“METEROnopoly,” said Peter Thiel, Silicon Valley’s answer to Darth Vader, “is the condition of every successful business.” This aspiration is widely shared by Gamman, the new acronym for the Valley giants: Google, Apple, Microsoft, Meta, Amazon and Nvidia. And the advent of AI has sharpened everyone’s appetite to reach that blessed state before others get there.
One symptom of their anxiety is the way they have been spending inordinate amounts of money on the 70 or so generative AI startups that have mushroomed since it became clear that AI was going to be the big thing. Microsoft is said to have invested $13bn (around £10.4bn) in OpenAI, for example, but was also the lead investor in a $1.3bn funding round for Deepmind co-founder’s startup Inflection, Mustafa Suleyman. Amazon invested $4 billion in Anthropic, the refugee-founded startup OpenAI. Google invested $500 million in the same company, with the promise of $1.5 billion more, and unspecified sums in A121 Labs and Hugging Face. (Yes, I know the names don’t make sense.) Microsoft has also invested in Mistral, the French AI startup. Etc. In 2023, of the $27 billion invested in AI startups, only $9 billion came from from venture capital companies – who until recently had been by far the largest funders of tech startups in Silicon Valley.
What’s going on here? After all, the tech giants have their own “fundamental” AI models and don’t need what the minnows have built or are building. And then the coin falls: We’ve seen this playbook before: Established companies spot potential competitors in their infancy and buy them up. Google bought YouTube in 2006, for example; Facebook bought Instagram for $1 billion in 2012, when it only had 13 employees, and then acquired WhatsApp in 2014 (for what seemed like an incredible sum at the time: $19 billion).
With the 20/20 vision of hindsight, we now realize that these were all anti-competitive acquisitions that should have been resisted at the time, but were not. That’s why it’s so comforting to know that there is at least one regulator – the UK Competition and Markets Authority (CMA) – that seems determined to learn from that history.
in a speech given to a meeting of American antitrust lawyers In Washington just over a week ago, CMA CEO Sarah Cardell announced that she was determined to ensure that markets for fundamental AI models were supported by fair, open and effective competition, as well as a strong consumer protection. Their concern was that the growing presence of a small number of giant companies along the AI value chain (the sequence of steps required to convert inputs into usable products) could shape these markets in a way that harms competition. and reduce choice and quality for businesses. and consumers.
It listed three key risks to competition: that companies that control critical inputs to developing fundamental models could restrict access to protect themselves from competition; that powerful incumbents could exploit their positions in consumer or business markets to restrict competition in model deployment and thus distort choice; and that associations between key actors could reinforce or expand existing market power throughout the value chain.
It also warned that the CMA would take steps through its formidable investigative powers (including merger control reviews, market investigations and potential appointments under new digital competition legislation) to assess and mitigate competition risks from the new technology.
Hearing a top regulator speak like this about the tech industry was truly extraordinary. Cardell was pointing out that rather than waiting for problems to arise before acting, the CMA would henceforth be proactive, striving to get ahead of the game rather than lag behind the tech moguls who believe – as the saying goes – in move fast and break things. The CMA, he said, was already preparing for the task based on what it has learned so far in dealing with technology platforms. His goal would be to take a holistic view of the entire value chain of AI model deployment, rather than focusing solely on individual parts of the chain. And it will use its merger review powers more aggressively to assess the competitive implications of AI partnerships and investments.
Stirring things, huh? But in some ways, it is normal for one of the few British institutions that seems capable of using post-Brexit freedoms as an opportunity for creativity and innovation. And any tycoon who might be tempted to dismiss Cardell’s speech as simply fiery rhetoric should reflect on the CMA’s recent record: its Solid investigation into Microsoft’s acquisition of Activision Blizzard, For example; or the way it forced Meta to ditch Giphy, an online database and search engine that allows users to search and share animated gif files. Cardell may keep a lower profile than Lina Khan, his counterpart at the FTC in the United States, but he is clearly serious too: beware of would-be monopolists.
after newsletter promotion
what i have been reading
Design of the times
Adam Tooze has written an extraordinary essay in his Substack about the famous Bauhaus design school, its historical legacy and the enigmatic career of one of its students, Franz Ehrlich.
The computer says go
Timothy Burke has written about the role of AI in productivity and project management in an elegant Substack essay which is based on an idea first articulated by Fred Brooks in his book The month of the mythical man.
Big data is watching
Henry Farrell meditates on a new literary genre, the thriller of apocalyptic systemsand its influence on nonfiction, in a blog post about his Substack Programmable Mutter.