The company told Bloomberg News Tuesday (April 28) that it is “firing on all cylinders” despite a report that said it had missed internal targets, and that it is still seeing growing demand from business customers and its new advertising operation.

“The mood internally is incredibly positive,” the company said in a statement.

This came after a Wall Street Journal reported that OpenAI had fallen short of several internal targets as competitors encroached on its territory. The same story described tension between company CEO Sam Altman and Sarah Friar, OpenAI’s finance chief.

OpenAI told Bloomberg the report was “prime clickbait.”

Meanwhile, a separate report from The Information said OpenAI is projecting a giant shift in subscription revenue.

In the last three years, that report noted, the startup has pulled in the bulk of its revenue from consumers’ $20-per-month ChatGPT subscriptions. Now, however, the company is anticipating that a cheaper, ad-supported subscription tier will attract new users but also lead existing subscribers to downgrade.

The report, citing internal projections, said OpenAI is hoping it will generate more revenue by selling ads to more users than depending on its existing flagship monthly subscription service, ChatGPT Plus.

At the start of the year, OpenAI had projected that consumer subscribers to ChatGPT Go would see a 36-fold jump to 112 million this year, The Information said. Executives have thus forecast that the number of subscribers to ChatGPT Plus will drop 80% to about 9 million. Subscribers to the most expensive Pro plan will double but will still account for less than 1% of the total.

Nevertheless, the company sees revenues more than doubling to $30 billion this year and reaching $284 billion in 2030, with ads bringing in more than a third of its revenue.

The Information argued that these forecasts illustrate how reliant OpenAI is on consumers, even as it focuses on landing enterprise clients.

In other AI news, PYMNTS wrote earlier this week about the forecasting problem confronting enterprise customers of AI providers.

“Traditional software costs tracked headcount,” that report said. “AI costs track activity. A single employee can generate thousands of AI interactions in a day. Another may trigger none. An automated process can run continuously without anyone watching the bill.”

As covered here, enterprise AI invoices have begun to resemble utility bills more than software subscriptions. Charges are connected to model activity, not employee count. Finance teams designed around stable annual renewals now find themselves managing a cost structure with no previous reference point.