AI’s power-and-chip boom is rewriting markets—are we heading for the next bubble?

Across late June 2026, multiple outlets converge on a single market question: is the AI boom becoming a self-fulfilling asset bubble, or the start of a durable new economic order. Bloomberg’s Ed Cole (Man Group) argues it is still too early to confirm bubble dynamics, but he flags that the traditional 60/40 portfolio framework may be outdated as inflation, reshoring, and globalization backlash reshape return drivers. The Financial Times frames a “new AI-based world order” where one factor increasingly dictates global returns, while BIS-style warnings (via InvestingLive) compare the buildout to dot-com and railway manias, emphasizing bust and recession risks. Separately, the Financial Times reports a record $200bn M&A surge in the US power sector as companies race to build energy infrastructure for data centers, and Guardian coverage highlights how chipmakers underpinning AI have propelled Asia-Pacific stock markets higher in the first half of 2026. Geopolitically, the cluster points to a shift from “AI as software” to “AI as strategic infrastructure,” where electricity, compute supply chains, and capital allocation become instruments of national competitiveness. The US power-sector deal wave suggests governments and regulators may increasingly treat grid capacity and generation buildout as strategic assets, tightening the link between industrial policy and financial markets. Meta’s AI research chief Dawn Song, cited by SCMP, pushes the frontier toward “AI agents” that deliver economically valuable work across real-world domains, which raises the stakes for labor markets, productivity policy, and regulatory oversight. Meanwhile, the broader “backlash against globalization” and reshore efforts discussed by Man Group imply that industrial geography—where chips, servers, and energy are produced—will matter more than in prior cycles, benefiting regions with resilient supply chains and penalizing those exposed to bottlenecks. Market implications are already visible in equities and deal activity. The US $200bn M&A boom in power infrastructure implies elevated demand for grid equipment, transformers, transmission services, and construction capacity, with second-order effects for industrial metals and energy-related capex. Chipmakers tied to AI hardware have seen valuations surge—Guardian notes some have tripled or more—supporting a risk-on tilt in Asia-Pacific markets and likely boosting semiconductor-related indices and exchange-traded exposure. If BIS’s “bust risk” framing gains traction, the impact could spill into rates and credit: investors may reprice duration, widen risk premia, and rotate away from traditional balanced allocations, particularly if inflation proves sticky. Currency effects are not explicitly quantified in the articles, but the direction of travel is clear: AI-linked equities and infrastructure beta are being bid, while portfolio diversification assumptions are under pressure. What to watch next is whether the AI buildout transitions from speculative momentum to measurable cash-flow conversion without triggering a macro downturn. Key indicators include power-sector capex announcements and permitting timelines for data-center energy buildouts, semiconductor order visibility, and whether chip valuation multiples compress as earnings catch up. On the policy side, monitor how regulators respond to AI agents’ expansion into economically valuable real-world tasks, and whether labor-market disruption leads to new training, wage, or automation rules. For bubble-risk confirmation, look for credit stress signals, rising defaults in leveraged infrastructure financings, and evidence that inflation and reshoring dynamics keep real yields elevated enough to challenge long-duration growth assets. The escalation trigger is a simultaneous slowdown in AI-related demand plus tightening financing conditions; de-escalation would be faster-than-expected earnings delivery and stable power-grid delivery that reduces bottleneck fears.

Geopolitical Implications

• 01

  The strategic center of gravity for AI is shifting toward electricity and compute supply chains, increasing the geopolitical weight of grid resilience and industrial policy.

• 02

  Reshoring and globalization backlash may harden industrial geography, rewarding countries and regions with secure semiconductor and energy inputs while penalizing bottlenecked supply chains.

• 03

  AI agents expanding into real-world economic domains will likely intensify regulatory competition and cross-border standards debates, affecting market access and compliance costs.

Key Signals

• —US power-sector permitting and grid-connection timelines for data centers
• —Semiconductor order books and earnings guidance versus valuation multiples
• —Credit spreads and default risk in leveraged infrastructure and AI-adjacent financings
• —Regulatory moves on AI agents’ autonomy and labor-market transition policies
• —Inflation trajectory and real-yield direction affecting long-duration growth assets