OpenAI Ads Push, Margin Gains, and Amazon Talks

This update is a roundup of same-day reporting from the linked sources below, with editorial context from the CPJ Stock Desk.

Three separate reports in the past week paint a picture of OpenAI pushing hard on revenue diversification, with improving unit economics and a possible new cloud partnership entering the conversation.

Key points

OpenAI is making progress on introducing advertising inside ChatGPT, according to The Information.
Compute margins reached 70% as of October 2025, up from 52% at end of 2024 and roughly double the January 2024 rate.
Margin improvement is driven largely by better economics on business-facing products rather than consumer subscriptions alone.
Jim Cramer flagged reported Amazon-OpenAI talks, though the scope and status of any deal remain unclear from available sourcing.
All three data points arrive as OpenAI continues its transition toward a for-profit structure, keeping IPO-readiness questions front of mind.

What is the advertising push actually about?

OpenAI appears to be in active discussions about showing ads inside ChatGPT, per a report from The Information cited by Seeking Alpha. The sources stop short of confirming a launch date or format. That said, the framing of “making progress” suggests this has moved beyond early-stage brainstorming into something more concrete, at least internally.

Advertising would represent a significant strategic shift for a product that has, until now, monetized primarily through subscriptions (ChatGPT Plus, Team, Enterprise) and API access. The logic is straightforward: ChatGPT has a very large user base, and ad-supported tiers could expand that base further while creating a revenue stream that scales with usage rather than requiring users to pay directly. The risk, equally straightforward, is user experience and brand perception. OpenAI has positioned ChatGPT as a premium, trusted tool. Poorly executed advertising could undercut that positioning quickly.

For investors watching the IPO trajectory, an ad business adds a new revenue line but also a new variable to model. Advertising revenue is typically tied to engagement metrics and ad market cycles, neither of which OpenAI has publicly disclosed in detail.

Why the margin improvement matters more than it might look

The compute margin figure is the most concrete financial data point in this week’s news. According to The Information, as reported by Fortune, OpenAI’s compute margins hit 70% in October 2025. That compares to 52% at the close of 2024 and roughly 35% in January 2024. The trajectory is steep and sustained.

Compute margin, broadly, reflects how much revenue OpenAI retains after covering the infrastructure costs of running its models. A jump from 35% to 70% in under two years suggests meaningful gains in model efficiency, better pricing on compute contracts, or a revenue mix shift toward higher-margin products. Fortune’s reporting points to business sales as a key driver, implying enterprise and API customers carry better margins than consumer subscriptions.

This matters enormously in the context of a potential public offering. OpenAI has faced persistent questions about whether it can convert its enormous revenue growth into actual profitability. A 70% compute margin does not mean the company is profitable overall. Sales, marketing, research, and headcount costs are substantial and not captured in that figure. But it does indicate that the core product economics are improving faster than many outside observers expected, which is a prerequisite for any credible path to operating profitability.

What should investors make of the Amazon reports?

This one warrants caution. Jim Cramer mentioned Amazon-OpenAI talks on air, characterizing it as one of a pattern of broad technology partnership announcements he labeled “Lazy Susan deals.” He called Amazon a “serious company” in the context, but the available summary does not specify what the deal would cover, whether it involves AWS infrastructure, Alexa integration, investment, or something else entirely.

OpenAI already has a deep infrastructure relationship with Microsoft Azure, which complicates the picture of any Amazon cloud arrangement. It is worth treating this report with appropriate skepticism until more specific terms emerge. Partnership announcements in AI have become frequent enough that the signal-to-noise ratio is genuinely low. A real Amazon deal with material financial terms would be significant. A vague collaboration announcement would be much less so.

The bigger picture heading into 2026

Taken together, these three reports suggest OpenAI is actively working to widen its revenue base (advertising), demonstrate improving unit economics (compute margins), and potentially add new distribution or infrastructure relationships (Amazon). Each piece individually is incomplete. The advertising push has no confirmed timeline. The margin data is from October and lacks full-company profitability context. The Amazon talks lack confirmed scope.

Still, for a company widely expected to pursue a public offering within the next few years, the direction of travel matters. The narrative heading into 2026 is one of a business trying to prove it can grow revenue and improve margins simultaneously, which is exactly what public market investors will want to see in any eventual S-1.