Arman Eshraghi on Market Hype Cycles, Cryptocurrencies, & How Fund Managers Really Make Decisions
Arman Eshraghi is an Honorary Professor of Finance and Investment at Cardiff Business School & a Visiting Researcher at the Cambridge University Endowment for Research in Finance and the Cambridge Centre for Alternative Finance.
By Aiden Singh, February 17, 2026
Introduction
Professor Arman Eshraghi is Chair of Finance and Investment at Cardiff Business School, where he leads the Cardiff Fintech Research Group and the fintech theme at the Digital Transformation Innovation Institute. His research investigates how financial technology, behavioral biases, and human psychology shape the decisions of investors, fund managers, and firms, with a focus on market dynamics, corporate finance, and investment management.
Professor Eshraghi combines interviews, empirical analysis, and interdisciplinary methods to uncover the hidden forces driving markets, from cryptocurrency speculation and AI adoption to fund manager psychology and behavioral anomalies in trading. His work examines how real-world conditions, such as sleep and attention, interact with human decision-making to produce patterns that challenge conventional financial theory.
In our conversation, Professor Eshraghi and I explore the mechanics of market hype, the psychological and behavioral drivers of cryptocurrency trading, the cognitive tensions faced by professional fund managers, and how human behavior continues to shape financial markets in ways that defy traditional models.
- - - - - - -
Market Hype Cycles
Aiden Singh: I would like to begin by discussing hype in financial markets more broadly, and how it can be modeled. Could you explain the dynamics of market hype cycles?
Arman Eshraghi: Fascination with new technologies or, more broadly, anything that is new and shiny, has been a feature of human behavior for centuries. Historically, we have observed financial bubbles driven by hype. A classic example is the Dutch tulip bubble of the seventeenth century, when the prices of tulips - particularly exotic varieties - became extremely inflated, ultimately leading to a localized financial market collapse.
This illustrates a broader facet of human nature: the tendency to be captivated by new and shiny developments. More recently, we can observe similar patterns with technological innovations. For instance, during the dot-com bubble at the turn of the millennium, companies could benefit merely by associating themselves with the internet without fundamentally changing their operations. Research has documented cases in which companies rebranded themselves as “.com” entities, even without an online presence, and stock prices reacted abnormally, increasing by roughly 75 percent within 10 days of the cosmetic rebranding. Of course, many of these companies eventually failed.
With the rise of fintech, we have seen comparable phenomena with blockchain and cryptocurrency. While these innovations have been instrumental in advancing financial systems and improving operational efficiency, the novelty of blockchain in 2017 sparked a wave of fascination. Many companies engaged in cosmetic rebranding to associate themselves with blockchain, often with minimal substantive adoption.
One particularly striking example is the Long Island Iced Tea Corporation, which renamed itself Long Blockchain Corporation, announcing plans to explore blockchain in its supply chain. The market reacted strongly, with the share price rising sharply, before later correcting and drawing regulatory scrutiny from the SEC for misleading statements.
In more recent years this pattern has reappeared in the form of what is now commonly called AI washing. This refers to situations in which firms that have long relied on quantitative techniques or relatively simple algorithms repackage those tools as artificial intelligence in order to project greater sophistication than is actually warranted. In practice, nothing fundamental has changed in how these models operate, yet the language used to describe them shifts dramatically to align with whatever technological label is currently in vogue.
There are many manifestations of this behavior, one of the most visible being corporate rebranding. Companies increasingly describe themselves as “AI first,” “AI driven,” “AI native,” or “data-driven,” often extending this language to machine learning and related buzzwords. Some firms go so far as to rename themselves or add AI related suffixes to their websites and online presence.
These strategies are especially common among small cap firms and speculative technology companies, where such signaling can generate an immediate reaction in the market, including short term jumps in share prices or spikes in trading volume. The phenomenon is even more pronounced in private equity and venture backed startups, where firms position themselves as the next Nvidia, Alphabet, or OpenAI and make sweeping claims about their technological capabilities. In these settings, verification is far more difficult because private companies face fewer disclosure requirements, making it easier for exaggerated narratives to persist.
