Warning about the hype: Why the AI ​​revolution is still stalling in many companies

Jobs in transition: Why AI won't replace us, but will divide the job market

The silent redistribution: Who will benefit from the AI ​​miracle in 2026 – and who will lose out?

Artificial intelligence has definitively left the experimental lab. By 2026, we are no longer talking about fascinating gimmicks or isolated pilot projects, but about a far-reaching foundational technology that is reshaping the global economy. As tech giants pump trillions into new data centers and infrastructure in an unprecedented supercycle, traditional companies face the mammoth task of deeply integrating smart agents into their business processes. But the rapid rise of intelligent machines comes at a price: power grids are reaching their limits, the labor market is polarizing, and the distribution of value creation is being fundamentally renegotiated. This is a detailed look at a crucial transitional year that will determine the winners and losers of AI capitalism – and reveals why the greatest challenge ultimately lies not in the technology, but in human adaptability.

When machines think for themselves: How artificial intelligence is recasting the economic foundation

Billion-dollar bets, power shortages and the silent redistribution of added value – why 2026 will be the litmus test of AI capitalism

Artificial intelligence is not a passing technological fad, but a fundamental, foundational technology—in economic terms, a general-purpose technology, comparable to the steam engine, electricity, or the internet. Such technologies permeate virtually all economic sectors, transforming production processes, organizational structures, and even the nature of work itself. The World Bank explicitly categorizes AI in its 2026 World Development Report, describing it as a technology with the potential to rewrite the development paths of entire economies. By 2026, this permeation will have reached a scale that extends far beyond experimental pilot projects. Companies will be investing heavily to embed AI not merely as a tool for increasing efficiency, but as a core element of their operational and strategic architecture.

The current phase is characterized by a transition to agentic systems that no longer merely respond to requests but independently plan, execute, and orchestrate across multiple systems. Gartner expects that by the end of 2026, around 60 percent of AI adoptions in companies will include agentic capabilities. This distinguishes today's AI from previous waves of automation, which primarily affected physical or routine tasks. Instead, it is now deeply impacting knowledge-intensive areas, from software development and financial decision-making to supply chain management. The economic assessment must therefore consider both the enormous productivity gains and the associated distributional effects. While some economists speak of a potential acceleration of global growth, others warn of a concentration of profits among a few technology providers and a temporary increase in structural unemployment in certain occupational groups.

Historically, with earlier foundational technologies, it often took decades for the full economic effects to become measurable. The electrification of factories required not only new machinery but a complete redesign of factory layouts and work organization. The situation is similar with AI today. While many organizations have adopted AI tools, deep integration into business processes, adaptation of organizational structures, and the necessary workforce training are lagging behind. This explains why, despite billions in investment, the macroeconomically measurable productivity gains have so far been more moderate than task-based laboratory studies would suggest. Nevertheless, early indicators point to an acceleration effect, particularly in sectors with high data availability and digital maturity.

The capital hunger of a trillion-dollar machine

The AI ​​market is growing at a breathtaking pace. Market research firm Gartner estimates global AI-related spending for 2026 at around $2.52 trillion, representing an increase of approximately 44 percent compared to the previous year. By far the largest single item is infrastructure, meaning specialized servers, processors, networks, and data centers. According to Gartner, the building of AI foundations alone will drive spending on AI-optimized servers up by 49 percent in 2026, with this area accounting for roughly 17 percent of total AI spending. Major hyperscalers such as Microsoft, Amazon, Google, Meta, and Oracle are planning aggregate capital expenditures of around $600 to $690 billion for 2026, an increase of approximately 36 percent compared to the previous year, with roughly three-quarters of that going toward AI infrastructure.

This scale will only become tangible over several years. McKinsey predicts that nearly seven trillion US dollars will need to be invested in data centers worldwide by 2030 to meet the demand for computing power, of which around 5.2 trillion will be for AI-capable capacity alone. The Swiss investment bank UBS anticipates that hyperscalers will spend around 4.1 trillion US dollars on AI infrastructure between 2026 and 2028, almost tripling the 1.3 trillion invested in the preceding six years. A revealing indicator of this overzealousness is the observation that cloud companies are now reinvesting virtually every dollar of cloud revenue directly into AI infrastructure.

