JPMorgan Chase’s recent analysis on AI infrastructure investments highlights a stark reality for the sector: to achieve even a modest 10% return on the trillions in projected spending through 2030, AI products and services would need to generate approximately $650 billion in annual revenue perpetually.
This figure underscores the immense scale required to justify the buildout, amid warnings of a potential “AI bubble” if demand doesn’t keep pace. Global AI-related capital expenditures like data centers, chips, and compute are forecasted to total $5–7 trillion over the next decade, driven by hyperscalers like Microsoft, Google, and Amazon.
In 2025 alone, these firms are on track for ~$350 billion in AI infra spend—a 60%+ year-over-year jump. At 10% ROI: $650 billion/year ~0.58% of global GDP. Lower hurdle 6% ROI, Drops to $360 billion/year.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
Higher hurdle 12% ROI; rises to $810 billion/year. JPMorgan illustrates the challenge with everyday equivalents: Equivalent to ~$35/month from each of the world’s 1.5 billion iPhone users. Or ~$180/month from each of Netflix’s 300 million subscribers.
However, the bank stresses that corporations—benefiting from AI-driven productivity gains—would bear most costs, not consumers directly. Early adopters already report $35+/month in time savings per user.
Echoing past tech overbuilds like 2000s telecom fiber, JPMorgan notes a $1.4 trillion funding gap for data centers alone, power constraints limiting new capacity to 122GW through 2030, and the danger of idle infrastructure if AI adoption slows.
Even leaders like OpenAI’s Sam Altman have flagged excess capacity concerns. This projection contrasts with JPMorgan’s own aggressive AI push: The bank is investing $18 billion in tech for 2025 up $1 billion YoY, with over 175 AI use cases live, yielding $1.5 billion in savings from fraud prevention, personalization, and efficiency.
Globally, JPMorgan sees “astronomical” compute demand but cautions that end-user value must accelerate to avoid a bust. The report, shared widely on platforms like X, has sparked debate—some view it as a bubble signal, while others including JPMorgan analysts rebutting skeptics like Michael Burry argue AI’s productivity upside makes the math feasible.
Sam Altman, is one of the most influential voices in AI, blending optimism about its transformative potential with pragmatic concerns about risks, infrastructure, and governance. While stressing the need for responsible scaling, democratic oversight, and alignment to avoid misuse.
Altman often describes AI progress as a “gentle singularity,” a gradual but exponential shift toward superintelligence that empowers humanity rather than overwhelming it. He views AGI (artificial general intelligence) as achievable and imminent, but downplays its drama:
My guess is we will hit AGI sooner than most people think and it will matter much less. Superintelligence, he predicts, could arrive by 2030, enabling breakthroughs beyond human limits. Altman is bullish on 2025–2027 as a pivotal period of rapid advancement, outpacing recent years.
Altman sees AI development as an exponential curve, with 2025 marking the entry of AI agents into the workforce—autonomous systems handling cognitive tasks like coding or analysis, boosting company output.
He outlines ambitious internal goals: an automated AI research intern by September 2026 running on hundreds of thousands of GPUs and a full AI researcher by March 2028. By 2026, AI could generate “novel insights,” accelerating discoveries in fields like medicine and physics.
In a recent X post, he shared OpenAI’s latest report on progress, highlighting recommendations for scaling responsibly. He predicts that by 2035, individuals could access intellectual capacity equivalent to the entire 2025 global population.
AI agents join workforce; small discoveries possible. Transforms knowledge work (e.g., coding, analysis); economic output surges. Automated AI intern; novel insights from AI. Speeds scientific breakthroughs; recursive self-improvement begins. Abundance in intelligence/energy; “anything else” becomes possible.
Universal access to vast intellect. Democratizes genius reshapes society, work, and creativity. Altman is “determinedly optimistic,” arguing AI will elevate humanity through abundance: cheaper intelligence nearing the cost of electricity, turbocharged economies, and solutions to grand challenges like curing diseases.
He envisions a “Cambrian explosion” in creativity via tools like Sora, where AI democratizes art and entertainment. AI agents will act as “virtual coworkers,” enhancing productivity without fully replacing humans. In a July 2024 X post, he stressed AI’s national security value: “AI progress will be immense from here, and AI will be a critical national security issue.
He advocates for U.S.-led coalitions to ensure AI remains “democratic” and benefits all, preventing authoritarian monopolies. While hopeful, Altman acknowledges dangers: a potential “AI bubble” akin to the dot-com era, driven by surging investments (e.g., OpenAI’s $1.4 trillion compute commitments over eight years).
He warns of misuse by rogue actors (e.g., cyberattacks) and societal harms like job displacement or AI addiction. His “doom score” isn’t zero, but he focuses on mitigation: layered safety value/goal alignment, reliability, robustness.
In a November 2025 X thread, he clarified OpenAI’s stance against government bailouts, emphasizing market accountability: “If one company fails, other companies will do good work.” He calls for technical alignment and societal adaptations, like universal compute access as a “human right.”
On user impacts, he worries about over-reliance (e.g., AI as “therapist” reinforcing delusions) and advocates treating “adult users like adults” while measuring long-term well-being. OpenAI plans 30 gigawatts of compute, with ambitions for 1 gigawatt weekly by reducing costs potentially halving capital expenses.
Altman pushes for U.S.-built fabs, energy, and data centers to maintain competitiveness, viewing it as essential for economic edge. Revenue projections: $20B annualized run rate in 2025, scaling to hundreds of billions by 2030, funding via equity, debt, and AI cloud sales.
He critiques uneven distribution, favoring “techno-capitalism”: encourage wealth creation but widely share benefits to raise both floor and ceiling. OpenAI’s 2025 restructure—to a public benefit corporation governed by a nonprofit—aims to attract capital while prioritizing humanity’s benefit, with $25B committed to health and AI resilience.
In his “Gentle Singularity” essay, he envisions a future of “wildly abundant” ideas and energy, with AI enabling personalized lives and resilience through widespread distribution. Reflecting personally, he sees AGI as “the most important technology humanity has yet built,” worth the “painful” effort despite work-life trade-offs.