IT crisis and current impact on the digital economy

Hardware Guide » Buying Guides » IT Crisis and Current Impact on the Digital Economy

Information technology crises are no longer isolated events that occur from time to time: they form a kind of common thread running through the recent history of the digital economy. From the Y2000K bug to the global blackout caused by a simple security patch, and including financial bubbles linked to the internet, cryptocurrencies, and now artificial intelligence, each shock highlights how dependent we are on a few providers and an increasingly complex infrastructure.

In recent years, the world has experienced a series of challenges, including pandemics, hardware disruptions, the rise of cloud computing and AI, new cybersecurity regulations, and massive threats . This has led to a scenario where three fronts converge: the economy (the AI ​​bubble and stock market volatility), core technology (chips, memory, data centers, legacy systems), and risk management (cyberattacks, global downturns, the 2038 effect). Understanding how these fronts intersect is key to anticipating what might go wrong next and, above all, how to prepare.

Hardware crisis driven by Artificial Intelligence

The current wave of generative AI has exposed a structural problem: training and running giant models requires enormous data centers equipped with highly specialized and expensive hardware, and this is straining the entire technology supply chain. What began as a stock market rally for manufacturers like NVIDIA, AMD, and memory suppliers has ended in a perfect storm of shortages, soaring prices, and a diversion of resources from the consumer market to the corporate and AI sectors.

In this context, AI-focused data centers require massive computing power, enormous amounts of memory and storage, and ultra-fast networks . GPUs like the NVIDIA H100/Blackwell, AMD Instinct accelerators, or Google TPUs operate with reduced precision but handle colossal volumes of operations, far exceeding what a CPU can efficiently manage. Added to this is a relentless demand for DRAM and HBM memory to power these GPUs, as well as high-capacity SSD and HDD storage infrastructures , connected via 400 Gbps or higher networks with advanced Ethernet or InfiniBand.

All of this has led to a large portion of the chip and component stock being directed towards AI , relegating home PCs, gaming, and other consumer devices to a secondary role. It's not that raw materials are lacking; the bottleneck lies in the factories: building a new semiconductor plant costs billions and takes between three and five years, so supply isn't keeping pace with the growing demand for AI.

The immediate consequence is a sharp increase in electricity consumption and environmental concerns . Next-generation data centers demand enormous power, reigniting the debate about the sustainability of the current large-scale AI model and its impact on national energy grids.

GPU Market: From Revenue Paradise to Chronic Shortage

The graphics card sector has been the visible epicenter of this crisis . History has repeated itself with variations: first it was cryptocurrency miners, now it's AI projects absorbing the production of high-end GPUs. Manufacturers have discovered that it's much more profitable to sell large batches of accelerators to data centers than to dedicate capacity to consumer gaming models, and this shift in priorities has left home users behind.

The problem isn't just that many GPUs are being manufactured for AI , but that the production lines themselves have been redesigned to maximize that segment. As companies prioritize the professional and data center range, the availability of PC models is decreasing , and the few units that do reach stores come with inflated prices. The gaming market, once one of the driving forces of the graphics industry, is now overshadowed by AI.

This diversion of resources has a side effect: it lengthens the equipment upgrade cycle . Many users choose to stick with their current GPU or turn to the second-hand market, fueling a parallel economy where ex-mining or ex-data center cards with a history of intensive use abound. The buying experience suffers, and the perception that "upgrading your PC has become a luxury" spreads.

Price escalation and market reorientation in RAM

While GPUs are grabbing headlines, the real hemorrhage is happening in DRAM . Since late 2025, major manufacturers—Samsung Electronics, SK hynix, Micron Technology—have been increasingly shifting production capacity toward server DRAM, HBM memory, and high-performance AI solutions. The result: fewer chips available for PCs, mobile phones, smart TVs, routers, and virtually any device that relies on conventional RAM.

This shift has been amplified by the cyclical nature of the DRAM market . After several years of overproduction and falling prices, many manufacturers cut back on product lines and slowed investments. Just then, the AI ​​boom hit, catching the sector with less strength than needed. The adjustment was abrupt: supply was slow to react while demand exploded, leading to historic price increases and shortages of consumer memory modules.

The situation has been so intense that some iconic brands in the consumer market have fallen by the wayside. The case of Crucial, Micron's brand for end users , is a prime example: the company decided to stop producing RAM and SSDs under that brand from 2026 onwards to concentrate its efforts on more profitable segments. For the average consumer, this translates into fewer reliable options and a market dominated by expensive products or niche alternatives.

