A recent study by the renowned analysis firm Gartner reveals that by the end of 2025, at least 30% of generative AI business projects currently under testing will be ruthlessly abandoned by enterprises. This startling statistic was announced at the Gartner Data & Analytics Summit held in Sydney, Australia, delivering a significant blow to the AI industry.
Why is this the case? The answer is high costs! Gartner has found that early adopters of generative AI are struggling with rising expenses. Imagine, deploying an AI project could cost between $5 million and $20 million—no small sum!
Image source: Image generated by AI, provided by the image licensing service Midjourney
For instance, if you want to design a customized generative AI model, such as fine-tuning the Llama model with industry-specific data, the initial investment could range from $5 million to $6 million, with an additional annual cost of $110,000. Developing a model from scratch would require a hefty $20 million! Even a simple document search feature would start at $750,000.
No wonder enterprises are looking glum! Gartner states that regardless of the ambition, generative AI requires a higher tolerance from businesses, as it is an indirect investment and should not be expected to yield immediate results.
Analyst Rita Sallam from Gartner said, "Last year, AI was hyped so much, and now executives are impatient to see returns. However, businesses are finding it difficult to justify and realize the value of AI."
However, there is not entirely hopeless. Gartner surveyed 822 business leaders and found that after adopting AI, 15.8% of enterprises saw increased revenue, 15.2% saved costs, and 22.6% experienced improved productivity. Although the numbers are not high, they at least prove that AI has some utility.
So, business owners, if you want to master AI, you must carefully calculate the costs! Analyze the business value and total costs, not only looking at direct returns but also considering future impacts.
Key Points:
🚨 Gartner predicts: 30% of generative AI projects will be abandoned by the end of 2025, primarily due to high costs.
💰 Staggering costs of AI projects: Custom models can reach $5-6 million, with full development costing up to $20 million.
📊 Mixed results from AI applications: About 15% of businesses see increased revenue or cost savings, 22.6% experience improved productivity.