Apparently generative AI isn’t all that profitable

  • According to a Wall Street Journal report, generative AI products aren’t generating enough revenue to turn a profit.
  • The immense demand for compute resources means that generative AI platforms will have to charge users more for access.
  • This could limit the number of customers, further impacting revenues from the emerging technology.

Every year since 2020 has been marked by a big tech product that captured the imagination of the public. First, we had non-fungible tokens, which are dead in the water now. Then we had aspirations of interconnected online realms in the metaverse, and now we have generative artificial intelligence (generative AI).

Since ChatGPT rose in popularity toward the end of 2022, generative artificial intelligence has only become more popular with firms incorporating the tech into their products and solutions. Now, however, a report from Wall Street Journal alleges that these solutions may be struggling to turn a profit.

Citing a source with “knowledge of the figures”, the publication reports that Microsoft has lost money on one of its generative AI products. Further to that, Zoom is trying to implement the technology on a budget with an in-house solution and Adobe is implementing usage caps as well as charging users based on their use of its AI tech.

The report reveals that, unlike traditional software, AI doesn’t enjoy the same economies of scale. This is because each query and prompt for a user requires extreme amounts of processing power.

For example, Microsoft charges users $10 for its GitHub Copilot but on average, the company is losing $20 per users with some costing it as much as $80 per month.

This cost will come down as the cost to compute decreases, but in the meantime, the companies developing AI solutions will need to increase the cost of access. This has the downside of users turning away from the service. For example, Microsoft will charge an additional $30 for access to an AI-infused version of Microsoft Office 365, a service that costs $13 for the base version.

With that in mind and considering what generative AI can do right now (write emails, generate presentations and other basic tasks) that’s a lot of dough to be spending on things that ultimately still need to be checked and verified by a human.

We are still in the early days of generative AI’s development, but already there is controversy surrounding how these platforms are trained, what data is used and how the technology will affect the employability of human beings.

As the hype fades, generative AI looks like the latest metaverse. That is to say, a solution that drew in billions in investment only for it to be ignored as soon as the new hotness was revealed.

Of course we simply can’t tell what will happen with generative AI in the future but things don’t look promising.

[Image – M Rezaie on Unsplash]


About Author


Related News