A similar dynamic is evident in public company earnings calls, where managers increasingly saturate their communications with AI related language when discussing quarterly results. Traditional statistical or econometric models are reframed as AI systems, and references to artificial intelligence appear repeatedly throughout these discussions. Textual analyses of earnings call transcripts suggest that heavy use of such buzzwords can partially offset weak fundamentals in the eyes of investors, which is a troubling development. Even when performance indicators are mediocre, the rhetorical emphasis on AI appears to sustain enthusiasm and market attention.
Regulators have not been entirely silent on this issue. The former SEC chair Gary Gensler warned explicitly about AI washing, and subsequent SEC leadership has also issued statements cautioning against misleading claims.
Nonetheless, regulatory bodies remain behind the curve when it comes to effectively monitoring and enforcing standards in this area. Policing the boundary between legitimate technological innovation and opportunistic relabeling is difficult, and enforcement mechanisms have yet to fully catch up with the speed and creativity of market participants.
These practices extend beyond operating companies to the asset management industry as well. Mutual funds and other investment vehicles increasingly market themselves as AI driven, particularly in areas such as portfolio construction and risk management, often overstating the role that artificial intelligence actually plays in their decision making processes. Taken together, these examples suggest that while AI currently dominates the technological imagination, the underlying dynamics are familiar.
The Gartner Hype Cycle is a model widely used by management consultants, CTOs, and technology enthusiasts to map a technology’s journey from obscurity to maturity. At its core, it builds on the concept of the technology S-curve, which shows that any technology - whether overhyped or not - begins with minimal public attention because most people are unaware of it. Over time, awareness grows, visibility increases, and the technology either matures or is eventually replaced by something better. Attention initially rises slowly, accelerates as adoption grows, and eventually levels off, forming the characteristic S shape.
Though, technology adoption is not purely mechanical; human psychology plays a central role. Investors, like all humans, respond emotionally to new and shiny developments, and that response tends to follow an inverted U pattern: initial excitement rises to a peak and then gradually wanes.
Overlaying this human element onto the S-curve gives us the Gartner Hype Cycle: a technology begins in obscurity, surges to a peak of expectations, dips into a trough of disillusionment, and then climbs again toward steady growth before finally plateauing. The cycle unfolds in five stages: the innovation trigger, the peak of expectations, the trough of disillusionment, the slope of growth, and the plateau of productivity.
For example, in 2017, 3D printing was all the rage and widely hailed as a revolutionary technology. Many companies jumped on the hype train, but it only took a few years for 3D printing to move through the five phases of the Gartner Hype Cycle. Today, the technology has plateaued, and most investors no longer get excited about it. You can buy a 3D printer on Amazon for a couple of hundred dollars; it is no longer seen as cutting-edge like quantum computing.
What our research shows is that stock prices of companies adopting these hyped technologies follow a very similar pattern. When a company claims to embrace a technology undergoing the five phases, its stock prices tend to mirror those phases. This was the first time robust evidence demonstrated that stock prices track the Hype Cycle phases regardless of the specific technology.
Investors, of course, are always on the lookout for hidden gems, companies that might become the next Amazon or OpenAI. However, they often misinterpret signals and get played by companies engaging in “technology washing,” or overstating their adoption of emerging tech.
This is particularly common among small, lesser-known companies outside major technology hubs: for instance, a small AI startup in South Dakota claiming it has developed a new large language model that will surpass OpenAI within months. Such bold claims attract investor attention, producing abnormal market activity, rapid price increases, and subsequent declines. Technologies that move quickly through the Hype Cycle tend to attract even more investor interest.
Take AI, for example. In just a few years, we have seen multiple generations of ChatGPT, from the very first version to the current fifth iteration. Investors get very excited by this rapid evolution.
By contrast, some technologies are “slow burners.” Quantum computing, for instance, requires highly sophisticated physics and will take much longer to go mainstream. Most investors, seeking returns within one or two years, lack the patience for such long-term bets.
Interestingly, investors are often drawn to the more intangible or cryptic technologies. The more difficult a technology is to understand, the more alluring it becomes, because people assume it is the “missing piece” in their portfolio. Early blockchain investors, for instance, often could not explain how blockchain works or the role of cryptography, but the opacity made it more appealing. By contrast, tangible technologies, like debit or credit cards, are easier to understand and therefore feel less exotic or special. Emerging financial technologies like blockchain, NFTs, and AI benefit from this psychological effect.
