The expense of AI services may lead industry leaders to adopt subscription-based models
Google is reportedly planning to introduce fees for AI-enhanced search features, marking a significant change to its revenue model. This move is seen as a necessary response to the high costs of providing such a service. Experts believe that this shift will compel other industry leaders to adopt subscription models to sustain their services. According to the Financial Times, Google’s proposal involves offering the new search feature exclusively to users of its premium subscription services, which are already required for access to AI assistants in other Google tools like Gmail and its office suite.
Google is testing a new search experience with selected users, utilizing its generative AI to provide direct, single-answer responses in a conversational style similar to ChatGPT and other competitors.
Heather Dawe, chief data scientist at UST, notes that AI search is more costly to compute than Google’s traditional search processes. Therefore, by charging for AI search, Google aims to recoup these expenses.
There’s a significant focus in AI on the substantial cost of computing power required to train advanced generative models. Amazon, for instance, reportedly spent $65 million on a single training run last year, according to engineer James Hamilton, who predicts the company will soon surpass the $1 billion mark in training costs.
OpenAI and Microsoft recently revealed plans to construct a $100 billion data center dedicated to AI training. Additionally, in January, Mark Zuckerberg expressed his intention to invest at least $9 billion solely in Nvidia GPUs.
However, according to Brent Thill, an analyst at Jefferies investment firm, the cost of training AI represents just a fraction of the overall expenses in this sector. Thill stated in a briefing note that the majority of AI computing expenditure is currently allocated to running, rather than training, models. He also noted that over 90% of AI computing costs are currently focused on inferencing, the process of querying an AI model. This trend is attributed to the rapid growth in inferencing costs, driven by the increasing deployment of models and tools into production.
He went on to say, “Some companies have opted to price new Gen AI features on a monthly basis, assuming that higher fees will offset usage costs. Alternatively, others have chosen a per-usage pricing model to mitigate cost risks. Some have even integrated these features into existing plans in the hopes of driving user growth.”
Competitors in the AI search space are offering similar subscription plans. Perplexity, an AI-powered search engine, does not display advertisements but offers a $20 monthly “pro” tier that grants access to more advanced AI models and unlimited usage.
On the other hand, some companies are still offering their products at a loss. Microsoft’s Bing provides free access to its AI features but requires the use of the company’s Edge browser. Arc, a browsing and search startup, offers its products for free to users and plans to generate revenue in the future by charging companies for business features.
