Is OpenAI Running Out of Cash!?

Are you wondering what’s really going on at OpenAI — and whether any of it should change how you’re thinking about your AI search strategy?

You’re not alone. In the space of a few weeks, OpenAI shut down its video generation tool Sora (reportedly costing a million dollars a day to run for just 500,000 users), closed its instant checkout feature for eCommerce, and saw its early advertising tests publicly dismissed as a flop by Walmart’s VP Daniel Danker. For a company approaching 900 million weekly active users, the pace of pullbacks has raised serious questions about where OpenAI is heading — and how quickly it needs to get there.

For marketing leaders, none of this is just industry gossip. If your business is already showing up in ChatGPT results — or planning to — the strategic direction OpenAI takes next will directly affect where your brand gets visibility, how ads might soon appear alongside AI answers, and which platforms deserve your optimisation budget. This article breaks down what’s actually happening, what it means for marketers, and what you should be doing about it right now.

Executive Summary

OpenAI is losing roughly three times more than it earns. While it isn’t in immediate financial danger, reports suggest the company is targeting profitability by 2029 — and it has a much more pressing deadline before then. With an IPO planned for Q4 2026, OpenAI needs to demonstrate a credible revenue model to public investors. That model, increasingly, looks like advertising.

Two senior hires from Meta — Dave Duggan as VP and Head of Global Ads, and Fidji Simo as CEO of Applications — signal that ad revenue is now OpenAI’s top commercial priority. But early beta tests have underperformed, and competitors like Google’s Gemini and Anthropic’s Claude are gaining ground in both consumer and enterprise markets.

Meanwhile, OpenAI has stepped back from trying to own the full eCommerce transaction — handing that role to Shopify through its new agentic storefronts — and is cutting products like Sora that drain cash without generating meaningful returns. For marketers, the takeaway is clear: ChatGPT remains the dominant AI referral platform by a significant margin, but the landscape is shifting fast, and the businesses tracking their AI search performance now will be best positioned to adapt.

How Much Money Is OpenAI Actually Losing?

The numbers are stark. Reports indicate that OpenAI is spending approximately three times what it earns. While the company has a substantial cash reserve, it’s far from infinite — and the direction of travel matters more than the current balance sheet.

Charlie Marchant, CEO of Exposure Ninja, puts the Sora situation into perspective: “If it’s leaking a million dollars daily on Sora for just 500,000 users, it’s clearly not worth the investment for them.”

The broader context makes this even more significant. OpenAI has instructed its lawyers to begin preparations for an IPO expected in Q4 2026 — rumoured at one point to be among the largest in history. Luke Nicholson, COO of Exposure Ninja, explains the pressure this creates: “I think that they want to have the public investors. They want some good news for those guys. They want to be able to point to some revenue increases.”

Profitability itself isn’t expected until around 2029, according to available reports. But OpenAI doesn’t have until 2029 to start showing progress. It has until Q4 this year to demonstrate a working revenue engine — and that engine is increasingly looking like advertising.

Download Our 2025 State of AI Search Report

See how AI is reshaping search, which brands are already winning, and 2026 predictions for your AI Search strategy.

Download my free copy

Why Is OpenAI Betting Everything on Advertising?

The short answer is that OpenAI’s competitors have already cornered the enterprise market — and advertising is the natural monetisation model for a mass-market consumer platform.

Consider the positioning. Anthropic’s Claude has established a strong foothold with developers through Claude Code and with enterprise users through tools like Claude Cowork — an agent that can access your browser and complete tasks in the background. Microsoft has Copilot deeply integrated across its entire product suite. Google has Gemini embedded within the workspace tools that millions of businesses already use daily.

As Luke puts it: “Everybody apart from OpenAI knew that OpenAI was gonna have to go for ads. If you are the mass market, then the way that you monetise that mass market is by serving the mass market ads.”

The clearest signal of this strategic shift is in the hiring. OpenAI has brought in Dave Duggan as the new VP and Head of Global Ads, and Fidji Simo as CEO of Applications. Both are Meta veterans who were instrumental in building Meta’s advertising platform into the revenue powerhouse it is today. You don’t hire people with that pedigree unless advertising is your absolute top priority.

OpenAI has also launched a beta ads feature. At least one Exposure Ninja client is currently on the waitlist to test it. The aim is clear: get ads generating sustainable revenue ahead of the Q4 IPO, because without a functioning ad platform, the valuation could fall well short of expectations.

Charlie highlights just how significant the competitive challenge is: “Platforms like Google Ads, Bing Ads, Meta Ads — they are comparatively much more sophisticated. They’ve had so much more time in market, they know what they’re doing. There’s really clear analytics.”

