OpenAI's ambitious ad revenue goals look like AI-generated money, analysis says

[Alfonso Maruccia]

AI Fail: Not long ago, AI companies such as OpenAI had nearly sworn they would never show advertising on their chatbot services. Even CEO Sam Altman described ads as a last-resort business model and said pairing them with AI felt uniquely unsettling. Now, everyone is doing the exact opposite. But according to recent research, industry players are unlikely to earn as much as they ambitioned from chatbot advertising.

A new market analysis from consulting firm Emarketer casts serious doubt on OpenAI's advertising revenue targets. The company began testing ads in ChatGPT in February, adding another revenue stream as it looks for ways to offset the enormous cost of operating its AI services. Emarketer's forecast suggests that OpenAI's internal projections far exceed what the broader US chatbot advertising market is likely to support.

OpenAI said it expects to generate $2.5 billion in advertising revenue by the end of this year, while setting its sights on a massive $100 billion revenue target by 2030. Emarketer's study suggests that OpenAI's advertising revenue could fall 90% short of its 2030 target, and that's just the first piece of bad news in a long series of setbacks for AI companies looking to generate meaningful revenue from advertising.

Emarketer estimates that standalone chatbot services from OpenAI, Microsoft, Google, Amazon, and others will collectively generate less than $1 billion in US advertising revenue this year. OpenAI, by comparison, expects to generate $2.5 billion on its own. By 2030, Emarketer sees the market reaching just $5.4 billion.

That forecast does not cover all advertising associated with AI. It excludes much of the spending on conventional ads placed alongside "AI search" content, such as Google's AI Overviews. Emarketer expects the broader US AI advertising market to reach $68.25 billion by 2030, but says most of that spending will remain outside standalone chatbot conversations.

Analysts concede that OpenAI's advertising revenue goals are not impossible, just highly unlikely unless massive changes in customer behavior turn the traditional advertising market upside down.

First, OpenAI could indeed see its advertising revenue skyrocket if advertisers completely abandon search engines such as Google, social media, and other traditional channels in favor of a fully chatbot-based advertising experience.

Then, the ChatGPT maker would need to overcome the competition in this new ad-driven business to reach its expected $100 billion revenue target within five years. In theory, OpenAI expects advertising to account for a third of its projected future revenue.

Will math and financial speculation finally overcome market realities, turning the chatbot business into a money-printing machine, just as AI companies have promised for years? AI evangelists such as SoftBank chief Masayoshi Son are still willing to make enormous bets on that future. OpenAI, meanwhile, is reportedly planning roughly $600 billion in compute spending through 2030, underscoring just how much revenue it will need before the economics begin to work.

Permalink to story:

OpenAI's ambitious ad revenue goals look like AI-generated money, analysis says

 

[torvuswalds]

AI has enormous potential, but I hope we don't repeat the internet's biggest mistake: monetizing attention above all else. The moment ad revenue becomes the primary metric, user trust starts competing with shareholder expectations, and history suggests shareholders usually win.

 

[Lowhangingfruit]

Direct to consumer ads are not a meaningful business for OpenAI. They are an AI infrastructure company. They provide the models that every other software company uses. That's how they make money. From B2B transactions.

 

[psycros]

Direct to consumer ads are not a meaningful business for OpenAI. They are an AI infrastructure company. They provide the models that every other software company uses. That's how they make money. From B2B transactions.

Apparently that's not really paying off, if all the reports are to be believed. The expenditures are enormous and we're already seeing big companies moving AI in-house. Why would you pay to support colossal data centers when your business only needs a small percentage of that capability? Just buy it and manage it yourself, setting your own standards for usage and behavior.

 

[NumberNine]

OpenAI was surprised by demand when they put the first demo up for ChatGPT3. They assumed they'd have a much longer path to relevance, but everyone and their mom started reaching out and throwing money at OpenAI, far beyond their wildest expectations. Governments even got involved where they could find benefit.

This was the pivotal point, the key moment where they alone could unilaterally chart their own path. What did they do? Yet another subscription service. Now with ads. A gated attraction, just like the rest of the internet.

 

[Theinsanegamer]

OpenAI was surprised by demand when they put the first demo up for ChatGPT3. They assumed they'd have a much longer path to relevance, but everyone and their mom started reaching out and throwing money at OpenAI, far beyond their wildest expectations. Governments even got involved where they could find benefit.

This was the pivotal point, the key moment where they alone could unilaterally chart their own path. What did they do? Yet another subscription service. Now with ads. A gated attraction, just like the rest of the internet.

So what should they have done then? How do you suggest they should make money to fuel this enterprise while making use of the technology?

 

[toooooot]

Human patience and toleration of ads has a limit. When everything screams in your face buy me buy me, a potential customer becomes deaf to it. There are already too many people who have zero tolerance for ads. They treat it like an endless source of attention. Like if they could show ads 24/7 for as long as there was a pair of eyes. Less is better, for advertisers too.

 

[p51d007]

AI has enormous potential, but I hope we don't repeat the internet's biggest mistake: monetizing attention above all else. The moment ad revenue becomes the primary metric, user trust starts competing with shareholder expectations, and history suggests shareholders usually win.

When HASN'T a company in the last forever years NOT tried to monetize something?