TECH NEWS – A Wall Street analyst is already looking ahead to a potential bursting of the artificial intelligence bubble. OpenAI has unveiled its GPT-5 large language model (LLM), which has received mixed feedback and sparked broader debates about sustainable monetization for these resource-intensive chatbots and LLMs.
Users have highlighted multiple drawbacks that limit GPT-5’s usefulness: significantly longer response times, unnecessarily complex response structure and syntax, and inability to retain context despite a larger context window. On Twitter, Philoinvestor posted an analysis comparing Netflix’s monetization path with that currently followed by AI chatbots and LLMs.
AI chatbot distributors currently have one recipe for sustainable revenue: increase average revenue per user (ARPU) by balancing free and paying users, while cutting costs. Unlike Netflix — whose personalized content raised the opportunity cost of switching platforms — most AI chatbots and LLMs are not meaningfully differentiated in raw capability, allowing users to switch freely between platforms.
The difference between Chatbots & the Netflix setup.$NFLX Netflix went from breakeven to printing cash in two ways:
– Increasing its # of users and
– increasing its revenue PER user.That resulted in higher total revenue and higher contribution margin at the user level… pic.twitter.com/5vOkJ7IrPt
— Philoinvestor (@philoinvestor) August 10, 2025
Even as costs rise to run increasingly complex models, distributors are competing to retain users by offering unnecessary extras, eroding overall cost-effectiveness. AI chatbots may never become the revenue engines their creators envision.
Meanwhile, Bank of America analyst Michael Hartnett pointed to a key signal that could precede an AI bubble burst: “Expect concentrated US stock returns (Mag7, AVGO, ORCL, and PLTR making up 80% of S&P 500 gains since Trump’s ‘Liberation Day’) to continue until tech credit spreads widen. That will be the signal that AI cash burn is threatening the overbuilt AI trade. The same thing happened in the second half of 1999, and it was the subsequent recession that triggered the productivity boom of the 2000s.”
Roughly 250 data centers are currently under construction in the US. This massive AI investment surge is already impacting GDP, which could grow by as much as $624 billion (2.08% of GDP) in 2025. According to Hartnett’s thesis, overspending will reduce capital efficiency, potentially leading to the collapse of the current AI mania.
Leave a Reply