A Moment of Excitement and Unease
Public excitement around artificial intelligence has grown rapidly. New models appear every few months. Companies feel pressure to declare themselves “AI-ready”. Investors chase deals at historic speed. At the same time, conversations continue in policy circles, boardrooms and research institutes: are we living through an AI bubble? The short answer is that parts of the sector show clear signs of excess. But that framing alone misses something deeper. Beyond technology, AI is becoming an economic and geopolitical force. That makes this moment different from earlier periods of hype.
The Scale of the AI Wave: What the Numbers Tell Us
UNCTAD estimates the global AI market at US $189 billion in 2023, rising to US $4.8 trillion by 2033, a pace matched by very few technologies in history.
McKinsey’s 2025 State of AI report notes that 88% of organisations now use AI in at least one function, yet only one-third have scaled it meaningfully across the enterprise.
Infrastructure is also expanding rapidly. A McKinsey compute analysis projects that global data centre investment needed for AI could reach US $6.7 trillion by 2030.
These numbers show strong momentum but uneven maturity.
Why Many Believe an AI Bubble Is Forming
Observers argue that investment has grown faster than revenues. In 2025, global venture funding for AI start-ups crossed US $89.4 billion, nearly one-third of all VC funding. (SecondTalent)
Nearly 46% of all global VC funding in Q3 2025 went into AI, and a handful of companies closed US $500 million+ rounds. (Crunchbase)
Product differentiation remains weak—many firms build similar assistants, agents and automation tools.
Infrastructure remains expensive, with companies signing large compute contracts before stable business models emerge.
This combination of exuberance, repetition and high cost resembles the early dot-com years: real innovation mixed with fragile foundations.
Why This Cycle Is Fundamentally Different
Two structural forces distinguish today’s moment.
First, governments are directly involved.
The U.S. CHIPS Act, the EU AI Act, China’s national AI strategies and India’s digital public infrastructure model show that states are steering AI development. They subsidise computing, shape regulation and invest in research. This was not the case during earlier tech booms.
Second, AI is already delivering real value.
In logistics, manufacturing, finance and pharma, AI systems reduce error rates, speed up research and automate repetitive tasks.
In India, 47% of enterprises now have multiple AI use-cases in production, according to EY–CII.
These gains are early but concrete, showing that AI is not merely speculative.
Capital Flows: Strategic Money Meets Speculative Money
By mid-2025, AI start-ups captured 51% of all global venture funding, indicating a historic reallocation of capital toward AI. (ET Enterprise AI)
But not all money is the same.
Strategic capital includes sovereign funds, public-sector AI programmes, Big Tech’s data-centre expansion and national investments in semiconductor supply chains.
Speculative capital includes high-valuation start-ups with limited revenue and “hype rounds” that prioritise speed over substance.
Strategic capital builds lasting infrastructure; speculative flows swing with market mood. Distinguishing the two gives a clearer picture of AI’s real momentum.
AI as Geopolitics: Technology, Power and Influence
AI is now an instrument of geopolitical leverage.
Control over chip fabrication, cloud infrastructure and research talent determines which nations can train and deploy frontier models.
The U.S.–China rivalry over advanced chips, Europe’s regulatory role and the rise of middle powers like India, Singapore and the UAE shape the global AI landscape.
AI is now linked to national competitiveness, economic security and diplomatic positioning.
Speculation and Excess — But on Top of Real Foundations
Speculation exists and cannot be ignored. Some start-ups add “AI-powered” labels without meaningful capability. GPU hoarding has pushed up hardware costs. Some valuations rely more on AGI timelines than present performance.
But speculation does not erase underlying progress. Bubbles may burst, but infrastructure, talent and institutional capability usually remain.
The real risk is not AI itself, but the misallocation of capital within a rapidly maturing market.
The Indicators That Actually Matter Now
To understand where AI is truly heading, it helps to track five signals rather than funding headlines.
Enterprise scaling
Meaningful change occurs when companies move beyond pilots and rebuild workflows using AI. McKinsey notes that only a minority of global organisations are scaling, even though adoption is widespread.
Compute economics
GPU costs are a major barrier. Prices now range from US $1.25 to US $10.50 per GPU-hour across providers, with reserved H100 clusters offered at around US $2.40 per hour. (GetDeploying, Lambda)
If costs fall, AI becomes more democratic. If they stay high, power consolidates.
Regulatory clarity
The EU AI Act (enforced 2024, fully applicable by 2026) and India’s 2025 AI governance guidelines reduce uncertainty.
Clear rules encourage real adoption and temper speculative behaviour. (EU Digital Strategy; MeitY Guidelines)
Vertical AI models
Domain-specific AI is beginning to outperform general models. A 2025 study estimates vertical AI could generate US $344 billion in value across major industries. (SymphonyAI)
This marks a shift from novelty to productivity.
Stable monetisation
Some leading AI firms generate over US $1.1 million in annual recurring revenue per employee, far above SaaS benchmarks. (BVP State of AI 2025)
Steady pricing models and recurring revenue are strong signs of real industry maturity.
These indicators reveal long-term strength better than market sentiment ever could. When all five move in the right direction, the sector is on solid footing regardless of valuation cycles.
A Measured Conclusion: Beyond the Bubble Frame
Some start-ups will fail. Many valuations will correct. But the underlying forces such as state involvement, infrastructure investment, enterprise value and geopolitical competition will not disappear.
The more important question is not whether AI is in a bubble, but whether institutions are prepared for the structural transition already underway.
AI has shifted from being a frontier curiosity to becoming an integral part of the global economic backbone. And that shift will outlast the inevitable noise of the market.