To summarize, there is a strong interplay between investor psychology and technology adoption. Our research indicates significant AI washing is happening today. We have developed a framework for detecting and quantifying these cases and plan to engage with regulators to explore potential enforcement regimes to curb misleading claims.
Aiden Singh: Where do you think we currently are in the AI hype cycle?
Arman Eshraghi: If you had asked me six months ago, I would have said we were probably at the peak of inflated expectations. Since then, however, there have been more cautionary statements regarding the hype itself. For example, a simple keyword search for “AI hype” shows that the term is being mentioned more frequently in the financial press.
There is clearly more caution now compared to six months or a year ago, but it still pales in comparison to the overall level of enthusiasm and overreach. I would therefore say that we are still in the hype phase, albeit less intensely than a year ago.
This is not to suggest that the technology is not fascinating or, in some ways, revolutionary. AI has already transformed aspects of daily life, and I personally use many of these algorithms in my academic work, where they have been quite impactful.
My concern, from a finance perspective, is that many companies are rebranding themselves as AI-driven without meaningful machine learning actually occurring. These labels can be misleading, as much of the underlying code has existed for decades and does not truly “learn” in the AI sense.
The prevalence of this kind of AI washing indicates that significant hype remains. Where this will ultimately go is difficult to predict, much like riding a bubble, as the outcome is unknown until the episode has run its course.
I do expect the technology itself to continue advancing rapidly. Some constraints may emerge, particularly physical ones such as energy, given that AI is highly energy intensive. Countries are investing heavily in energy infrastructure in response.
At the same time, the human psychology component suggests that the hype itself is unlikely to dissipate anytime soon. As people witness more revolutionary applications of AI, the enthusiasm and accompanying hype will continue.
- - - - - - -
Cryptocurrencies
Aiden Singh: Continuing the discussion on hype cycles, let’s talk about cryptocurrencies. Given that they’ve undergone repeated cycles of euphoria and downturns, do you think cryptocurrencies possess intrinsic value?
Arman Eshraghi: That is a very difficult question to answer. If you think about traditional currencies - for example, the dollar in your pocket - does it have intrinsic value? In a classic sense, it does not, because it is no longer backed by physical assets as it once was. Yet in another sense, it does, because it is backed by the United States government, its power to tax, and its economic and military strength.
Cryptocurrencies are similar in that their value depends on context and perception. Take Bitcoin, the largest cryptocurrency: in the traditional economic sense, it does not have intrinsic value. However, it is widely used both as a speculative asset and, to some extent, as a store of value, which gives it a form of practical value. Some argue that its value also derives from the computing power and infrastructure that support the Bitcoin blockchain.
In short, cryptocurrencies like Bitcoin are assets or currencies without intrinsic value in the classical sense, but the market has assigned them value. Historically, despite volatility, the price of Bitcoin has generally trended upward over the past decade, and in recent years it has stabilized around certain levels. This stability suggests that the market perceives value in the innovation and that people continue to buy and hold it.
In addition, in recent years we have seen the corporate sector engage with Bitcoin. They are buying Bitcoin tokens to diversify corporate assets or to leverage the underlying technology. Based on this, I would not predict an imminent crash. Many have predicted crashes over the years, and those predictions have proven incorrect.
Ultimately, it is very difficult to determine a precise economic “intrinsic value” for Bitcoin. Its worth is largely determined by market perceptions and utility rather than classical economic fundamentals.
Aiden Singh: Are there any valuation techniques we use for traditional assets that could be applied to cryptocurrencies, or is it essentially impossible to assign them a value?
Researchers have attempted to apply models used in valuing equities and stocks to cryptocurrencies, but with only limited success. Fundamentally, cryptocurrencies are a different kind of asset, and in some cases it may not even be appropriate to use traditional valuation models.
Part of the finance literature approaches cryptocurrencies in an asset-pricing context, treating them like other assets where one can create risk factors and attempt to estimate value. However, this approach is hotly debated and controversial. In short, while there have been numerous attempts to value cryptocurrencies using traditional methods, there is no universally accepted formula or model that produces a definitive figure. That is the state of the field today.
Aiden Singh: Some advocates describe cryptocurrencies as digital currencies, while others have referred to them as digital gold. Do you think either of these descriptions is accurate?
Arman Eshraghi: They are partly correct, in the sense that cryptocurrencies are clearly digital assets.
