The AI Investment Cycle: Platforms, Infrastructure, and Markets | Global Conference 2026

Global AI Investment Scale and Trajectory

• Current annual AI capital expenditure (CapEx) is approximately $1 trillion, combining investments from the "big four" hyperscalers and emerging "neoclouds."
• Projected total CapEx over the next five years is estimated at an additional $7 to $8 trillion.
• AI contribution represents roughly 33% of a 3% annual global GDP growth rate (adding ~$3 trillion/year), excluding downstream service value.
• The investment landscape is characterized by roughly ten concurrent "Manhattan projects" (five in the US, five in Asia/China) with no expected abatement in the next four to five years.
• Market cap restructuring has seen $8 trillion in incremental value added to the compute and intelligence layers YTD, while the services/application layer has lost $1–2 trillion.

The Three-Layer AI Stack Valuation

• Compute Layer (Bottom): The physical layer (chips, data centers, power, cloud) is the primary driver of current market cap, with nine of the ten largest global companies building their own chips.
• Intelligence Layer (Middle): Three foundation labs (SpaceX, Anthropic, OpenAI) are individually valued at $1 trillion or more.
• Services/Application Layer (Top): This sector, representing two-thirds of global GDP, remains unproven ("jury is out"), leading to a market "rewiring" away from traditional SaaS models.

Evolution of Software and Business Models

• The SaaS business model is declared "dead" as infinite, low-cost code generation collapses the price of software; however, total software volume will exceed historical norms due to non-stop agent-driven code creation.
• Competitive moats based on traditional software duration are being questioned; future value will derive from industries applying AI rather than software vendors themselves.
• Historical software developer growth has been 30x in 25 years (1 million to 30 million); AI is projected to cause an exponential increase in code output over the next 20 years.
• Infosys emphasizes a shift from experimentation to "production-grade" adoption, focusing on 10–15 high-impact business outcomes rather than hundreds of proofs of concept (POCs).

Financial Infrastructure and the Machine Economy

• A critical market gap exists: current financial rails are human-centric and lack support for machine-to-machine (M2M) micropayments required by AI agents.
• Blockchain and decentralized networks are identified as the necessary infrastructure to enable native digital payments for AI agents, potentially utilizing the HTTP 402 (payment required) protocol standard.
• Solana Foundation President Lily Liu highlights the risk of centralized AI control, advocating for open-source, decentralized models to distribute intelligence across the global population rather than concentrating it among two or four governments.

Market Structure, IPOs, and Valuation Dynamics

• BlackRock anticipates three new trillion-dollar companies (OpenAI, Anthropic, and potentially others) entering the public market, intensifying market concentration via "power laws."
• There is a backlog of roughly 1,000 pre-AI era unicorns that are expected to never achieve an exit due to the new economic reality.
• Upcoming AI IPOs are expected to exert significant liquidity pressure and drive fast-track inclusion in major indices, favoring passive "pro-growth" strategies.
• Citadel Securities data indicates that 45% of current US equity allocation is in AI-related assets, with the "Magnificent Seven" comprising 35% of S&P 500 market cap.

Sovereign AI and Geopolitical Considerations

• Foundation model development is bifurcated, with the US and China acting as the only two nations capable of building "sovereign intelligence."
• However, the physical "compute layer" (data centers) remains a sovereign decision point for all nations, with potential hubs identified in Malaysia, the Philippines, Spain, and the Middle East.
• Cybersecurity and data privacy are described as boundary-agnostic threats, requiring a security layer that operates independently of geopolitical borders.

Human Capital and Cognitive Impacts

• Infosys reports that entry-level jobs will not be eliminated but will evolve; the company continues to hire entry-level talent to ensure a "human-in-the-loop" for domain understanding.
• Experts warn of potential "cognitive atrophy" where reliance on AI (e.g., full self-driving cars) degrades human muscle memory and critical thinking skills.
• Countermeasures include conscious disengagement strategies and maintaining human agency in critical decision loops.

Future Outlook and Buzzwords

• Scott Rubner (Citadel): Predicts "Equity Technicals" and the rise of automated, 24/7 trading strategies as the dominant market theme.
• Lily Liu (Solana): Anticipates a global debate pivoting from Universal Basic Income (UBI) to Universal Basic Ownership (UBO) and Universal Basic Opportunity.
• Dennis Gada (Infosys): Forecasts "Security" as the primary focus, with AI being used to solve AI-induced vulnerabilities and protect enterprise data.
• Tony Kim (BlackRock): Warns of a near-term period of uncertainty and economic "rewiring," followed by a long-term era of unprecedented possibilities in longevity, space, and disease cure.