Why the AI Bubble is Overhyped: A Contrarian View from Viktor Shvets

The notion that the current AI revolution is a speculative bubble overlooks its structural advantages and tangible returns in key sectors. While critics warn of overvaluation and excess, Viktor Shvets, Head of Global Desk Strategy at Macquarie Capital, argues that AI's disruptive potential—and its already proven ROI in healthcare, finance, and manufacturing—justifies strategic optimism. Here's why the “bubble” narrative is overblown, and where investors should focus.

Historical Precedents: Bubbles as Catalysts, Not Cul-de-Sacs

History shows that technological revolutions require asset bubbles to accelerate innovation. Carlota Perez's framework, cited by Shvets, explains that each wave—from the steam engine to the internet—involved overcapitalization and debt accumulation, only to later enable transformative productivity gains. The dot-com bubble of the late 1990s, for instance, laid the groundwork for the internet's eventual $5 trillion+ economy.

Shvets applies this logic to AI: today's $500 billion in cumulative capex and R&D by U.S. hyper-scalers (up 85% since 2021) are not wasted investments but a necessary step to build foundational infrastructure. Unlike past bubbles, AI's ROI is materializing faster. Startups like OpenAI and Cursor are reaching $200 million in annual revenue in under two years—a pace unmatched in previous cycles.

The Myth of Overvaluation: Sectors Delivering Tangible Gains

Critics cite inflated valuations and speculative hype, but Shvets points to sectors where AI is already driving measurable returns:

1. Healthcare: AI diagnostics reduce errors and costs. Companies like IBM Watson Health and DeepMind (Alphabet) are improving outcomes in radiology and drug discovery.

2. Finance: AI algorithms optimize trading, risk management, and customer service. Firms like Vanguard and BlackRock are integrating AI to outperform passive benchmarks.
   

3. Manufacturing: AI-powered robotics and predictive maintenance cut downtime. Companies like General Electric and Siemens report 20-30% efficiency gains in supply chains.

Structural Advantages: Speed, Scalability, and Societal Adaptation

Three factors mitigate bubble risks:
- Accelerated Innovation Cycles: AI's ROI is measured in quarters, not years, reducing capital strain.
- Low Marginal Costs: Unlike physical infrastructure, AI scales effortlessly. A chatbot costs pennies per user, enabling “good enough” solutions to disrupt industries.
- Societal Adaptation: While job displacement is real, humans are pivoting to roles emphasizing creativity and empathy—sectors AI cannot easily replicate. McKinsey estimates AI could lift global productivity to 5% by the late 2030s, reversing decades of stagnation.

Investment Implications: Focus on Spillover Sectors and Agile Firms

The contrarian case isn't to ignore risks but to allocate capital where AI's spillover effects are strongest:
- Healthcare: Invest in AI-driven drug discovery (e.g., BenevolentAI) and telemedicine platforms.
- Defense: AI is revolutionizing logistics and decision-making. Track firms like Raytheon Technologies, which uses AI for predictive maintenance in defense systems.
- Undervalued Startups: Prioritize agile firms leveraging AI's scalability, such as Cohere (natural language processing) or DeepMind (healthcare algorithms).

Avoid clinging to legacy tech giants like Microsoft or Amazon unless they prove agile in AI integration.

Conclusion: The AI Revolution is Here—Don't Miss the Real Opportunity

The fear of an AI bubble misses the point: this is a paradigm shift akin to the Industrial Revolution, not a fleeting fad. While overcapitalization exists, the speed of ROI and AI's structural advantages—paired with abundant global capital—mitigate prolonged market strain.

Investors should focus on sectors where AI's impact is tangible and undervalued, not follow the herd. The “bubble” narrative will fade as productivity gains materialize. As Shvets concludes, “The risk isn't over-investment in AI, but under-investment in a world where it's the only path to relevance.”

This article reflects a contrarian perspective informed by historical precedents and current data. Past performance does not guarantee future results.