In a striking echo of past technological frenzies, the artificial intelligence sector is hurtling towards a predictable climax. The parallels with the dot-com bubble of the late 1990s are uncanny, prompting urgent reflections on what comes next. When speculative bubbles burst, they often leave devastation in their wake, but they also create fertile ground for renewal and improvement. The critical question is whether we will seize this opportunity to build a more equitable and responsible technological landscape.
A Familiar Pattern of Speculation and Concentration
Rewind to December 1999, a time when tech investors were intoxicated by the promise of rapid riches. Back then, a simple website and a flashy Super Bowl advertisement were mistakenly seen as foolproof tickets to wealth. This era was characterised by a dangerous conflation of spending with genuine growth and marketing with sustainable business models. The inevitable crash followed swiftly, wiping out a staggering $1.7 trillion in market value and inflicting a $5 trillion blow on the broader economy.
Fast forward to today, and history appears to be repeating itself with artificial intelligence at the centre. The current AI boom exhibits eerily similar traits. A concerning 80% of stock gains in 2025 are concentrated in just seven corporate giants: Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla. These behemoths are engaged in a fierce battle for dominance over the entire AI stack, encompassing hardware, software, data, energy, and critical infrastructure.
This isn't merely a contest for market share; it's a struggle over who will shape how billions of people across the globe learn, create, and perceive the world. Such extreme concentration of power should alarm everyone, as it risks stifling innovation and centralising control in the hands of a few.
The Flawed Economic Logic Behind the AI Frenzy
Much like the dot-com era, today's AI valuations are soaring without clear, sustainable paths to profitability. Many companies are peddling the fantasy that AI will seamlessly replace human workers, despite evidence showing that 95% of AI experiments within firms fail to reach production. Instead of focusing on developing public-interest tools that genuinely expand human potential, a significant portion of the industry is generating what author Cory Doctorow terms "productive residue"—a deluge of synthetic media, misinformation, and deepfakes.
The core issue isn't artificial intelligence itself; it's the prevailing economic model driving its development. This model treats technology as an extractive industry, hoarding vast amounts of data, consolidating corporate power, and externalising harm to society. The so-called AI arms race is motivated less by genuine innovation and more by a desire for domination, prioritising profit over human welfare and ethical considerations.
The Viable Alternative: Open-Source and Mission-Driven AI
Fortunately, a different and more sustainable economic model already exists and is gaining traction. Around the world, a growing community of open-source developers and mission-driven companies is diligently building shared infrastructure for trustworthy AI. Their work focuses on creating systems that are transparent, auditable, and adaptable to local needs, proving that groundbreaking innovation does not require monopolistic control over data.
This movement is exemplified by pioneering companies that blend strong values with commercial competitiveness. Hugging Face, for instance, operates the world's most widely used open-source machine-learning model and dataset hub. Flower AI enables decentralised, federated learning, challenging the dominance of centralised big models. Oumi offers a fully open-source platform for building and deploying custom AI models on local infrastructure, moving away from reliance on closed cloud systems.
These initiatives are not speculative gambles; they represent the seeds of a more sustainable and pluralistic technology ecosystem. They embody a double-bottom-line economic model for tech—one that values both mission and money, aiming to create public good alongside financial returns.
Learning from History: The Post-Bubble Opportunity
If history serves as a reliable guide, the current AI frenzy will likely conclude with a significant market correction, similar to the dot-com crash. However, this potential collapse should not be viewed as an endpoint but rather as the beginning of a new chapter. In the aftermath of the last bubble, open-source building blocks like the Linux stack rose from the ashes to challenge and ultimately surpass proprietary systems like Windows.
Research indicates that such open-source components have generated an astonishing $8.8 trillion in value over the past two decades. New studies suggest that startups and other businesses could unlock tens of billions in additional value by switching from closed AI platforms to open-source models. The potential for creating immense value through open, collaborative approaches is truly enormous.
When the AI bubble eventually pops, society will face a critical choice. We can either rebuild the same monopolistic, extractive model, or we can use this pivotal moment to design an economy that is pro-human and values-driven. This means championing open models, advocating for transparent governance, and ensuring equitable participation in the value that AI creates.
Envisioning a Human-Centric AI Future
It also means refocusing on what people genuinely desire from technology: robust privacy, strong security, personal agency, and simple joy. The true promise of AI lies not in its infinite scale but in its ability to make our lives easier, richer, and more creative without sacrificing individual choice or human dignity.
This vision is already becoming a reality in some quarters. Experiments with privacy-protecting, open-source models for applications like browser and email assistants are showing promising improvements. Imagine a future where individuals and communities can host small, local AI models that are energy-efficient, privacy-preserving, and tailored to their specific needs. A future where developers build tools collaboratively rather than competitively, and where innovation is measured by public good rather than mere market share.
This is not a utopian fantasy. If we act now—by building AI that is open, transparent, and rooted in shared human values—we can ensure that the next era of technology expands human freedom instead of constraining it. The dot-com crash ultimately gave us the modern, more open web. The next correction could gift us an even better technological landscape, provided we have the courage to fundamentally rethink the economics of innovation.
In the end, the choice rests with us. We can passively allow a handful of powerful corporations to own and dictate our collective future. Or, we can take an active role in shaping what we build, embracing collaboration and shared ownership to create a technology ecosystem that benefits all of humanity.