Meanwhile, the emergence of Chinese memory manufacturers like CXMT and YMTC is gaining traction. CXMT is developing very fast DDR5 modules—around DDR5-8000—while YMTC is focusing on high-density NAND Flash with technologies like Xtacking 4.0, capable of stacking many layers and achieving capacities of up to 8 TB in SSDs. Brands like Netac, Asgard, KingBank, and Gloway are already integrating these chips, although giants like Kingston and Corsair are still observing the trend with caution.

SSD and HDD Crisis: Storage Isn't Saved Either

The explosion of AI and cloud services isn't just devouring GPUs and DRAM; it has also set off alarm bells in the SSD and hard drive market. AI models need to be trained on massive amounts of data and require mixed storage environments: ultra-fast NVMe SSDs for hot processing and high-capacity nearline HDDs for historical datasets or less frequently accessed information.

After a period of NAND flash memory oversupply, manufacturers—again including Samsung, SK hynix, Micron, and others—cut production to stabilize prices. When AI began to drive demand for high-density enterprise SSDs, factories weren't ready to respond quickly . The result: stock shortages, especially for high-end drives shared between servers and high-end consumers, and further price increases.

In the HDD sector, companies like Western Digital and Seagate have found themselves in a similar situation. Large contracts with cloud and AI providers have absorbed a significant portion of the available capacity, to the point that some models have their stock already committed months before they even leave the factory. Although many users believe the entire market is shifting towards SSDs, hard drives remain irreplaceable in terms of cost per terabyte for cold mass storage.

For the end user, this translates into a familiar experience: less real variety in stores and a feeling that everything is more expensive . A 1 or 2 TB SSD that was affordable just a few years ago is now prohibitively expensive again depending on the model, and high-capacity HDDs for backups are no longer the bargain they once were.

Impact on the consumer: prohibitively expensive hardware and a shift to the second-hand market

All these changes in GPUs, RAM, and storage converge on the same point: the wallets of home users and small businesses . Building or upgrading a PC in 2026 has become an exercise in patience and restraint: exorbitant prices, lack of stock in key components, and a feeling that the consumer market has been relegated to the sidelines by AI and large data centers.

The pressure isn't limited to computers. The rising cost of DRAM and NAND flash memory is also impacting smartphones, routers, televisions, game consoles, and any device that requires memory or storage. Many consumer electronics manufacturers have had to adjust their product lines, raise prices on models, or cut specifications to avoid further price increases, which ultimately means fewer features for more money.

Given this situation, more and more users are turning to the secondhand market, refurbished products, or even DIY projects . We've seen initiatives as extreme as projects in Russia to manufacture handcrafted RAM modules due to scarcity and cost. At the same time, there's a growing trend towards alternative suppliers, especially Chinese ones, that promise chips geared towards the consumer sector neglected by the major traditional manufacturers.

This shift towards less established alternatives comes with its own risks: uncertainty about quality, limited warranties, and questionable compatibility . But when the official market becomes inaccessible, many users are willing to take on some exposure if it allows them to keep their equipment running or expand its capacity without going broke.

AI bubble and tensions in financial markets

While the hardware sector is experiencing its own turmoil, the financial side of the tech ecosystem is also in turmoil. Enthusiasm for AI has driven stock market indices to record highs, but beneath the surface, a brutal reshuffling is underway: traditional software and some established tech companies are facing growing investor distrust.

The catalyst has been the perception that many tasks performed by traditional software can be replaced by cheaper, more automated AI services. Legal model-based tools like Anthropic's Claude, productivity assistants, process automation, and intelligent data analytics platforms are prompting markets to question whether traditional software companies will be able to continue monetizing their products in the same way.

In recent months, a historic gap has opened up between the software sector and the S&P 500 , with a difference of more than 20 percentage points, a deviation comparable only to what was experienced with the bursting of the dot-com bubble in 2000-2001. Companies such as Oracle, ServiceNow, and AppLovin have seen their stock market value fall by tens of points, while others such as Palantir, Intuit, and Workday have also suffered sharp declines.

At the same time, there is a rotation of capital towards more “tangible” sectors such as energy, heavy industry, and basic materials, which have registered significant gains in the same period. Many investors prefer to take refuge in businesses linked to the traditional economic cycle while they wait for the dust to settle on AI and for it to become clear who truly generates sustainable value.

Parallels with the dot-com bubble and the current rise of AI

The situation is strikingly reminiscent of the dot-com bubble of the late 90s and early 2000s. Back then, the internet was the technology poised to change everything; today, AI takes center stage. In both cases, there has been a cocktail of unrealistic expectations, inflated projects lacking a solid business model, and a flood of investment funds eager to get in on the next big thing.

Just as companies like Lycos, Terra, and Boo.com emerged then and disappeared as quickly as they were inflated, we now see a proliferation of AI startups promising revolutionary solutions without demonstrating a clear path to profitability. The engine of stock market growth in major economies is once again concentrated in a single sector, and much of the capital is driven more by fear of missing out than by fundamental analysis.

However, there are important differences. Today's AI has real and profitable applications already in place : cloud services, process automation, specialized semiconductors, AI-integrated software in enterprise suites, and more. Furthermore, the technological and financial ecosystem is more mature, with more sophisticated risk analysis tools and a far more advanced digital infrastructure than in 2000.

Also significant is the influence of major figures like Elon Musk, Mark Zuckerberg, and Jeff Bezos , who not only lead key AI projects but are also closely linked to economic and political power. For many analysts, their capacity to influence makes a total collapse like the dot-com bubble less likely, although it doesn't prevent a "clean-up" of unsound companies—something that, on the other hand, is considered necessary to stabilize the market.

The Y2000K bug and the next big risk: the 2038 bug

Time-related crises in computing are not just a historical anecdote. The famous Y2K bug became the symbol of the first major global technological scare: it was feared that on January 1, 2000, systems that stored the year with two digits would interpret "00" as 1900, causing failures in banks, airlines, electrical grids, and all kinds of critical infrastructure.

In practice, the impact was minimal thanks to massive investments in prevention , estimated at over €200.000 billion globally. Governments and businesses reviewed, patched, and updated millions of lines of code and systems, averting the predicted chaos. In hindsight, many questioned whether the threat had been exaggerated, but experts agree that the risk was real and that preventative measures were what kept the world running on New Year's Eve.

In response, some systems resorted to temporary "patches," such as changing the reference year to 2020 to buy time, assuming that by then modern solutions would have been migrated. This workaround has taken its toll: on January 1, 2020, parking meters in New York stopped accepting payments , and some video games, like WWE 2K20, began to crash if the console's date wasn't changed to 2019.

A new challenge is already looming on the horizon: the Year 2038 problem . It affects systems that use 32-bit time counters—very common in Unix-based technologies and many embedded devices—which count the seconds from January 1, 1970. When that counter reaches its maximum value, on January 19, 2038, time will "turn around," and many machines will interpret the date as December 1901, which could trigger cascading failures.

The most vulnerable sectors are those that depend on critical infrastructure: banking and payments, telecommunications, air transport, energy, and industrial networks . Addressing this involves much more than simply updating hardware: it requires reviewing algorithms, operating systems, firmware, and applications that work with times and dates, relying on mathematical concepts such as modular arithmetic and numerical representation.

Global blackouts and the risk of relying on a single provider

Beyond date-related bugs, one of the most illustrative recent episodes of systemic risk has been the massive computer blackout caused by a faulty security update . A simple patch from a third-party cybersecurity provider was enough to disable millions of Windows computers and paralyze a substantial part of the global digital economy for hours.

Although the incident affected "only" a small percentage of the total number of devices, its impact was monumental because Microsoft and its platforms are at the heart of much of the global infrastructure . Airports were overwhelmed, financial services were disrupted, and companies were unable to access their systems. The episode made it clear that technical robustness and investment in security are not enough when everyone depends on the same link in the chain.

Security experts and academics have pointed out that these kinds of events highlight the enormous risk of putting all your eggs in one basket . Basing critical operations on a single vendor or ecosystem makes a domino effect almost inevitable when one fails. For many small businesses, diversification is difficult due to cost and complexity, but the current level of concentration is reaching alarming levels.

Some advocate for fostering an ecosystem with more medium and small-sized providers , so that the failure of one doesn't bring down half the planet. However, this diversification also opens up new attack surfaces and multiple potential vulnerabilities, complicating defense. Finding the balance between resilience and complexity becomes one of the major strategic challenges of the next decade.

The major IT challenges for businesses in the short term

This entire context of crisis, bubbles, and dependencies directly impacts IT teams within organizations , who operate in a kind of permanent state of emergency. Beyond purely technical issues, they must manage a growing list of challenges that affect the business as a whole.

