Despite Big Tech jumping on generative AI en masse, it’s far from a cash cow for them as it stands. According to research by the Wall Street Journal, Big Tech is incurring hefty losses when offering the solutions.
Big Tech has fully embraced generative AI. Nevertheless, the companies in question are failing to actually make a return from the new solutions. They are still incurring heavy losses when offering generative AI implementations, WSJ concludes.
One example cited by the business newspaper is Microsoft’s GitHub Copilot. This developer-oriented solution now has more than 1.5 million users.
The tech giant charges $10 per user per month for GitHub Copilot, but Microsoft loses an average of $20 per month per end user. Heavy operational costs are to blame. Generative AI solutions demand an enormous amount of computing power just to run them. This is apart from all the development costs put into training the underlying models.
Another problem that WSJ has identified is that generative AI models have a large capacity and therefore many capabilities. Currently, end users are only tapping into a small portion of their potential.
Demanding more money
Providers of generative AI solutions are trying to boost the disappointing revenues with a number of adjustments. For example, they’ve been hiking prices for the solutions.
Microsoft, for example, plans to charge an extra $30 for an AI version of Microsoft 365. The cheapest version of the productivity suite is now $10, but the AI version will allow automatic composition of Excel sheets, emails and PowerPoint presentations and other operations.
Google also wants to charge customers an extra $30 for AI features in its productivity solutions. OpenAI charges $20 a month for using GPT-4 in ChatGPT.
Limitations generative AI models
In addition, providers may want to deploy less powerful AI models for simpler tasks. Microsoft is developing simpler models for Bing and for search only. For example, based on open-source models.
In imaging tool Firelfly, for example, Adobe opts for credits. If the credits run out, the tool’s operation is slowed down.
In short, large tech companies must quickly determine how they are going to make money from generative AI services.
Tip: European telco operators call on Big Tech to pay for upgrades