Organizations are removing manual fallback procedures in pursuit of efficiency, scale, and automation. Financial markets are dominated by algorithmic trading, just as logistics networks are coordinated by software. As dependence on automation grows, fewer people retain the authority required to intervene when systems fail.
READ MORE: Ex-Lord Mayor “alarmed” over government's response to Atlantic undersea cable risks
The danger is not that automation makes mistakes. Humans make mistakes too. The risk is that highly automated systems can fail at machine speed while reducing opportunities for human intervention.
This challenge is becoming more acute as agentic AI enters the workplace. Researchers and industry leaders warn that organizations are handing decision-making authority to systems whose reasoning is difficult to inspect and whose behavior can be unpredictable.
As these systems become embedded in critical infrastructure, finance, and business operations, the question is no longer whether humans remain in the loop, but whether they remain capable of taking control when required.
The efficiency gains are immense. But when automated systems fail, organizations may discover that the opportunity for human intervention has disappeared.
Debasement of human capital
Large-scale systems, such as human societies, do not run on technology alone. They run on people who understand it.
Every critical system depends on specialists whose knowledge has often been accumulated over decades: mainframe engineers, power-grid operators, industrial control system experts, semiconductor engineers, network architects, and countless others. Much of this expertise is undocumented, difficult to transfer, and impossible to replace quickly.
Across industries, automation and AI are taking over entry-level work traditionally performed by junior staff. These roles have historically served as training grounds for future experts. If fewer people learn the fundamentals, fewer people will eventually develop the deep knowledge required to maintain the systems on which civilization depends.
The problem is already visible in sectors that rely on aging technologies. Banks, governments, and insurers continue to depend on mainframes and COBOL systems built decades ago, while the pool of specialists capable of maintaining them continues to shrink. Similar concerns exist across energy, manufacturing and critical infrastructure, where experienced engineers are retiring faster than they can be replaced.
As organizations rely more heavily on AI-generated code, automated operations, and autonomous agents, the risk is that fewer people develop the expertise needed to understand the systems operating beneath them.
Knowledge is a dependency. When it disappears, rebuilding can take decades.
Natural System Shocks
Some of the greatest threats to humanity are well-known and inevitable.
Floods, pandemics, earthquakes, volcanic eruptions, and extreme weather events have shaped human history for millennia. Unlike many technological risks, their existence is not disputed. The only uncertainty is when they will occur and how damaging they will be.
The problem is that civilization has become increasingly dependent on fragile technological systems that are vulnerable to natural shocks.
A severe geomagnetic storm similar to the 1859 Carrington Event could damage satellites, disrupt GPS services, interfere with communications, and potentially affect power grids across multiple continents. Governments and researchers have repeatedly warned that a major space-weather event could produce economic damage measured in trillions of dollars.
Natural disasters become systemic risks when they collide with complex technological societies. A drought becomes a food crisis. If a cyberattack brings down a major shipping company at the same time, a supply chain disruption can cause a famine.
The Bronze Age Collapse was shaped in part by environmental shocks interacting with societies already under strain. The lesson remains relevant today.
Nature does not need to defeat civilization. It just needs to test it.
Risk swarms
The greatest systemic risks rarely strike alone.
The Bronze Age Collapse was not caused by a single event. War, famine, migration, environmental change, and political instability interacted with one another until entire societies lost the ability to adapt.
Researchers describe this phenomenon as a polycrisis: a situation in which multiple problems become entangled, amplifying one another in ways that make the whole far more dangerous than the sum of its parts.
Modern societies may be particularly vulnerable to this dynamic. A cyberattack can occur during an energy shortage. A natural disaster can disrupt supply chains already weakened by geopolitical tensions. A financial shock can coincide with a major technology outage. Each event increases the pressure on systems already operating under strain.
When multiple shocks arrive simultaneously, organizations and governments can find themselves fighting several crises at once while losing the capacity to respond effectively to any of them.
This may prove to be the defining challenge of the Silicon Age. Not a single existential threat, but swarms of interacting risks that overwhelm systems designed to deal with problems one at a time.
Get in touch if you're working in systemic risk and have a story to share.
