##  [# AI Bubble Defense: Why Investors Need to Actively Manage Tail Risk](/sections/advisor-center/ai-bubble-defense-why-investors-need-actively-manage-tail-risk) 

 

# AI Bubble Defense: Why Investors Need to Actively Manage Tail Risk

 

 

AI expectations are soaring, but so are AI bubble risks. Learn why liquidity is the key to gauging AI risk and a strategy for managing the fat left tail in markets right now.



 

 

 

 

 [![ETF.com](/sites/default/files/styles/author_image_icon/public/2023-12/etf.com_400x400.png?itok=LO34LQE1)](/contributors/etfcom-staff) 

[By ETF.com Staff](/contributors/etfcom-staff)

 Nov 03, 2025

 Edited by: ETF.com Staff

 

 

 [ + Follow ](/etf/login) 

     Share  <a class="a2a a2a_button_email"> Email </a><a class="a2a a2a_button_linkedin"> LinkedIn </a><a class="a2a a2a_button_facebook"> Facebook </a><a class="a2a a2a_button_x"> X (Twitter) </a> 

 

 

 

 

 

 

 

 

  
            googletag.cmd.push(function() {
                googletag.display('js-dfp-tag-article_page_302x26');
            });
    
    

 

 

  

 



 

 

  Loading 

 

 



 
            let duration = 0, percent = '0', new_percent = '';
            jwplayer("etfVideo").setup({
                playlist: "https://cdn.jwplayer.com/v2/media/E7F7R2gC",
                aspectratio: "16:9",
                autostart: "viewable",
                ga: {}
            });
            jwplayer().on('time', function(e) {
                duration = e.duration
                if (e.position >= duration * 0.25){
                    percent = '25';
                } 
                if (e.position >= duration * 0.50){
                    percent = '50';
                } 
                if (e.position >= duration * 0.75){
                    percent = '75';
                } 
                if (e.position == duration){
                    percent = '100';
                }
                if (new_percent != percent){
                    // GTM Datalayer - Video Event
                    dataLayer.push({
                        'event': 'video',
                        'video_id': 'E7F7R2gC',
                        'video_percent': percent,
                        'video_url': 'http://www.etf.com/markdownify/node/133625' 
                    });
                    new_percent = percent;     
                }
            });
         

 



 

 

The AI investing play continues to prove enormously beneficial for investors. However, today's soaring markets carry substantial embedded risk. Cem Karsan, founder and CIO of Kai Volatility Advisors, breaks down the issues, opportunities, and how to gauge AI bubble risk with Dave Nadig, President and Director of Research, in a recent [Excess Returns](https://www.youtube.com/channel/UCPYvx_y92dvI1PSdiho0ALw) episode. The following is a snippet of their conversation centered on the AI bubble and liquidity.

**Transcript**

## Opening: On Black Swans and Swimsuits

**Karsan:** You have to be resigned to the fact that you also cannot be swimming without a swimsuit, in the words of Warren Buffett, right? You need to understand that there are black swans and things at any time that can happen.

 
            googletag.cmd.push(function() {
                googletag.display('js-dfp-tag-in_article_unit');
            });
    
    

 

 

## AI Risk: A Problem of Scope 

**Karsan:** A lot of these AI investments that they're making are big, structural. These are not going to play out in the next year or two. They're going to play out in the next five plus years, and I think the expectations assume that we're going to stay above 60,000 feet for the next five to 10 years. And at 70, 80,000 feet right now, I don't think that's very good for the next five to 10 years.

**Nadig:** What's your canary in the coal mine then? I mean, obviously price is an obvious one. Things go down 20% – that's a signal. But is there anything in your day-to-day that you look at to help calibrate the, you know, “Well, if we’re not five years, but we're not saying ‘Katy bar the door, it's all going to hell in a handbasket today.’” How are you gauging that on an ongoing basis? How do you play in that, “Well, it's not tomorrow, but it's not 10 years”? How do you play in that space?

**Karsan:** Yeah, liquidity is the best predictor in the short term. Right? And now looking forward and trying to determine – or being able to catch when that liquidity turns precisely – is incredibly difficult. And you can't just look at one thing. You have to think of the things that may drive those liquidity outcomes. And put probabilities on them, and have early indicators for when those things might start to go.

And so you have to have kind of a risk dashboard, and think about kind of what those things are. And you can't, by the way, predict all of them. You have to be resigned to the fact that you also cannot be swimming without a swimsuit, in the words of Warren Buffett, right? You need to understand that there are black swans and things at any time that can happen. We may be able to predict with some – the gray swans and some of the other things that come along the way – but there is always going to be things that you cannot predict and price. And knowing that you're that far off the ground, you need to use that as a risk management tool.

## Balancing AI Bubble Risk and Opportunities

**Karsan:** The left tail is fat. We know that. And so knowing that, that part of the distribution, guess what: is underpriced. Doesn't mean you shouldn't be net long, that you shouldn't be playing long. By the way, I think in the short term, the right tail is also underpriced. It doesn't mean go short. It just means manage that left tail. And so I think at baseline you need to think about the world distributionally and hedge those risks at baseline.

But I think in between there you can also manage probabilities and understand when there's a better chance of something or not and be dynamic and active along the way, and have that dashboard to help understand, okay, maybe there's a higher probability. Maybe that one percent probability is done up to one-and-a-half or two given X, Y, and Z. And so maybe we need to be a little bit more aggressive on that tail or lighten up a little bit for some short period until this X, Y, Z shows something better. But I think that's the general approach. I can get into dashboard items and things that we're concerned about, but there's quite a few.

**Nadig:** Most of them are around liquidity?

**Karsan:** All of them have an *effect* on liquidity.

*Don’t miss the rest of this in-depth conversation on AI risk, liquidity, and options. To see the full discussion,* [***go here***](https://youtu.be/qBoeOAViXJI?si=AZvGib7vHnPeweSa)*.*



 

 

 [ + Follow ](/etf/login) 

 [ETF.com Staff](/contributors/etfcom-staff) 

 

 

  etf.com is the single source for ETF intelligence. We provide real-time ETF news and analysis to educate investors and drive financial knowledge in…   [View Bio](/contributors/etfcom-staff)

 



 

 


 Related Topics  [Artificial Intelligence](http://www.etf.com/topics/artificial-intelligence) 

 [Advisor Center](http://www.etf.com/topics/advisor-center-0) 

 [ETF.com Videos](http://www.etf.com/topics/videos)