Though, central bank digital currencies are also digital, which immediately raises the question of what actually makes something a currency. Traditionally, a currency is something you can use in everyday transactions such as buying groceries, paying for services, and settling routine payments. At least for now, you generally cannot use Bitcoin at the supermarket, which makes its status as a currency more debatable. This is one reason why many regulators prefer the term “crypto assets” rather than cryptocurrencies.
The comparison between cryptocurrencies and gold has also been widely debated. Gold has a long history as a reliable safe-haven asset. It has been used as an inflation hedge and as a store of value during recessions and periods of economic stress. In that sense, gold has a much stronger and more established claim to being a safe haven. Cryptocurrencies, by contrast, have displayed safe-haven characteristics in some episodes but not in others, which makes their claim considerably weaker and far less consistent.
Some critics argue that gold itself has no intrinsic value, but that argument is not entirely convincing. Gold does have practical applications: it is used in electronics, dentistry, and other industrial contexts, and it is an excellent conductor. While these uses may not fully explain its market value, they do provide at least some intrinsic foundation. The same cannot be said for cryptocurrencies, which lack comparable non-financial applications.
Ultimately, labels like “currency,” “digital gold,” “safe haven” are useful as conversation starters, but that is largely where their usefulness ends. Cryptocurrencies are a distinct category of assets. They share certain features with gold, with other commodities, and with fiat currencies, but they are also fundamentally different from all of them in important respects. Understanding cryptocurrencies requires recognizing both these partial similarities and the ways in which they do not neatly fit into existing asset classes.
Aiden Singh: What do we know about the different types of traders who participate in Bitcoin markets, how they behave, and how their behavior impacts the market?
Arman Eshraghi: Typically, cryptocurrency traders tend to be more short term and speculative. We know they have a higher risk appetite and are more willing to take risks, and they are also much more strongly driven by what they themselves describe as FOMO - the fear of missing out. Many individuals buy or trade cryptocurrencies simply because they know someone who has done well and they do not want to be left behind.
While FOMO also exists in equity markets, it appears to be significantly more potent in the cryptocurrency space. Part of the reason is that equity markets offer a vast universe of stocks, which makes participation more diffuse. In contrast, cryptocurrency markets are often dominated by a small number of highly visible tokens, typically the largest and most talked-about ones. That makes the decision to enter the market feel simpler and more immediate.
Cryptocurrency traders also tend to be younger and more technologically inclined, or at least less intimidated by technology. There is often a perception that familiarity with digital tools gives them an edge, or at least lowers the psychological barrier to entry.
In one of our papers, we analyze the trading profiles of cryptocurrency traders and classify them into different groups based on their degree of speculation and risk-taking. One of our core conclusions is that cryptocurrency markets are fundamentally different from equity markets, and the behavioral patterns we observe are distinct as well.
For example, pump-and-dump schemes are far more prevalent in cryptocurrency markets than in traditional equity markets. Especially among smaller or alternative tokens, we frequently see assets created essentially out of thin air, aggressively promoted by their creators, and then sold off to relatively unsophisticated investors. This pattern repeats itself, in part because the cryptocurrency space has historically faced less regulatory scrutiny. In that sense, it continues to resemble a financial ‘wild west,’ where these dynamics persist largely unchecked.
- - - - - - -
How Fund Managers Make Decisions
Aiden Singh: Let’s shift to professional investors, fund managers in particular. You interviewed 51 fund managers about their investment decision-making processes and found that both conventional and behavioral finance fall short of capturing how these decisions are actually made. What did you uncover?
Arman Eshraghi: This is joint work with two of my collaborators. We analyzed interviews and transcripts with many fund managers from around the world, delving deeply into their decision-making processes. Often, the conversations focused on how they manage portfolios worth hundreds of millions - and in some cases, billions of dollars - how they decide what to buy and sell, the timing of these decisions, and how they cope with the anxiety of managing other people’s money.
One key takeaway is that traditional finance models do not account for the strong emotions involved in managing money. Behavioral finance models, while an improvement, often reduce these emotions to basic labels such as fear and greed. Our conclusion was that this is far too simplistic. There is a much wider range of emotions triggered by investing on behalf of others.
Fund managers experience panic and anxiety over missing targets, underperforming benchmarks, facing investor redemptions, and navigating competition with other managers.
