Why AI Stocks Are Falling: Hype Cycle Theory Explains the Generative AI Downturn

Hype Cycle Theory on Why AI Stocks are Significantly Down

Listen to this blog in a 4-minute podcast

Armand here, RPost’s armadillo product evangelist. Over the last weeks, main AI stocks have dropped significantly off their highs. By main ‘Generative AI’ infrastructure stocks, I am considering the infrastructure to deliver generative AI, and the public companies with major investments like chips (Nvidia, AMD…), infrastructure operating these (Amazon, Oracle, Microsoft, Google…), and models (Microsoft, Meta, Google…). 

Now, when you look at Intelligent Applications using AI, values are significantly high and durable. By ‘Intelligent Applications’ I am considering companies like Palantir, CrowdStrike, and RPost with its Raptor™ AI security intelligence.

Why, might you ask, are the major AI infrastructure stocks down and the intelligent application stocks maintaining or growing value?

Some blame the infrastructure stock dip on tariffs for infrastructure raw materials, others on the new model entrants in China like Deep Seek creating uncertainty; but I have a different theory. Consider the Gartner Hype Cycle for Artificial Intelligence (see chart).

The Hype Cycle tracks the evolution of development of technology innovation from inception, to the inevitable inflated expectations, the challenging trough of disillusionment (the realization that achieving those expectations in the technology is going to be more challenging or take longer than expected) and then the slope of enlightenment where the hard work is put in to achieve the success from the technology at scale. Finally, the plateau of productivity, when the technology is humming and mainstream.

If you consider the Gartner Hype Cycle for Artificial Intelligence (2024) you will notice that Generative AI reached the peak of hype in late 2024 and is now on the downslope with people realizing that the technology is incredible, but the game changing applications are for widespread use at the enterprise level are still not there. Sure, those that need quick research, drafts of briefs written, a helper to edit documents or messages, or even a friend who does not sleep yet loves to brainstorm – these are great applications but are not enterprise-wide game changers for most companies.

The technology will surely get there – and fast – but this realization that it is not here yet as was expected may be what is causing these AI infrastructure stocks to dip. The AI infrastructure cost is significant, and to fund this R&D, operations costs, and continuous powering-up investment, AI infrastructure companies are going to need well-paying enterprise organizations creating mass efficiencies with killer AI applications. 

At the application level (not simply an AI wrapper that makes your prompts pre-written, but truly complex applications that use a variety of aspects of AI), these companies are and will continue to soar without reprieve. On the cybersecurity use of AI, you additionally have urgency.

You see, as discussed this week at the Gartner CIO Leadership Forum in London, cybercriminals have found the killer app for AI. It is the ease of generating an AI clone of someone who is influential in an organization, messaging and joining web meetings disguised in these AI impersonations. And as AI advances, these AI clones will be unleashed as agents with contextualized personalities.

Palantir, CrowdStrike, and RPost are leaders in use of ‘good guy’ AI to counter cybercriminal ‘evil empire’ AI. RPost’s Raptor™ AI security model, assistant, and agent can see when cybercriminals or threat actors have effectively deployed sleeper cells (compromised devices, applications, or email accounts) at third parties (outside of one’s network) to power up their reconnaissance with aim to gather **context** -- who is communicating with whom about what when and in what communications style. 

Context is king, for these sophisticated threat actors.

Context is the needed food for these threat actors to power-up their nefarious AI clones and impersonations. Without context, the AI clone will not be believable. With context, these impersonations will fool even those most savvy.
  
Back to the Hype Cycle. Intelligent applications (e.g., RPost’s Raptor™ AI security) are powerful tools solving challenging problems and are on the Hype Cycle Slope of Enlightenment. Hence, growth in value.

The AI infrastructure stocks centered on core Generative AI, as you can see in the chart, is having a wakeup call that results are coming but are not here **yet**.

But surely, they will be soon.

Request a demo or discussion about RPost’s Raptor™ AI security. (CLICK TO CONTACT US FORM)

You might also like