This supercycle is driving demand for semiconductors, especially graphics processing units (GPUs) and specialized AI chips, but also for energy, cooling technology, and high-performance networks. The effect extends deep into the supply chain: It is estimated that by 2026, up to 70 percent of the world's memory chips will be consumed by AI data centers, forcing the three major manufacturers, Samsung, SK Hynix, and Micron, to redirect scarce capacity to high-margin memory chips. At the same time, new financing instruments are emerging, and a growing share of investments is being financed through debt, significantly increasing indebtedness in the technology sector and raising new systemic questions.

Despite this boom, the distribution of value creation remains unequal. A significant portion of the profits is concentrated among the providers of the underlying technology, while many traditional companies are, for the time being, primarily bearing the costs of infrastructure and integration. Expectations in the capital markets are correspondingly high. Goldman Sachs observes that rising investment forecasts are increasingly accompanied by growing investor selectivity, with investors scrutinizing more closely which companies will actually benefit. This shift from pure infrastructure enthusiasm to demonstrable value creation at the application layer marks a stage of market maturation.

From code to container: AI is transforming entire industries

In software development, AI is fundamentally changing the way we work, and its adoption is now almost ubiquitous. Current industry analyses show that around 97 percent of organizations are using or evaluating AI in software development, with the greatest benefit in pure code generation, while planning, design, and maintenance still require human judgment. This doesn't reduce the need for skilled developers, but rather shifts it. Routine tasks are automated, while complex system architecture, the critical evaluation of AI-generated code, security aspects, and creative problem-solving are gaining in importance. Significantly, in teams with high AI usage, the time spent on code reviews increased by 91 percent because more generated code requires more intensive review, and 66 percent of developers cite the phenomenon of being almost, but not quite, right as their biggest frustration. AI thus accelerates code writing but shifts the bottleneck to quality assurance.

In finance, AI enables faster, more widely supported, and more transparent decisions. Algorithms analyze massive datasets in near real-time, identify patterns that elude human analysts, and simulate scenarios at high speed. This democratizes certain analytical tools but also carries the risk of systemic errors if models are based on similar training data or fail to adequately represent rare extreme events. In industry and logistics, AI significantly accelerates supply chain optimization, but many organizations remain in the pilot phase because issues of integration into existing systems, data quality, and change management have not yet been resolved.

SAP exemplifies this transformation and the gap between announcement and availability. At its Sapphire conference in May 2026, the Walldorf-based company presented its vision of the Autonomous Enterprise, in which AI agents take over business processes end-to-end, while humans retain control over critical decisions. SAP positions Joule as the central orchestration and interaction layer and has announced more than 50 domain-specific assistants and over 200 specialized agents; the stated goal is to deploy around 400 agents by the end of 2026. At the same time, SAP is leveraging the AI ​​boost to accelerate the migration of legacy systems to the cloud, with agent-based tools expected to reduce effort by more than 35 percent. However, the reality of usage remains sobering: A significant portion of the announced agents are still in preview or early access status, Joule is tied to cloud contracts, and industry analyses estimate that only around three percent of SAP customers are using Joule in production. This discrepancy between strategic vision and operational penetration is symptomatic of the entire field of enterprise AI in 2026.

Managed AI Platform - Image: Xpert.Digital

Here you will learn how your company can implement customized AI solutions quickly, securely and without high entry barriers.

A managed AI platform is your all-inclusive, worry-free solution for artificial intelligence. Instead of dealing with complex technology, expensive infrastructure, and lengthy development processes, you receive a ready-made solution tailored to your needs from a specialized partner – often within just a few days.

The key advantages at a glance:

⚡ Rapid implementation: From idea to ready-to-use application in days, not months. We deliver practical solutions that create immediate added value.

🔒 Maximum data security: Your sensitive data stays with you. We guarantee secure and compliant processing without sharing data with third parties.

💸 No financial risk: You only pay for results. High upfront investments in hardware, software, or personnel are completely eliminated.

🎯 Focus on your core business: Concentrate on what you do best. We take care of the entire technical implementation, operation, and maintenance of your AI solution.

📈 Future-proof & scalable: Your AI grows with you. We ensure continuous optimization and scalability, and flexibly adapt the models to new requirements.

