Engines & LLMs
Analyst Predicts AI Shakeout: Is an LLM ‘Extinction’ Coming?
Get ready for a reality check in the AI world. According to market analysts at Gartner, the booming market for large language model (LLM) providers might be heading for a major shakeout – think of it as an “extinction phase” where only the strongest survive.
Why the grim prediction? It boils down to money. Building and running these powerful AI models costs an absolute fortune, requiring massive investment in computing power and data (what analysts call capital expenditure or “capex”). Gartner predicts that, much like the cloud computing market eventually became dominated by giants like AWS, Microsoft Azure, and Google Cloud, the LLM space will likely consolidate around just a few major players who can afford the steep price tag.
Right now, many AI companies are focused purely on getting users and building market share (“adoption”) rather than turning a profit. John-David Lovelock, a VP Analyst at Gartner, stated, “Adoption right now is the most important thing… there is extinction coming. The market will not be able to support this number of model providers that we currently have.”
Gartner also suggests that most businesses (the CIOs making tech decisions) will prefer to buy AI features built into software they already use, rather than trying to develop their own complex AI models from scratch. They’ll look for reliable, off-the-shelf solutions.
However, don’t expect a sudden crash like the dotcom bubble burst. Lovelock suggests this will be a “slower pruning” rather than an overnight collapse. While the overall spending on generative AI is predicted to soar (reaching $644 billion in 2025), the number of companies actually providing the core LLMs is expected to shrink significantly.
Our Take
Okay, so Gartner thinks a bunch of the AI companies building these big language models might go bust soon? Makes sense when you think about it. We see tons of cool AI demos, but building and running things like ChatGPT or Gemini costs insane amounts of cash for computing power.
It feels like only the giants with super deep pockets (your Googles, Metas, OpenAIs, Anthropics) can really afford to play the long game here. This sounds like the AI gold rush might be settling down, leading to fewer, bigger players calling the shots. Probably means more focus on actually making money from AI soon, too.
This story was originally featured on The Register.