For eCommerce businesses in particular, Google Ads is already deeply embedded in their revenue model. Persuading those advertisers to shift budget towards a platform that is still working out its fundamentals is a significant ask.

Why Did OpenAI’s Early Ad Tests Flop?

The early signals have not been encouraging. Walmart’s VP Daniel Danker publicly criticised the ChatGPT ad beta, with the core complaint being that shoppers simply weren’t completing purchases inside the ChatGPT interface. They preferred to navigate to Walmart’s website instead.

Luke identifies the root cause as a design philosophy problem: “Previously OpenAI has underestimated how difficult it is to make ads a functioning part of the overall system. Early on when OpenAI started talking about ads, they were really, really focused on the user and how the ads wouldn’t really interfere with the experience.”

The tension here is one that every ad-supported platform eventually has to confront. Make ads too unobtrusive and they don’t perform — which means advertisers won’t spend. Make them too intrusive and you damage the user experience that built your audience in the first place. Every successful ad platform in history has had to navigate this balance, and most have taken years to get it right.

Luke frames the urgency: “This is a huge piece of work. It actually has a massive set of implications for our business. And we’re on a bit of a timeline here because we’re burning our money so quickly and we want to get these ads up and running, generating revenue in a sustainable way ahead of the Q4 IPO.”

For marketers considering whether to test ChatGPT ads when access becomes available, the lesson is worth noting. The attribution and analytics infrastructure that makes platforms like Google Ads and Meta accountable simply doesn’t exist yet in ChatGPT. That doesn’t mean it won’t — but it does mean early adopters need to go in with realistic expectations about what they’ll be able to measure and prove internally.

What Does the Shopify Agentic Storefront Deal Mean for Retailers?

One of the most interesting developments is what OpenAI has chosen not to do. Last week, OpenAI closed its instant checkout feature — the system that allowed eCommerce purchases to happen directly inside ChatGPT conversations. Almost simultaneously, Shopify announced its agentic storefronts, which allow users to browse and buy products within ChatGPT — but with Shopify handling the actual merchant process, owning the customer data, and managing the transaction.

This is a meaningful strategic retreat. Rather than trying to become the merchant, OpenAI is stepping back into the role Google has played for decades: a discovery platform that connects users with products, while the transaction happens elsewhere.

Luke sees this as part of a broader correction: “The idea was we would never have websites anymore because everything that you would do, you would just do via a chatbot. I think that excitement has now faded a little bit and people are realising, you know what? The internet is pretty great.”

For eCommerce businesses, there’s a practical consideration worth flagging. Shopify’s agentic storefronts are currently opt-out — meaning if you’re a Shopify seller, you’re included by default. The cost is an additional 4% per transaction on purchases made via ChatGPT. For Google’s AI Mode, Gemini, and Microsoft Copilot, there are no additional fees — at least for now.

Charlie highlights the potential friction this creates: “I think some retailers, especially in the US where card and Visa fees are already high, are gonna be like, fine, 4% more per transaction. But I think there’s gonna be a lot of retailers who are like, that is huge.”

If your profit margins are already tight — and in many retail verticals they are razor-thin — that 4% could be a dealbreaker. And as Luke points out, the competitive pressure from fee-free alternatives means OpenAI can’t afford to overcharge: “OpenAI are not the only game in town, and one of the problems that they face is anywhere that they drop the ball even slightly, they’ve got people nipping at their heels.”

The practical step here is simple: if you’re on Shopify, check whether agentic storefronts are enabled on your account. Run the numbers on whether that 4% works for your margins. If it doesn’t, consider opting out and focusing your AI commerce visibility on the fee-free alternatives instead.

Who Is Winning the AI Market Share Race?

Despite all the turbulence, ChatGPT remains dominant. Industry-wide figures put its market share at roughly 64–68% of AI chatbot usage. Gemini has grown steadily through late 2025 and early 2026, now commanding around 20% — a significant improvement, but still a long way from catching ChatGPT.

What’s more telling, though, is what Exposure Ninja is seeing in its own client data. Across B2B, B2C, eCommerce, and lead generation websites in the SME and mid-market space, ChatGPT accounts for closer to 80% of AI referral traffic and revenue — even higher than the headline industry figures suggest.

Luke shares the detail: “Both in January and February we saw Gemini’s users rise in general on average across our client portfolio, getting up to around 20%. I’d be interested to see if we can see that hit 25% in March.”