They must engage directly with companies to understand what is really happening, and trust emerges as a central theme. Even if a company presents solid financial statements, fund managers, particularly fundamental managers, emphasize the importance of human-to-human trust, looking the CEO in the eye to gauge credibility. Quantitative fund managers, who primarily rely on algorithms, also highlight trust, though in different ways.
Conviction is another critical theme. Fund managers must believe they can outperform the market, even though, statistically, most cannot consistently do so. While exceptions like Warren Buffett exist, they are rare. Yet to charge fees and attract investors, fund managers must maintain confidence in their ability to succeed despite the odds.
This creates a fascinating psychological tension. Fund managers must hold two contradictory notions simultaneously, acknowledging that on average they cannot outperform the market while convincingly presenting themselves as capable of doing so. Understanding how fund managers navigate these conflicting emotions was a key insight from our research.
Aiden Singh: Do you think fund managers are fully aware that, on average, they cannot consistently outperform the market, or is there a degree of cognitive dissonance? Might they genuinely believe that they can beat the market, despite the evidence to the contrary?
Arman Eshraghi: We think there is cognitive dissonance at play. On one level, these managers are very smart people. They read the studies, they read the academic literature, and the consensus in mainstream finance is that it is extremely difficult to consistently outperform the market. While the market is not perfectly efficient, most markets today are quite informationally efficient, and they have become even more efficient in recent years due to high-frequency and algorithmic trading. When arbitrage opportunities arise, quantitative managers with access to massive computing power and sophisticated algorithms tend to act faster and outperform traditional stock pickers.
These managers are not unsophisticated; they know this. Yet, what we observe is that they are very adept at managing this cognitive dissonance. They park the knowledge at the back of their minds and cultivate a sense of self-belief that they can outperform the market, even though deep down they know they cannot. They are skilled at holding these two realities together and managing the anxiety that it generates. In short, on one level, they believe they can beat the market, but if one digs deeper, they likely understand that, in reality, they cannot.
- - - - - - -
The Influence of Sleep on Financial Decisions
Aiden Singh: We’ve talked about market hype, the short-term trading behavior of bitcoin traders, and the cognitive dissonance of fund managers.
You’ve also found some interesting results regarding how sleep can affect traders and financial markets. Could you tell us a little more about that research and what you found?
Arman Eshraghi: That is another example of how human psychology affects financial decisions. If we were machines, none of these factors would matter. In reality, however, financial decisions are influenced not only by economic factors but also by the conditions under which they are made. For example, research has shown that environmental cues, such as listening to happy music, can make investors more bullish and buy more than they sell.
Our recent research has focused on the impact of sleep on financial decisions. On days when people have had poor or insufficient sleep, they tend to sell more stocks than they buy. The reason is that selling decisions are generally faster, whereas buying decisions are slower and more cognitively demanding. There is a classic paper called “Selling Fast and Buying Slow,” a play on Daniel Kahneman’s Thinking Fast and Slow, that documents this pattern.
Buying requires evaluating a broad set of options: choosing a sector, an asset class, and then selecting specific stocks. For fundamental investors, it may also involve examining company fundamentals and conducting valuations.
Selling, by contrast, is typically a quick comparison between the current market price and the entry price, often influenced by whether an investor is sitting on a paper gain or loss. Selling is therefore a more intuitive decision.
Our study was partly inspired by personal experience. My colleagues and I noticed that after staying up late watching streaming shows, we traded differently the next morning, selling more than buying. This led us to ask whether we could measure the impact of sleep quality on financial markets in a systematic way.
We used Netflix as a case study. Netflix frequently releases highly popular series at midnight, often causing dedicated viewers to lose two to three hours of sleep. We observed that on the mornings following these releases, the market tended to dip slightly. After controlling for alternative explanations, this appeared to be a clear manifestation of sleep affecting financial decisions at a broad market level.
Previous research has also documented this effect, for example, around daylight saving time changes. When clocks move forward and people lose an hour of sleep, the market dips the following day. The limitation of such studies, however, is that they provide only one observation per year, whereas our approach leverages many instances throughout the year, making the effect more robust.
These findings reflect my broader interest in behavioral finance and how human psychology interacts with economic behavior.
——————
Editing by Harpreet Chohan.