More information here:

• Managed AI Platform

Few giants, high volatility

The AI ​​landscape is dominated by a handful of hyperscalers and technology giants that control both the infrastructure and the leading models. Their massive investment expenditures, however, also create dependencies across the entire supply chain, from chip manufacturers and energy providers to cooling and networking specialists. This is reflected in the stock markets as high volatility. Periods of euphoric price increases in AI stocks are repeatedly interrupted by corrections triggered by warnings of overheated development, regulatory risks, or doubts about the profitability of the enormous investments. The central open question is whether future AI revenues will justify the record-high infrastructure spending.

For investors seeking broad exposure to AI without taking on individual stock risks, thematic funds and exchange-traded funds (ETFs) have become attractive. However, experts urge caution, as not every company with AI in its name will benefit in the long run, and the increasing selectivity of investors is already evident in the divergence of individual share prices. The real winners are likely to be those who not only implement AI but also use it to create sustainable competitive advantages, whether through superior customer experiences, significantly higher operational efficiency, or entirely new business models. Established software providers with deep industry knowledge and high-quality company data may be better positioned than pure AI startups because their models are grounded in real-world business contexts.

Work in transition: Supplementation instead of replacement, for now

The impact on the labor market is more nuanced than many popular narratives suggest. The International Monetary Fund estimates that around 40 percent of employment worldwide could be affected by AI, with the figure rising to about 60 percent in advanced economies, around 40 percent in emerging markets, and about 26 percent in low-income countries. Being affected does not automatically mean displacement. So far, AI has augmented human labor rather than completely replacing it, and in roughly half of the exposed jobs, it is likely to make workers more productive rather than replace them. At the same time, the Fund warns that AI is likely to exacerbate overall economic inequality in most scenarios unless policymakers intervene.

The real challenge, therefore, lies less in the threat of mass unemployment than in accelerated polarization. Highly skilled workers who can effectively utilize AI will see significant productivity gains, while routine tasks at mid-level qualification levels will come under pressure. Young professionals entering the workforce in high-pressure fields will face narrower hiring windows, while experienced specialists can leverage their expertise through AI. In software development, overall demand remains high because generated code must be intensively tested, adapted, and integrated into complex systems, as evidenced by the sharp increase in review effort. Retraining and lifelong learning will thus become key economic policy priorities that will determine the fair distribution of the benefits of AI.

Electricity, water, and the physical price of intelligence

The enormous energy demands of data centers represent one of the most tangible practical limitations of the AI ​​boom. The International Energy Agency estimates the global electricity consumption of data centers in recent years at several hundred terawatt-hours and, in its baseline scenario, projects a doubling to around 945 terawatt-hours by 2030, which would then correspond to almost three percent of global electricity consumption. Data centers, artificial intelligence, and electromobility are among the most dynamic drivers of the globally increasing electricity demand, which, according to the IEA, is expected to grow by an average of 3.6 percent annually between 2026 and 2030. In some regions, this is already leading to local grid bottlenecks, rising electricity prices, and conflicts with municipalities that oppose new data centers due to noise, water consumption for cooling, and the additional strain on infrastructure. In the US, according to an industry estimate, data centers could account for around 38 percent of the net increase in electricity consumption by 2037.

These environmental and infrastructural limitations are forcing the industry to innovate in energy-efficient chips, liquid cooling, and site selection in regions with abundant renewable energy. Power density per server rack is increasing dramatically and, according to industry experts, will almost double annually, with rack power moving from around 227 kilowatts toward 400 kilowatts and beyond, necessitating new power distribution concepts. At the same time, new markets are emerging for green data centers, intelligent energy management systems, and specialized service providers. It is noteworthy that, according to the IEA, renewable energies and nuclear power are projected to account for roughly half of global electricity generation by 2030, while emerging economies, particularly China, will account for the majority of the additional consumption. Sustainability is thus becoming not only a risk but also a differentiating factor.

Law, power and morality: the regulatory framework

The European Union has established a risk-based regulatory framework with the AI ​​Act, which serves as a global benchmark. The so-called Digital Omnibus, published as Regulation 2026/1744 on July 24, 2026, and entering into force on July 27, 2026, specifically adjusted the timeline. The obligations for standalone high-risk systems listed in Annex III, which cover areas such as employment, education, law enforcement, creditworthiness, and biometrics, were postponed from August 2026 to December 2, 2027, while high-risk systems embedded in products, as listed in Annex I, were given until August 2, 2028. For high-risk systems used by public authorities, the deadline is even August 2, 2030. At the same time, August 2, 2026 remains an active date, as the transparency obligations under Article 50, such as the labeling of chatbots and deepfakes, largely apply according to the original timetable, although an extended deadline of December 2, 2026, has been granted for watermarking generative systems already on the market.