There’s an important nuance in the data, though. Gemini and Claude users appear to convert at a higher rate than ChatGPT users. The traffic from those platforms tends to be more qualified, more engaged, and more likely to result in a conversion. So while ChatGPT drives the sheer volume, the smaller platforms may deliver disproportionate value per visitor.

The bigger strategic question is whether ChatGPT’s dominance is sustainable. Consumer switching between AI platforms is genuinely difficult — most people default to the tool they already know, and changing habits takes real friction. But the enterprise market tells a different story.

Charlie sums it up: “Getting people to switch between apps, getting people to switch over to Gemini from ChatGPT, or to Claude from ChatGPT — I think it’s not that easy for the average searcher when we’re not thinking about work.”

The real vulnerability for OpenAI isn’t that consumers will actively choose a different chatbot. It’s that Microsoft, Google, and Anthropic are embedding AI features directly inside the enterprise tools people already use — Copilot within Microsoft 365, Gemini within Google Workspace, Claude within developer workflows. If users can get AI-powered assistance without opening a separate app, their ChatGPT usage naturally declines without them ever making a conscious decision to switch.

As Luke frames it: “OpenAI isn’t running out of cash, but it’s possibly running out of time in which to convince the rest of the world that all of the investment in it is worthwhile.”

What Should Marketers Be Doing Right Now?

The competitive dynamics between AI platforms are fascinating, but the practical question for marketing leaders is straightforward: what does any of this actually change about your strategy?

The honest answer is that ChatGPT is still the platform driving the vast majority of AI referral traffic and conversions for most businesses. That makes it the priority. But it’s not the only platform worth paying attention to — particularly as Gemini and Claude grow their share and appear to deliver higher-quality, higher-converting traffic.

If you haven’t already set up tracking for your AI search traffic, that’s the single most valuable thing you can do this week. In Google Analytics, you can create a custom channel group that separates out traffic from ChatGPT, Gemini, Claude, Perplexity, Microsoft Copilot, and other AI platforms. This gives you a clear view of not just how much traffic each platform sends, but — more importantly — which platforms drive actual conversions for your business.

Charlie recommends using Looker Studio for a clearer picture: “You can set it up so that it shows a lovely pie chart and you can see what the biggest part of the pie is for your referral traffic from those platforms, but also more importantly, from the conversions from those platforms as well.”

Once you can see the data, you can make informed decisions about where to focus your optimisation efforts. The businesses that are tracking this now have a genuine advantage — because whatever you hear about the market in general may not apply to your specific audience, your specific industry, or your specific customer base.

Luke reinforces this: “The most important thing to do is find out what your customers are doing and making sure that you’ve got a system in place to track that. Otherwise, whatever you hear in general may not apply to you specifically. It pays to know your numbers.”

Next Steps

This week:

• Set up AI search tracking in your Google Analytics.
• Create a custom channel group that separates ChatGPT, Gemini, Claude, Perplexity, and Copilot traffic.
• Look at both volume and conversion rate by platform to understand where your highest-value AI referral traffic is actually coming from.

Next 30 days:

• If you’re on Shopify, review whether agentic storefronts are enabled on your account.
• Run the numbers on the 4% per-transaction fee and decide whether it works for your margins.
• If it doesn’t, consider opting out and directing your AI commerce visibility towards the fee-free alternatives through Google AI Mode and Microsoft Copilot.

Next 90 days:

• Build an AI search visibility strategy that covers the platforms your data shows are actually driving results.
• Don’t optimise for every AI platform equally — let your analytics guide where you invest.

If you’d like expert guidance on where to start, request a free AI search audit from Exposure Ninja that covers traffic analysis, conversion data, and page-level priorities for the platforms that matter most to your bottom line.

In Conclusion

OpenAI isn’t about to go bankrupt. It has substantial cash reserves and is approaching 900 million weekly active users. But it is under real pressure — to cut costs, build a revenue model that justifies its valuation, and do both before a Q4 2026 IPO that the entire tech world is watching.

For marketers, the practical implications are clear. ChatGPT is still the dominant source of AI referral traffic and conversions, and that’s unlikely to change overnight. But the ground is shifting beneath it. Gemini is growing. Claude is carving out the enterprise and developer markets. And OpenAI’s pivot to advertising means the way brands appear inside ChatGPT is about to change — probably sooner than most marketing teams are prepared for.

The businesses that will navigate this transition best are the ones that already know their numbers. They know which AI platforms are driving traffic, which are driving conversions, and where the highest value per visitor actually sits. If you don’t have that visibility yet, start there. Everything else flows from it.

Is your marketing underperforming?

Request a website and marketing review and our team will tell you how to improve your marketing.

Request a Free Review