“Technology firms are promoting novel AI technology capable of generating business memos or computer code, but they are still grappling with how these products can translate into profitability. These generative AI tools are untested and come with substantial operating costs, demanding robust servers equipped with expensive chips that consume significant power. Giants like Microsoft, Google, Adobe, and others in the tech industry are experimenting with various strategies for creating, marketing, and pricing their AI offerings.
In some instances, companies have incurred losses on their early generative AI products. For instance, Microsoft reportedly experienced financial losses on one of its initial generative AI offerings. As a response, Microsoft and Google are launching AI-backed updates to their software with higher price points. Zoom Video Communications, on the other hand, has attempted to manage costs by occasionally employing a simpler in-house-developed AI. Adobe and similar companies are placing restrictions on monthly usage and implementing consumption-based pricing models.
The cost of running these AI models has left many customers dissatisfied with the expenses involved. This presents a challenge for companies to determine how AI can be utilized and how much customers are willing to pay for it. Chris Young, Microsoft’s Head of Corporate Strategy, acknowledges that it will take time for both companies and consumers to determine the optimal use cases for AI and its associated pricing.
Developing and training AI products is a lengthy and expensive process, often requiring years and hundreds of millions of dollars. AI lacks the economies of scale seen in standard software, as it demands intensive computations for each query. Increased usage by customers also escalates infrastructure costs, potentially leading to losses for companies charging flat fees for AI services.
Some companies utilize highly powerful AI models, which demand substantial computational resources and strain computer processors more than standard software or cloud services. Microsoft, for example, employs OpenAI’s GPT-4, one of the largest and most expensive AI models available. Such models are costly to operate, which can affect a company’s profitability.
Companies are optimistic that generative AI costs will decrease over time, similar to trends observed in other technologies like cloud storage and 3-D animation, thanks to advancements in hardware and innovation. OpenAI, for instance, reduced its charges for its older AI this year, with customers now having to pay for the latest versions. Despite the uncertainty surrounding revenue models, the market for AI-related companies has witnessed significant growth in valuation. OpenAI, for instance, is reportedly in talks with investors for a share sale that could value the company at up to $90 billion, a substantial increase from earlier this year. However, as enthusiasm for AI continues, company executives are expected to closely scrutinize the associated costs in the near future.”