OpenAI's Ad Run Rate, Outcome Pricing, SoftBank's $5.5B Bid

Three separate developments dropped Monday, each pointing at the same underlying pressure: OpenAI needs to turn its massive cost base into durable, diversified revenue before any path to public markets can close.

Key points

• OpenAI’s advertising unit has reached a $1 billion annualized run rate, roughly seven months after ChatGPT ads launched in February 2026.
• That pace looks impressive in isolation, but the company is tracking to miss its $2.4 billion 2026 ad-revenue projection by a wide margin if growth does not accelerate sharply in Q4.
• OpenAI is piloting outcome-based pricing with major enterprise clients, charging only when AI tasks are completed successfully rather than on a per-token or subscription basis.
• SoftBank’s SB Energy unit has offered $5.5 billion to cement its data-center partnership with OpenAI, deepening the financial entanglement between the two organizations.
• All three stories together sketch a company that is simultaneously chasing advertising scale, restructuring enterprise contracts, and locking in infrastructure commitments at extraordinary cost.

Does the ad business have room to close the gap?

A $1 billion annualized run rate is a real number. Reaching it in roughly seven months shows that advertisers are willing to test ChatGPT as a channel. The problem is the math. To hit $2.4 billion for the full year, OpenAI would need to dramatically accelerate in the back half, and the report from The Information signals the company may fall short. That gap matters for investors tracking OpenAI’s broader revenue projections, which have repeatedly been revised or missed. Advertising was positioned as a high-margin complement to subscriptions and API revenue. A shortfall here does not sink the business, but it does raise questions about the reliability of forward guidance the company has shared with backers.

One structural issue is that ChatGPT ads are still relatively new inventory. Advertisers tend to ramp slowly on unfamiliar formats, and Q4 seasonality could provide a genuine tailwind. Whether that is enough to close a gap of more than $1 billion in annualized terms in a single quarter is unclear.

What does outcome-based pricing change for enterprise customers?

The shift to pay-only-on-completion contracts is a meaningful commercial pivot. Under a traditional token or API pricing model, the customer absorbs all execution risk: if the model fails a task, they still pay for the compute. Under outcome-based pricing, OpenAI takes on a share of that risk. That makes the product easier to justify internally for enterprise buyers, particularly in cost-sensitive areas like customer support automation.

The catch for OpenAI is margin discipline. If the company prices these contracts too aggressively to win deals, and model failure rates are higher than expected, the economics deteriorate quickly. The report notes that rivals are moving toward similar structures, which means competitive pressure is at least partly driving the decision. Still, for an enterprise sales motion, outcome-based pricing is a credible signal that OpenAI believes its models are reliable enough to stake revenue on performance.

How does SoftBank’s $5.5 billion offer fit the bigger picture?

SoftBank’s relationship with OpenAI is already substantial, running through the Stargate joint venture and SoftBank’s own balance sheet commitments. SB Energy’s $5.5 billion offer to secure the partnership adds another layer to that dependency, this time on the infrastructure and energy side. Data center capacity is a genuine constraint for frontier AI training and inference, and locking in a well-capitalized partner with energy assets has obvious operational logic.

From an investor standpoint, the number underscores how capital-intensive OpenAI’s buildout remains. The company is not just spending on research and headcount; it is anchoring long-term infrastructure deals worth billions. That creates durable cost obligations that need to be serviced by revenue growth, which loops back directly to the advertising shortfall and the outcome-pricing experiment. The three stories are connected: OpenAI is building an expensive foundation and is still working out which revenue models will pay for it.

Nothing in these reports constitutes investment advice, and this site has no affiliation with OpenAI.