#2. Is AI a Monopoly game ?
Unleashing endless possibilities, 20 May
Dear readers,
I am very glad to write to you for the first time after the “official” launch of my newsletter Singularity some days ago! As I said at this moment, I think that the effects of AI are largely underestimated, and we need to prepare the society and the “corporate” world to navigate in this new environment.
As I am French and European at the same time, the challenge seems to be higher for us, as Europe seems to be in a sort of mid-tech trap. In this respect, the mantra of Singularity is clear: BE AI-READY, and it will be possible by creating some bridges on both side of the Atlantic to improve things collectively.
Today, I am going to speak about a major concern for us European and at a lesser extent for American: the fact that AI can be monopolized or, even hijacked (for the most critical of us) by a small number of big companies, especially in the United States. In a sense, Microsoft, Nvidia and Meta can become a sort of superpowers in themselves.
This concentration of economic and financial power can even be totally new and unprecedented in the history, even if there have always been dominant players and sectors over time.
But, with AI, things are different. This technology seems to favor, more than any others, extremely rich players because AI requires a huge number of resources and can create a vast wealth in return if things work out. In this context, it’s not unthinkable to see some companies exceed $10,000 billion of market cap in the future.
For instance, almost 6 years ago, Apple became the first company to reach a value of one trillion of dollars. Today, it’s three times more. Another impressive figure: the market capitalization of Nvidia has increased of $1 trillion from its value at the beginning of the year. One reason: its superchips are the best ones for giant scale AI. At the end of her keynote speech for the launch of GPT-4o, Mira Murati, CTO of Open AI, thanked Jensen Huang (the CEO of Nvidia) for “bringing the most advanced GPUs possible” to demo its model.
In a sense, and for some observers like Tim Wu, the digital economy looks more and more like a Monopoly board game. Is it true? Is the trend exacerbating with the rapid expansion and the uncommon features of artificial intelligence? The answer is clearly not black and white. If the analogy can be insightful in several ways, it is not a perfect comparison.
Here are the three main similarities between AI market and Monopoly :
Market dominance
In Monopoly, the object is to become the wealthiest player by buying, selling, trading, and collecting rend on properties. In the AI market, major tech companies aim to dominate by developing advanced technologies and acquiring vast amounts of data. In Monopoly and AI, there is a resource accumulation, or some kind of predation said the most critical people. In a sense, tech industries, in AI especially, are prone to monopoly and monopolist behaviors.
Competitive strategy
Both in Monopoly and in the AI sector, strategic decisions play a crucial role. Companies must decide where to invest, which tech to develop and how to outmaneuver the most important competitors. Recently, for example, Amazon completed a multibillion dollar bet on Anthropic, an OpenAI rival. In a new controversy move, Microsoft recruited the main experts, like Mustafa Suleyman (the co-founder of DeepMind) of InflectionAI without buying the start-up !
Winner-takes-all
The AI industry, like Monopoly, often sees a “winner-takes-all” dynamic where the leading companies gain significant advantages and smaller competitors struggle to stay in the contest. “Competition is for losers. If you want to create and capture lasting value, look to build a monopoly”, wrote Peter Thiel
Here are the three main differences between AI market and Monopoly :
Innovation, Research & Development
Over time, resources can be less important to develop some very competitive models and can be an opportunity for small players to scale up. For instance, GPT-4o is less expensive to run than its previous version. In addition, some big companies can suffer some internal problems to maintain an innovation culture and keep their advance in terms of technology. “The more successful things are, the more risk-adverse people become”, said Google CEO Sundar Pichai in a recent interview.
In this view, a kind of cultural battle seems to emerge between incumbent (Google) and new player (OpenAI). The following Sam Altman’s post on X is indeed an illustration and it triggered a lot of debates and – sometimes -mockeries (OpenAI presentation space is compared as a McDonald’s or Starbuck interior design):
Regulation & Ethics
Unlike Monopoly governed with fixed rules, the AI industry is subject to evolving regulation and ethical consideration. As explained by Tim Wu in a recent great essay, the rules that govern tech competition may evolve in the future, notably regarding the battle over commercialized AI. It is interesting to consider the case of the breakup of AT&T’s telephone monopoly in 1984 for example.
Collaboration vs. competition
First, small companies can leverage open source LLMs. They can build on existing technology without substantial investment. Second, small companies can negotiate some very lucrative arrangements with big companies. It seems that Apple and OpenAI will announce a partnership next month to use the startup’s tech on the iPhone, maybe at the WWDC 2024, the annual Apple Worldwide Developer conference (10 – 14 June). More explanations can be found in a newsletter published by
. Other example: in France, Mistral AI signed recently a deal with Microsoft.
In the next letter, I will present some European startup (French ones especially) very active in the AI field, and I will also propose some advice for small companies to be AI-ready.
>READ THE PREVIOUS EDITION “How far AI will go?”
Let’s spread the word :
Amaury