First, cybersecurity is becoming increasingly complex : AI-powered phishing, deepfakes, endpoint attacks, sophisticated ransomware… The attack surface expands with every new service, IoT device, or cloud integration. Security can no longer be viewed as a static configuration, but rather as an ongoing practice that requires segmenting networks, testing systems, training staff, and reacting as quickly as the attackers.

Added to this is the complexity of regulatory compliance . Frameworks such as GDPR, HIPAA, the NIST cybersecurity framework, and new AI regulations impose requirements that go far beyond mere paperwork: governance structures, constant monitoring, and collaboration between IT, legal, compliance, and business are essential to avoid falling behind or incurring penalties.

Another sensitive issue is cloud costs . Cloud adoption is almost universal, but controlling spending has become a major headache. AI workloads, containers, multi-cloud architectures, and unsupervised proof-of-concept projects can cause bills to skyrocket unpredictably. Without good governance, the cloud becomes a source of technical debt, wasteful spending, and traumatic budget cuts.

Furthermore, companies are undergoing constant digital transformation . There are no longer one-off modernization projects, but rather an endless succession of migrations, integrations, and platform changes. The risk is ending up with an ecosystem full of isolated tools, technical debt, and processes that don't deliver real value. IT leaders must establish a clear roadmap, aligned with business objectives and capable of scaling.

Data, legacy systems, and talent: the human bottleneck

Managing corporate data has become another major headache. The volume of information generated by organizations—system logs, user activity, customer data, IoT sensors—is constantly growing. It's not just about storing it, but also about classifying it, protecting it, complying with regulations, and extracting value from it. When this is done poorly, compliance problems, inefficiencies, and decisions based on erroneous information arise.

All of this becomes more complicated when dealing with critical legacy systems : old applications, obsolete hardware, and highly customized platforms that can't be shut down overnight because they support essential processes. Modernizing them without breaking anything requires careful planning, phased migrations, and often investments that clash with already strained budgets.

At the same time, the skills gap is widening . Technology is advancing faster than teams can keep up with training, and the job market can't keep up with demand for professionals in cloud architecture, advanced cybersecurity, or applied AI. Even when top talent is hired, retaining it is difficult in an environment where offers with better salaries, remote work, and opportunities for advancement are constantly appearing.

The consequence is that many IT departments are forced to do more with less, and often with outdated knowledge . To break this vicious cycle, it's necessary to combine external recruitment strategies, robust internal upskilling programs, and strong management support to allocate time and resources to continuous professional development.

As a backdrop, internal communication remains an Achilles' heel. Without coordination between IT and other departments , without clear documentation, and without knowledge sharing, projects end up being delayed, requirements are misinterpreted, and frustration skyrockets. In distributed or hybrid teams, this risk is multiplied.

Supplier risk, remote work, and preparing for an uncertain future

Another aspect that has gained prominence is the risk associated with external providers . Almost all companies now depend on third parties for software, cloud services, support, or business process outsourcing. Each of these partners introduces potential vulnerabilities: lax security practices, service interruptions, excessive contractual dependence, or regulatory non-compliance that ultimately affects the customer.

Managing this risk goes far beyond signing contracts: it requires ongoing assessments, clear security and compliance agreements , and close cooperation between IT, Procurement, Legal, and Security. The outage linked to a faulty patch has been a stark reminder of what happens when one of these links fails.

The rise of remote and hybrid work adds another layer of complexity. Ensuring access to corporate systems from home networks, managing endpoints scattered across the globe, and maintaining seamless collaboration across time zones requires well-thought-out policies, appropriate tools, and constant vigilance for potential security vulnerabilities.

Ultimately, technology leaders face the paradox of long-term planning in an environment that changes every few months . AI is accelerating, quantum computing is on the horizon, regulations are evolving, and supply chain crises appear and disappear. Rather than finding the perfect solution, the goal is to design flexible architectures, cultivate multidisciplinary teams, and frequently review strategy to adapt to whatever comes next.

What all these crises show—from the Y2000K bug to the AI ​​bubble, including the 2038 bug, global supplier failures, and the strain on GPUs, RAM, and storage—is that fragility and interdependence are now structural features of our digital world . The organizations that will emerge strongest will not necessarily be those with the most cutting-edge technology, but rather those that combine risk diversification, a culture of prevention, investment in people, and the ability to react quickly when, inevitably, something goes wrong again.