OpenAI is reportedly preparing to trial ads inside ChatGPT’s Android beta, as profit has significantly failed to catch up with the exploding user base and ballooning infrastructure bill.
An analysis by HSBC Global Investment Research concludes that OpenAI still isn’t likely to be profitable by 2030 — and warns the firm will need at least US$207 billion in additional funding to sustain its growth plans.
HSBC’s projection includes a scenario in which, by 2030, ChatGPT and related services reach roughly 44 percent of the world’s adult population (up from about 10 percent in 2025).
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).
Yet even under this optimistic adoption scenario — and after assuming robust subscription and enterprise-AI growth — the infrastructure and compute burden threatens to overwhelm revenues. HSBC models show roughly US$792 billion in cloud and AI infrastructure costs from late 2025 to 2030, with a data-center rental bill of about US$620 billion alone.
Longer term, commitments balloon further: compute obligations could reach US$1.4 trillion by 2033 under current deals.
That math explains why OpenAI reportedly is already seeking more funding, even while it experiments with ads. The ad test may be part of a broader strategy to diversify revenue streams beyond paid plans and enterprise deals — precisely because the compute and infrastructure costs are so staggering, they can swallow growth if not offset by fresh capital or new revenue models.
Why The Infrastructure Costs Are So Massive
The pressure stems largely from recent multiyear cloud-compute and data-center rental deals that lock in enormous capacity commitments. Among them are a roughly US$250 billion agreement with Microsoft and a US$38 billion deal with Amazon Web Services (AWS), combining to secure around 36 gigawatts of AI compute power by the end of the decade.
To put that into perspective: HSBC’s analysts estimate that even with such deals, OpenAI’s “free cash flow + other liquidity” through 2030 would remain negative, leaving the US$207 billion gap.
Because AI models — and their inference demand from millions of users — scale rapidly, compute costs grow rapidly. The larger and more widely used the models become, the more infrastructure you need: data centers, specialized hardware, cooling, electricity, bandwidth, and maintenance. For a platform forecasting widespread global adoption, this translates into high recurrent costs, not one-off capital expenditures.
This high fixed-cost structure helps explain the company’s pursuit of non-subscription revenue streams like advertising: they may be essential to closing the widening deficit.
Ads Inside ChatGPT: A Potential Survival Mechanism — with Trade-offs
That brings us back to the ad test in ChatGPT’s Android beta. For OpenAI, ads offer a way to monetize the conversation and search traffic generated by its hundreds of millions — soon billions — of users. If properly implemented, ads could tap into digital ad budgets just like search engines or social platforms do, but with far greater targeting precision, thanks to the conversational and contextual signals available.
At a time when infrastructure spending threatens to outstrip even aggressive revenue forecasts, the ad strategy may reflect more than a convenience — it could be part of a lifeline to keep the platform afloat without perpetual infusions of external capital.
But that approach carries risks: delivering ads inside conversational AI could erode user trust, blur the lines between unbiased information and commercial content, and raise privacy concerns — especially when the system already handles personal data. For a tool many use for work, research, education, coding, or writing, introducing commercial placements changes the value proposition substantially.
What This Means for the Broader AI and Tech Industry
OpenAI’s situation encapsulates a fundamental tension now emerging across the frontier-AI industry: scaling to global reach demands massive compute infrastructure, but monetizing that scale — via subscriptions, enterprise sales, ads — remains uncertain. HSBC’s analysis suggests that even with high growth, the cash flows are unlikely to catch up to the cost curve without continuous capital infusions or breakthroughs in monetization.
For investors, cloud providers, chipmakers, and other infrastructure partners, that raises critical questions: Is the current AI megacycle sustainable, or is the industry building what some analysts now call a “money pit with a website on top”?
It also means regulators, privacy advocates, and users need to watch closely. As AI platforms like ChatGPT play an increasingly central role in how people access information, learn, work, and create, their business models will shape not only the tools themselves but the incentives behind them — from what kind of content gets surfaced to how user data is used and monetized.