We are currently witnessing the rapid development of generative artificial intelligence (AI) technology. For example, large language models such as GPT and Claude have been developing rapidly since 2018. These models can generate text that resembles human speech, to visual generative AI (GenAI) such as DALL-E and Midjourney that are able to create images from text descriptions.

Other examples include AI models that can generate music, programming code, and even short videos .

This development then triggered a strong flow of capital flowing into AI startups . Venture capital funding in the United States for artificial intelligence startups increased significantly in the second quarter of 2024, reaching a total of US$55.6 billion or more than Rp905 trillion .

The figure includes the $6 billion raised by Elon Musk’s conglomerate xAI and the $1.1 billion raised by CoreWeave, signaling a strong increase in interest in the AI ​​sector.

However, many GenAI ventures are still in their early stages, with business models still figuring out their shape. Examples include AI startups focused on automated content creation or virtual assistants still figuring out how to generate stable revenue .

This then raised questions from analysts at several large institutions such as Gartner , Wall Street , and Goldman Sachs about the risk of AI becoming a trend in a bubble and then dissolving over time.

To answer this question, we can first look at two previous technology trends, namely dot-com and crypto. Both became bubbles because investor expectations turned out to be too high—exceeding the practical applications and reliability at the time.

The upside of dot-com market optimism

In the late 1990s to early 2000s, the world witnessed an important event: the dot-com bubble .

This era reflected excessive market optimism, fueled by the revolutionary potential of the World Wide Web. Alan Greenspan, in his term irrational exuberance , described how investors—enchanted by the Internet’s allure—pour huge sums of capital into fledgling Internet companies.

These startups, often without a proven business model or clear revenue stream, easily gained access to capital, primarily through Initial Public Offerings (IPOs). Some of the major companies that IPOed during the dot-com bubble period include BlackBerry, Broadcom Corporation, and Verisign .

The NASDAQ Composite Index, which is dominated by technology stocks, has risen more than 400% . It has soared from under a thousand points in 1995 to over 5,000 points in 2000, signaling a surge in investor interest.

Unfortunately, many of these companies had weak business foundations and were more focused on market share than profitability. In 2000, the dot-com bubble burst.

The dot-com bust was marked by a serious mismatch between inflated valuations and actual financial health. The Bloomberg US Internet Index plunged from $2.9 trillion (at current exchange rates equivalent to Rp47,000 trillion) to $1.1 trillion (Rp17,900 trillion), more than halving.

The stark contrast in companies like Cisco and Yahoo! illustrates the volatility of tech investing and the risks of overvaluation. Both reached high valuations at the peak of the dot-com bubble, only to see steep declines thereafter.

Cisco managed to survive and stay relevant, while Yahoo! eventually lost its market dominance .

The bursting of the bubble had a wide-ranging impact on the economy , including significant job losses, particularly in the telecommunications sector.

Despite the chaos, the dot-com bust had long-term benefits. The bursting of the bubble paved the way for a more realistic understanding of the potential of Internet business.

This era also became a filter that allowed companies with sustainable business models to survive in the technology industry. This historical episode emphasized the risks of trend-based overvaluation and highlighted the importance of strong business fundamentals.

The dot-com bubble is also an important lesson for emerging technology sectors, such as GenAI.

Crypto bubble: digital assets and speculation fever

The crypto bubble emerged as an alternative form of the dot-com bubble that emerged from a fever of speculation.

Unlike the dot-com bubble, the crypto bubble has different characteristics and impacts. The crypto bubble was driven by the innovative charm of blockchain technology and cryptocurrencies, resulting in an unprecedented surge in speculative investment.

One of the main drivers is the ‘ fear of missing out ‘ (FOMO) syndrome. Investors have been lured by stories of the surge in the value of cryptocurrencies, such as Bitcoin, which rose from around US$900 (Rp14.66 million) to almost US$20,000 (Rp325.87 million) in less than a year in 2017.

Those lured by the promise of extraordinary returns then invest their money in various digital assets and initial coin offerings (ICOs). Unfortunately, many investors ignore the risks and fluctuations in the value of these commodities.

However, the structure of the crypto market and its integration (or lack thereof) with the traditional financial system makes it different from the dot-com scenario. Unlike the dot-com crash, which had a broader impact on the global economy, the crypto crash had a more direct impact on digital assets. This impact is partly due to the relatively limited adoption of cryptocurrencies in mainstream finance when the bubble burst.

In addition, the regulatory landscape for cryptocurrencies is fraught with uncertainty. Unlike traditional securities that have well-established regulatory frameworks, cryptocurrencies fall into a regulatory grey area in some countries.

This creates additional risks for investors. For example, the lack of regulation makes it challenging to address issues such as fraud and market manipulation, which are common in the crypto space.

GenAI risks ballooning?

GenAI is an impressive trend because this technology is able to create new content—from text, images, to music—that previously could only be produced by humans.

Its ability to simplify production processes and provide automated solutions in a wide range of fields makes it very attractive to many industries, from marketing to art.

However, the quality and reliability of GenAI are still limited. These AI models require massive amounts of data and incredible computing power. This makes scalability—the ability of a technology to accommodate increased load—a major challenge.

In addition, legal and ethical issues such as intellectual property rights and algorithmic bias are also becoming more prominent. The risk of a global economic slowdown is also making investors more cautious, which has a direct impact on funding for AI startups.

Just like the dot-com bubble, the current GenAI trend may be heading in the same direction. Many AI companies are still operating with unproven business models and are heavily reliant on investor money.

For example, Microsoft and Adobe are struggling to profit from AI due to high investment . Microsoft loses an average of $20 per AI user, but the average cost is only $10. This makes investors somewhat skeptical and cautious about investing in the AI ​​market.

However, unlike the dot-com and crypto era, GenAI has shown significant real-world applications across sectors ranging from healthcare to finance .

The rapid advancement of AI technology and its diversified adoption across industries provide a more solid foundation, potentially reducing the risk of a bubble. For example, JP Morgan , one of the world’s largest investment banks, developed a GenAI-based conversational assistant to improve the efficiency of the company’s financial data analysis.

However, we cannot ignore the fact that GenAI is still in its early stages of development. There is still a lot of uncertainty about the technology’s ability to generate long-term profits.

Additionally, increasingly stringent regulations and environmental concerns about the energy requirements of AI data centers could be factors limiting the growth of this sector.

Famed British investor Jeremy Grantham, who predicted the dot-com bubble and the 2008 financial crisis, says the AI ​​bubble could “pop,” but the impact may not be as bad as the internet bubble.

Even if the AI ​​bubble bursts, Jeremy says, the technology’s intrinsic value will likely persist because of its broad and significant practical value.

The key to avoiding bubbles in the GenAI sector lies in balancing investment, speculation, and a realistic assessment of the technology’s current capabilities. The potential for sustainable practical applications should also be a consideration for investors.

The AI ​​market may be corrected, similar to past tech bubbles. However, AI technology has provided its value and utility. The foundation of today’s tech companies is also stronger than during the dot-com bubble .

Once corrected, AI is likely to remain a vital and integral part of the technology landscape. The sector has the potential to continue to grow and integrate into various aspects of business and society.

Lessons from the dot-com and crypto bubbles provide valuable guidance in navigating this new frontier. That’s why it’s imperative to invest wisely, have realistic expectations, and focus on sustainable growth and application.

AI companies and investors also need to ensure ethical and appropriate use, and find sustainable paths to maintain revenue levels. This is not only to avoid financial losses, but also to encourage responsible and sustainable innovation in the future.

Author Bio: Arif Perdana is Associate Professor in Digital Strategy and Data Science at Monash University