OpenAI's User Surge vs. HSBC's Compute Cost Warning: Balancing Growth and Financial Reality

OpenAI's trajectory paints an ambitious picture of dominance. Company leadership projects 220 million paying ChatGPT users by 2030 – roughly 8.5% of an estimated 2.6 billion weekly users – up sharply from 35 million (about 5%) at present. This explosive growth potential underpins forecasts of $129 billion in annual revenue according to reports. The path so far has been impressive: ChatGPT reached 800 million weekly active users by November 2025, generating $10 billion in annual recurring revenue. User engagement is similarly strong, with 5.8 billion monthly visits and 2 billion daily queries.

However, this growth optimism collides with stark financial headwinds warned by major financial institutions. HSBC executives, alongside General Atlantic, caution that Big Tech's annual AI infrastructure investments exceeding $380 billion could outpace revenue generation. They project global data center costs soaring to $5.2 trillion by 2030. OpenAI's own infrastructure deals, described as $1 trillion, exemplify the scale of spending driving these concerns. The HSBC CEO and others explicitly warn of a potential capital misallocation "bubble," drawing parallels to the dot-com era, where massive spending might not yield commensurate tangible returns.

The tension centers on sustainability. OpenAI's current $10 billion revenue stream must eventually cover vastly expanding compute costs projected to reach $5.2 trillion globally by 2030. While OpenAI's leadership focuses on user penetration and revenue scaling, HSBC highlights the fundamental risk: whether projected user growth and revenue can realistically absorb infrastructure expenditures at this unprecedented scale without straining profitability. This cost reality check tempers pure growth narrative enthusiasm.

Scaling Heights, Soaring Costs

OpenAI is scaling at an astonishing pace. ChatGPT's weekly active users doubled from 400 million to 800 million in under a year, hitting 700 million by August 2025. This user surge translates directly into revenue muscle: annual recurring revenue reached $10 billion by June 2025, with monthly revenue hitting $1 billion and a staggering 591.6% year-over-year growth from March 2024 to March 2025. The platform's dominance is undeniable, holding an 81.13% generative AI market share in March 2025 and 62.5% in AI subscription tools by August 2025.

This explosive adoption fuels massive investment. OpenAI plans to pour over $1 trillion into infrastructure through 2032 according to projections, projecting 220 million paying users by 2030. However, the path to profitability is clouded by compute costs. Despite the $200/month "Pro" tier designed for high-volume users, CEO Sam Altman confirmed it remains unprofitable due to the sheer cost of processing its users' 2.5 billion daily queries. HSBC analysts estimate total expenses could balloon to $620 billion in 2030, raising questions about the sustainability of current pricing models against these escalating operational demands.

While user growth and revenue momentum are undeniable, the high compute costs, especially for premium services, present a significant friction. OpenAI's massive capital commitments signal long-term conviction, but the immediate profitability of its high-value Pro tier remains elusive, highlighting the critical challenge of managing costs as usage scales.

Competitive Threats and Cost Constraints

Despite its early lead, OpenAI faces mounting competition that could erode its market position. Analysts project its consumer AI market share will fall from 71% today to just 56% by 2030, squeezed by rivals like Anthropic, Elon Musk's xAI, and Google's Gemini 3 platform. This contraction comes even as the company anticipates adding 220 million paying ChatGPT users by that year. The pressure is intensifying as major tech firms collectively plan $380 billion in AI capital spending for 2024 alone. While OpenAI has committed to $1 trillion in infrastructure investments through 2032, HSBC's CEO warns these massive outlays may outpace near-term revenue generation. Productivity gains and consumer willingness to pay are projected to materialize only after several years of continued investment.

The escalating compute costs present a particularly daunting challenge for all players in the AI race. McKinsey estimates data centers alone will require $5.2 trillion in spending by 2030, creating unprecedented capital intensity. OpenAI itself faces potential expenses reaching $620 billion in 2030 despite generating only $13 billion annually. This investment blitz strains energy infrastructure as AI operations demand more power than conventional data centers. Microsoft and other companies are already pursuing off-grid power contracts to secure reliable energy supplies, highlighting grid capacity limitations. While these investments aim to build future moats, analysts warn of potential misallocation risks in this early-stage sector. The extended timelines for revenue realization versus capital expenditure create sustained margin pressure, making the profitability trajectory inherently uncertain despite the long-term growth thesis.

Catalysts and Scenarios: Efficiency Gains vs. Bubble Risks

What matters now is whether efficiency breakthroughs can outpace the relentless demand growth driving AI infrastructure spending.

The potential for dramatic efficiency improvements exists, with McKinsey noting DeepSeek's reported 18x reduction in training costs as one example of technology that could mitigate the massive compute requirements fueling the AI expansion. This efficiency potential represents a critical variable in OpenAI's growth thesis, as lower costs could accelerate adoption and improve margins across the industry. However, this optimistic scenario faces a powerful countervailing force: the Jevons Paradox, where technological efficiency paradoxically increases overall consumption rather than reducing it.

The scale of projected infrastructure spending underscores why some analysts view current investment levels with concern. HSBC's CEO has warned that massive AI investments, including OpenAI's $1 trillion infrastructure commitments, may outpace revenue generation capabilities, with meaningful productivity gains and consumer willingness to pay materializing only after several years. This perspective aligns with McKinsey's broader projection of $5.2 trillion in data center costs by 2030, of which AI workloads account for 70% of growth. The sheer magnitude of these projected investments creates significant risk that capital allocation could become misaligned with sustainable economic fundamentals.

The reality likely sits somewhere between these competing forces. While efficiency gains could moderate cost pressures in the medium term, the underlying demand drivers-enterprise integration needs, generative AI adoption, and competitive infrastructure races-suggest spending will remain robust regardless of technological improvements. Investors should monitor whether productivity gains from these investments materialize as expected, as this will determine whether current spending levels represent strategic investment or speculative overreach.