Is AI the new dot com bubble?

Update January 2025: The Rise of Deepseek: Another Overhyped AI Bubble?

The AI landscape is once again at a crossroads, with the arrival of Deepseek reigniting discussions about whether the current wave of AI advancements is truly revolutionary or yet another iteration of the dot-com bubble—an industry booming on speculation rather than substantive value. While proponents of Deepseek tout it as a game-changer, its actual impact remains questionable, raising concerns that it might be more of a marketing marvel than a true breakthrough.

Deepseek’s Overstated Capabilities

Deepseek, a Chinese AI model positioning itself as an open-source alternative to Western models like OpenAI’s GPT, has been met with a mix of excitement and scepticism. While it claims superior performance and deeper contextual understanding, early user experiences suggest otherwise. The model struggles with consistency, often failing in areas where its competitors have already made significant strides. The technical refinements Deepseek boasts of are not necessarily revolutionary but incremental, making its real-world efficacy questionable.

The Open-Source Illusion: Not All That Glitters is Gold

One of Deepseek’s key selling points is its open-source nature. While the AI community generally champions open-source models for their transparency and customisation potential, Deepseek’s approach appears to be more of a strategic move than a genuine commitment to AI ethics and accessibility. Several reports indicate that while parts of the model are open, the training data and fine-tuning processes remain opaque. This raises serious concerns about the integrity of its claims and whether it truly differs from proprietary models that restrict user access under the guise of openness.

Furthermore, its reliance on Chinese data sources raises geopolitical and ethical questions about AI censorship, data privacy, and bias. If Deepseek is trained on data curated under government regulations, can it really claim to be neutral or objective? This is particularly relevant given the global AI arms race, where AI models are often reflections of their creators’ ideological and commercial interests.

The AI Bubble Redux?

Deepseek’s overhyped entry into the AI arena is reminiscent of the early 2000s dot-com bubble, where companies with little actual innovation rode a wave of exaggerated claims, only to collapse when the market corrected itself. Many AI models today—including Deepseek—suffer from the same problem: marketing-led hype rather than genuine technological breakthroughs.

Venture capitalists and AI evangelists continue to push the narrative that each new model represents an exponential leap forward, when in reality, most improvements are marginal at best. The obsession with releasing new AI models—many of which feel like repackaged versions of existing frameworks—only reinforces the idea that we are in a bubble. Like the dot-com bust, AI startups that lack real-world application or business sustainability will likely face a reckoning.

Deepseek’s Real Position in the AI Market

Rather than being an industry disruptor, Deepseek seems more like a regional alternative to existing AI solutions rather than an indispensable upgrade. Its functionality, so far, does not appear to outperform GPT-4 or Claude in significant ways, making it another example of AI incrementalism disguised as innovation.

While China’s AI ambitions are undoubtedly aggressive, Deepseek’s current state does not justify the hype surrounding it. If anything, it serves as yet another cautionary tale of how AI startups, regardless of their geographical origin, rely on speculative buzz to gain traction rather than delivering groundbreaking advancements.

Lesson in Critical AI Adoption

As AI continues to evolve, it is crucial to separate real innovation from marketing-driven illusions. Deepseek, like many AI models before it, thrives on the illusion of disruption. However, its actual impact on AI progress remains debatable. Unless it can offer something truly transformative—beyond the standard open-source rhetoric and overblown performance claims—it is unlikely to stand the test of time.

For businesses and individuals investing in AI, the lesson remains the same: hype does not equal value. Deepseek may be the latest AI model to capture headlines, but whether it can genuinely reshape the industry—or if it will merely be another name in the ever-growing list of AI experiments that failed to deliver—remains to be seen.

______________________________________________________________________________

Original Article follows below

The comparison between the current AI boom and the dot-com bubble of the late 1990s is compelling, but it demands nuance. Both eras showcase surging valuations and a gold rush mentality. In the dot-com bubble, speculative investments in unproven startups drove irrational exuberance, leading to catastrophic corrections. Similarly, AI companies like OpenAI now command staggering valuations, often with limited profitability, echoing that speculative zeal.

However, the fundamental differences are stark. Unlike the dot-com era, the AI sector is bolstered by established players such as Microsoft and Nvidia, whose diversified revenue streams and tangible applications—ranging from healthcare to autonomous vehicles—offer a more stable foundation. AI technologies have already proven transformative in multiple industries, distinguishing the current wave from the speculative promises of nascent internet firms.

Nonetheless, skepticism persists. Analysts like MIT economist Daron Acemoglu warn that AI’s impact on the economy may be overstated, cautioning against potential market corrections akin to the early 2000s crash. Conversely, institutions like Goldman Sachs argue that AI valuations are underpinned by strong fundamentals, countering the “bubble” narrative.

In January 2025, the truth likely lies somewhere in between. The parallels to the dot-com bubble are instructive, but the presence of established market leaders and demonstrated utility make this a more grounded—albeit still risky—sector. The prudent investor will remain cautious, acknowledging both the transformative potential of AI and the volatility that comes with any rapidly expanding industry.

Recent developments in the AI industry both support and challenge the idea of an AI bubble akin to the dot-com bubble of the late ’90s. On one hand, there are growing concerns that the rapid investment in AI may not be sustainable, with some experts suggesting a market correction is on the horizon. For example, Baidu’s CEO Robin Li predicts that only a small percentage of AI startups will survive the coming shakeout, comparing the current climate to the dot-com bubble where many companies were overvalued and lacked clear paths to profitability. He believes the AI industry will experience a “necessary process” of consolidation, much like the dot-com bust, where only the most innovative and profitable companies will endure.

Similarly, venture capitalists have observed that many AI startups are receiving substantial funding despite lacking customers, revenue, or clear market demand. This trend is reminiscent of the speculative investments during the dot-com era, where companies were funded based on future potential rather than proven business models. There are already signs of waning hype, with concerns over the ability of AI companies to generate profits despite large investments in infrastructure and technology.

On the other hand, the advancements in AI, particularly in large language models (LLMs) and practical applications in industries such as healthcare, logistics, and customer service, suggest that the current wave of AI development has more tangible benefits than many dot-com companies did at their peak. AI technologies have demonstrated the potential to significantly enhance efficiency and productivity in a way that early Internet companies could not achieve in the ’90s. This real-world utility may provide a more solid foundation for AI than what existed for many companies during the dot-com boom.

Ultimately, while the AI industry may face a correction due to overvaluation and market saturation, its technological maturity and real-world applications differentiate it from the dot-com bubble. However, the parallels in speculative investment and valuation inflation indicate that caution is warranted.