By Tom Dotan and Deepa Seetharaman

Tech companies are touting new AI technology that can spit out business memos or computer code. They are still figuring out how those products will generate a profit.

Generative artificial-intelligence tools are unproven and expensive to operate, requiring muscular servers with expensive chips that consume lots of power.

$Microsoft (MSFT.US)$ 0.78%increase;  $Alphabet-C (GOOG.US)$, $Adobe (ADBE.US)$ and other tech companies investing in AI are experimenting with an array of tactics to make, market and charge for it.

Microsoft has lost money on one of its first generative AI products, said a person with knowledge of the figures. It and Google are now launching AI-backed upgrades to their software with higher price tags.

$Zoom Video Communications (ZM.US)$ 0.95%increase;

has tried to mitigate costs by sometimes using a simpler AI it developed in-house. Adobe and others are putting caps on monthly usage and charging based on consumption.

A lot of the customers I’ve talked to are unhappy about the cost that they are seeing for running some of these models,” said Adam Selipsky, the chief executive of Amazon.com’s cloud division, Amazon Web Services, speaking of the industry broadly.

It will take time for companies and consumers to understand how they want to use AI and what they are willing to pay for it, said Chris Young, Microsoft’s head of corporate strategy.

We’re clearly at a place where now we’ve got to translate the excitement and the interest level into true adoption,” he said. 

Building and training AI products can take years and hundreds of millions of dollars, more than with other types of software. 

Artificial intelligence is taking on a larger role in white-collar work, with the ability to draft emails, presentations, images and more. Workers have already lost their jobs to the technology, and some CEOs are changing hiring plans.

AI often doesn’t have the economies of scale of standard software because it can require intense new calculations for each query. The more customers use the products, the more expensive it is to cover the infrastructure bills. These running costs expose companies charging flat fees for AI to potential losses.  

Microsoft used AI from its partner OpenAI to launch GitHub Copilot, a service that helps programmers create, fix and translate code. It has been popular with coders—more than 1.5 million people have used it and it is helping build nearly half of Copilot users’ code—because it slashes the time and effort needed to program.

It has also been a money loser because it is so expensive to run.  

Individuals pay $10 a month for the AI assistant. In the first few months of this year, the company was losing on average more than $20 a month per user, according to a person familiar with the figures, who said some users were costing the company as much as $80 a month.

Microsoft and GitHub didn’t respond to requests for comment on whether the service is earning money. The profitability picture for GitHub Copilot and other AI-powered assistants will change if computing costs come down.

Microsoft is going with a higher price for its next AI software upgrade. On top of regular monthly charges—starting around $13 for the basic Microsoft 365 office-software suite for business customers—the company will charge an additional $30 a month for the AI-infused version. The AI-powered feature can be instructed to compose emails, create PowerPoint presentations and build Excel spreadsheets independently.

Google, which is releasing a similar AI assistant feature for its workplace software, will also be charging $30 a month on top of the regular subscription fee, which starts at $6 a month.  

Microsoft, Google and others have gone with a flat monthly rate, betting that the higher additional charges will more than cover the average expenses of running the technology.

Adobe leaders unveiled Adobe Firefly, a family of generative AI models, at a summit in Las Vegas this year.

One of the reasons the services can be so costly is some use the most potent AI models available, which require more power and put more strain on computer processors to operate than standard software or cloud services.

Microsoft, for example, is using OpenAI’s latest software for its AI features. The version, called GPT-4, is among the largest and most costly AI models available. 

Using it to summarize an email is like getting a Lamborghini to deliver a pizza. 

They require massive compute power,” said Jean-Manuel Izaret, the head of the marketing sales and pricing practice at Boston Consulting Group. “They require massive intelligence.”

Microsoft has been exploring using less powerful and cheaper AI tools for its Bing search engine, including some built with

$Meta Platforms (META.US)$’ open-source AI software, said people familiar with the discussions. 

The Information, a technology-focused news outlet, earlier reported that Microsoft was considering other tools for Bing.

Zoom has developed a smaller, cheaper software for its AI assistant, powered by multiple models, including those made by OpenAI and Meta. It only uses the most powerful AI for the most difficult tasks.

Zoom doesn’t have to charge extra for the tool, which can summarize meetings and compose chat messages, because it avoids expensive AI when it can, said Smita Hashim, Zoom’s head of product, who has also worked at Microsoft and Google.

We are pretty amazed and shocked at the big price tickets we are seeing from some of the competitors,” she said.

Adobe uses a system of credits to help ensure its AI image generator, Firefly, won’t put it into the red. Once Adobe customers go over their allotted monthly credits, the company slows down the speed of the service to discourage overuse. 

“We are trying to provide great value but also protect ourselves on the cost side,” Adobe CEO Shantanu Narayen said. 

Companies expect generative AI to get cheaper over time, as happened with many technologies, including cloud storage and 3-D animation. New chips and other innovations will likely drive the processing costs down.

OpenAI lowered the price it charges for using its older AI earlier this year. Its free viral chatbot ChatGPT uses an older version of its software. Customers who want a ChatGPT powered by its latest version have to pay $20 a month. 

The uncertainty about moneymaking models scared off investors. The shares of companies connected to AI have soared this year. OpenAI is talking to investors about a share sale that would value it at
as much as $90 billion, roughly triple its level earlier this year.

The rising valuations reflect enthusiasm about the promise of AI, said May Habib, CEO of Writer, which builds generative AI tools for enterprise companies. She said executives will soon be taking a harder look at the costs.

Next year, I think, is the year that the slush fund for generative AI goes away,” Habib said.