The AI Boom: Innovation, Investment, and the Risk of a Bubble

August 29, 2025

Artificial intelligence (AI) has become the defining technological story of the 2020s. Generative models like ChatGPT, MidJourney, and Claude have captured the public imagination, fueled multi-billion-dollar corporate investment, and triggered widespread debate over the future of work, creativity, and economic growth.

But with climbing valuations and increasingly grandiose narratives, a central question emerges: are we in the middle of a speculative bubble similar to the dot-com boom of the late 1990s, or is AI a transformative revolution ? We look at the parallels, explore how AI might be different, and outline how Runbox is approaching AI in the midst of this development.

Blue balloon. Man hold needle directed to air balloon. Concept of risk.

The Runbox approach

At Runbox, we know there is a lot excitement and expectation around artificial intelligence. Tech companies are rapidly embedding AI into their email platforms – adding tools like smart replies, auto-drafting, and predictive search – to showcase innovation and keep users within their ecosystems.

While some of these advances are exciting, we also recognize that not every AI feature necessarily adds real value – and many can come at the cost of privacy, security, or simply overwhelming users with unnecessary complexity.

As a smaller, independent provider, our focus is different. We’re not chasing every new AI trend, but plan on using AI responsibly and in line with our Generative AI Policy, the GDPR and EU’s AI act.

Our preference will always be for open source software, genuine service, sustainability and human centered support. That means when you have a question or run into an issue, you’ll get a real person on the other end, not a chatbot.

Hype as a market force

The current AI boom rests on genuine technological progress. Advances in deep learning, access to massive datasets, and the rise of GPU-accelerated computing have made possible tools that rival human output in language, imagery, and even coding. According to the IDC, the overall global economy could see a boost of $19.9 trillion by 2030, or an increase of 3.5% to the global GDP – all due to the widespread adoption and use of artificial intelligence.

Graph showing AI percentage of total Venture capital funding in the us by statista

But market enthusiasm often runs ahead of fundamentals. Global venture capital investment in AI hit $49.2 billion in the first half of 2025, surpassing the total of $44.2 billion in 2024 and more than double the total ($21.3 billion) for 2023. This is a level of capital concentration seen in few other industries. Meanwhile, many of these startups remain pre-revenue, or rely on costly cloud computing without a clear path to profitability.

Source: Statista

The Dot-Com parallel

The dot-com era offers a cautionary tale. Between 1995 and 2000, internet-related companies experienced explosive growth. The NASDAQ index rose nearly 400%, and firms with little more than a website could secure multi-million-dollar valuations. By 2002, the bubble burst, wiping out $5 trillion in market value.

Despite the collapse, the internet itself did not disappear – instead, it matured. Out of that wreckage emerged Amazon, Google, and eBay, companies that built sustainable models and reshaped the global economy. The lesson: hype can destroy value in the short term, but the underlying technology will continue to march forward.

AI today mirrors this dynamic. Startups such as Anthropic, Inflection, and OpenAI have reached multi-billion-dollar valuations at record speed, echoing the trajectory of some dot-coms. Yet, as with the internet, the lasting winners may only become clear after a market correction.

The rise and fall of dot-com:

Internet bubble
Source: Quartz

Signs of a potential AI bubble

Analysts point to several bubble-like dynamics in the AI sector:

  1. Venture Capital Frenzy – In the first half of this year, more than half of all global venture funding in 2025 went to AI startups, and nearly two-thirds of funding in the US, leaving other sectors comparatively underfunded.
  2. Concentrated Valuations – Nvidia, the primary supplier of AI chips, saw its market value reach almost $4 trillion in July 2025, making it the world’s most valuable company. This concentration of value in hardware echoes Cisco’s dominance during the dot-com era.
  3. Speculative Narratives – Predictions of imminent “artificial general intelligence” echo the dot-com promise of limitless “new economies” that failed to materialize on the expected timeline.

Why AI might be different

Despite the parallels, AI may avoid the fate of being a pure bubble for several reasons:

  • Immediate Use: Unlike many dot-com startups, AI already delivers tangible value, such as automating customer service, optimizing logistics, accelerating drug discovery and diagnostic accuracy.
  • Infrastructure: The heavy capital investment in GPUs, data centers, and cloud infrastructure suggests deeper integration into the economy signaling a long-term commitment and growing reliance on AI technologies.
  • Cross-Sector Relevance: AI is not confined to one industry; it cuts across healthcare, finance, education, and creative sectors, shaping innovation and efficiency in diverse parts of the economy – from assisting doctors with medical imaging to helping banks detect fraud in real time.
Ibm 360 mainframe
IBM System/360

AI’s roots in machine learning

What we call artificial intelligence today didn’t appear overnight. It is the product of decades of steady progress in machine learning, tracing back to the mid-20th century. In the 1950s and 60s, researchers experimented with symbolic reasoning, early neural networks, and statistical approaches to pattern recognition. Though limited by the computing power of the time, these pioneering efforts laid the groundwork for what was to come.

The 1970s and 80s saw expert systems find use in fields like medicine and finance but proved limited, driving interest in more adaptable methods. With better algorithms and more data, neural networks gained renewed attention, and by the 1990s machine learning was powering speech and handwriting recognition as well as early recommender systems.

The innovations we see today are built on decades of research and experimentation, combined with transformative advances in hardware technology such as graphics processing units (GPUs), tensor processing units (TPUs), and large-scale cloud computing. These tools made it possible to train deep learning models on massive datasets, moving forward from theoretical ideas to the powerful AI applications that shape industries today.

The Path Forward

History suggests two things can be true at once: valuations may outpace reality in the short run, while the technology itself can prove transformative over decades. The internet bubble destroyed trillions in paper wealth but ultimately gave rise to the digital economy that defines the 21st century.

Is AI on a similar path? A correction may be inevitable, particularly for firms chasing hype without sustainable models. But the technology is not a fad. Instead, it is a platform shift – the long-term impact of AI on the global economy is likely to be profound. The enduring companies of the AI era may not be the loudest startups of today, but those building resilient infrastructure, safe governance frameworks, and business models grounded in real-world utility.

Want to read more about what’s happening in the world of AI? Here are some useful resources:

9 industries being revolutionized by AI

AGI Forecasts: Crystal Balls, Moving Goalposts, and the AI Experts Who Love Them

Long term implications of AI on global economic equilibrium

The global impact of AI: Mind the gap

Artificial intelligence H1 2025 global report

Stanford The 2025 AI Index Report

What do people expect from Artificial Intelligence?

More from the blog: