As development costs for artificial intelligence projects continue to rise, more AI projects are reassessing their monetization plans and examining how to balance the books on the back of the trillions being sunk into AI infrastructure.
At the same time, consumer sentiment around AI is declining as more AI options that seemingly offer little value get pumped into every other app. The AI influx has sparked broader concern about an AI bubble, and raised questions about a hyped technology that may not ultimately deliver on its promises.
In recent weeks, xAI signed a deal with Anthropic to rent out data center capacity for $15 billion per year, essentially assisting a rival in its development. Meanwhile, Meta is looking to charge AI developers for access to advanced models and capacity as part of its latest subscription offerings.
According to a new report from The Information, Meta is also looking to embed technical staff in corporate partners to help them integrate its AI models. In addition, CNBC reported that Meta may want to rent out data center capacity to other businesses if it ends up having more space than it needs.
The moves suggest that pressure is rising on AI developers to demonstrate actual value, as well as a pathway towards making real money. Because while the latest AI tools are a technical achievement, various reports have suggested that most businesses that have integrated AI systems aren’t seeing significant benefit. Furthermore, for most regular users the true value of AI is still not clear.
For example, a study published earlier this year by the National Bureau of Economic Research found that among nearly 6,000 CEOs, chief financial officers and other executives, the vast majority reported seeing little operations-level impact from AI.
As per the report: “Executives report little own-firm impact of AI over the last 3 years, with nine-in-ten reporting no impact on employment or productivity.”
The report also found that the same executives predicted AI would boost productivity at their firms by an average of 1.4%, raise output 0.8% and cut employment 0.7% over the next three years.
Those gains may be notable, but are less impressive when matched against the increasing costs of AI usage.
As reported by Business Insider, Uber Operations Chief Andrew Macdonald recently noted that it’s becoming harder to justify AI costs within the company. He said that while the company has consistently blown through its AI token usage budget, that higher token usage has not translated into a proportional increase in useful consumer features.
Meta’s own AI Superintelligence Chief Alexandr Wang acknowledged the AI backlash, as well as the challenge ahead of developers in terms of demonstrating the real-world value of their models.
In a recent interview with The Core Memory podcast, Wang said, “On some fundamental level we haven’t yet demonstrated, in a very real way, how this is actually a tool for personal empowerment or personal agency, or how it just makes people’s lives a lot better.”
Wang added that while many people find AI tools helpful, and some people find they make things a little better, overall user experiences are not overwhelmingly better. It’s different for developers, though. Developers, Wang said, have a very positive view of AI, because it’s changed the way they work.
“[Developers] can build so many more things faster, they can build entire projects over a weekend, and it’s just this incredible enabler of personal agency,” Wang said. “That moment hasn’t happened for everyone else in the world yet.”
This is seemingly the fundamental misalignment with the perception of AI. The development community is lauding the value of AI tools and driving massive AI investment, and in return is seeing huge benefits from these tools. However, regular users and businesses are not.
Is that because most people haven’t realized the power of AI tools to assist in their own tasks as yet, or is it that AI is fundamentally more aligned with certain tasks and functions, but practically useless for others?
Developers rely on replicated code patterns and systematic processes that can be replicated by AI tools. Those tools, in turn, are able to provide guidance and automation, within these processes. But AI tools are less valuable for an SMB looking to sell more stuff. Sure, AI ad targeting is driving better responses sometimes, but the practical use cases are limited. Meanwhile, the examples that consumers are being shown via AI content generation and smart search engines just aren’t impressive enough to justify the trillions of dollars in spend.
And again, add to this the fact that many businesses are reporting that they’re not seeing efficiency gains as a result of AI adoption, and the value gap does seem apparent. This could also lead to a rationalization of the AI industry.
Another consideration, as reported by CNBC, is the rising competition in the space, and the fact that China-based AI developers are rolling out less expensive models that could undercut these massive U.S. projects.
The overall picture does suggest an AI bubble, which has been inflated by massive hype, and still has a way to go before it’s able to demonstrate practical value to most users.