The omnibus thus shifts the requirements without abandoning the goal of making AI safe, transparent, and human-centered. Newly introduced are additional prohibitions against AI systems that generate non-consensual intimate depictions and abusive material, effective from December 2, 2026. Substantial fines of up to seven percent of global annual revenue continue to create significant pressure for compliance. Geopolitically, AI is intensifying competition between the US, China, and Europe, with export controls on high-performance chips, the establishment of proprietary supply chains, and the battle for talent shaping the landscape. In China, the demand for computing power is accelerating a separate supercycle in high-performance materials, which argues for a cross-border, diversified investment strategy. Ethical questions surrounding model bias, the protection of personal data, and deployment in sensitive areas remain unresolved, and companies that establish robust governance structures early on gain not only legal certainty but also the trust of customers and investors.

Not just a game of the big players

Despite the dominance of large technology companies, significant opportunities are opening up for small and medium-sized enterprises (SMEs) and emerging economies. SMEs can suddenly leverage accessible AI tools to access analytical capabilities previously reserved for large corporations. In developing and emerging countries, AI can help close productivity gaps, provided that local data ecosystems and digital skills are successfully established. The World Bank, in its 2026 Global Development Report, emphasizes that governments must create both analog and digital foundations and that AI applications should be designed to reach people with basic mobile phones and those lacking literacy or purchasing power, for example, through voice-based services. This brings to the forefront the question of whether AI reduces or exacerbates global inequalities.

New business models are emerging around AI as a service, agent-based platforms, data-driven ecosystems, and hybrid human-machine collaboration. Companies that not only implement AI but also reinvent their core processes around the technology are likely to make the biggest leaps forward. This requires the courage to cannibalize existing models and a culture of continuous experimentation. At the same time, the example of enterprise AI shows that there is a long road between purchasing a tool and actually creating value, a road that requires data quality, process restructuring, and cultural change.

What executives and investors should do now

Leaders must understand AI as a core strategic competency, not an isolated technology project. This means establishing clear AI governance, ensuring data quality and availability, developing talent with hybrid skills, and strategically shaping partnerships with technology providers. Experience from 2026 shows that the bottleneck is rarely the technology itself, but rather the organizational ability to integrate it into existing processes, as demonstrated by the low productive adoption rate of even mature agent platforms. Those who master integration into core and data processes transform lab productivity into measurable business success.

Investors should look beyond short-term waves of enthusiasm and focus on fundamental value creation, sustainable scalability, and responsible development. Given the increasing selectivity of markets and the open question of the profitability of record investments, diversification across the entire value chain seems advisable, from infrastructure and models to vertical applications. The concentration of profits among a few providers of the core technology is real, but the next wave of value creation is likely to emerge in the application layer, where industry knowledge and data access will determine success.

Between euphoria and disillusionment: the most likely path

The future of AI in business will be neither a pure productivity miracle nor a dystopian collapse. A realistic path seems to be one in which AI noticeably accelerates global growth, boosts productivity in many sectors, and creates new forms of value creation, while simultaneously entailing significant adaptation costs in the form of retraining, inequality, and regulatory fine-tuning. The decisive factor will be the speed and quality of societal and organizational adaptation, not solely the technical capabilities of the models. The discrepancy between the rapid pace of investment and the still-moderate productive penetration is not a contradiction, but rather the typical pattern of a basic technology in its maturation phase.

Companies and governments that invest early in skills, infrastructure, and responsible governance will benefit disproportionately, while others risk falling behind. The AI ​​revolution is therefore not inevitable, but a challenge to be shaped. It demands a high degree of foresight, adaptability, and ethical responsibility from all stakeholders. Only in this way can the technological potential be transformed into broadly shared economic and social progress. The developments of 2026, from the trillions in investments and the postponed regulatory deadlines to the physical limits of power grids, indicate that the global economy is in the midst of a crucial transition phase, the outcome of which remains uncertain.

Konrad Wolfenstein

I would be happy to serve as your personal advisor.

You can contact me at wolfenstein∂xpert.digital or

Just call me on +49 7348 4088 965 .

LinkedIn