Finally, there’s a challenger to Google. After Google saw off Altavista, then Yahoo! and Bing, and established itself as a monopoly, there really has not been very much challenge to what has often appeared an impassable position.
But then, of course, came generative AI, and a seeming revolution in both information retrieval and content creation occurred – particularly with the November 2022 launch of ChatGPT.
The frenzied signing up led to claims that this was the beginning of the end for Google, or perhaps that Google was already dead. As alternative systems arrive, the social media hype cycle moves into ‘ChatGPT killer’ mode as well. It’s understandable, they hype just gets more engagement, but is there really any truth in it? Is ChatGPT really a Google killer?
Google’s ‘weird bind’
The reality is at the moment Google is finding itself in a weird bind. Is it going too slow in AI releases (for users) or too fast (for regulators)? Open AI’s ChatGPT has been nipping away, eating up the LLM chatbot market to the tune of 80% and some people have called it a Google killer. But Google’s Gemini LLM is actually pretty good, and this is a long game.
The LLM fight is just the beginning of a new phase of information retrieval, content creation and the Internet itself. This is a seismic shift, but with all fundamental upgrades to the web, it’ll rely on accessibility to become ubiquitous:
- Web 1.0 – needed widespread broadband to become mainstream in the late 90s.
- Web 2.0 – needed smartphones and 3g+ connectivity.
- Web 3.0 – will require major chip and data centre infrastructure. The investments here are already happening, as indicated by Nvidia’s skyrocketing share price.
AI is partially the formulative layer of web 3.0. It is through AI that the spacial web and ‘Metaverse’ may be created. But even with AI, we’re early. It took essentially a decade from inception for social media to be widely used. In the early days there were some big casualties – MySpace and Friendster being cited by Mark Zuckerberg as eclipsed by Facebook, which now, as Meta, is essentially a monopoly as Google is in search. I don’t doubt that the AI wave will happen much faster than social media adoption, but my view is it will take several years for widespread adoption, to the tune of billions of regular users, to occur.
Google’s huge resources
Thus Google has an immense amount of resources. In 2023 it posted revenues of $305bn, and profits of $74bn, while its market cap is $1.93 trillion. By contrast the entire generative AI market is not forecast to reach this size by 2032.
Despite being backed with some huge investments. Open AI is likely to make near a $5bn loss in 2024, and has a lack of sustainable business model. Indeed, it is fairly clear from Sam Altman’s statements that OpenAI is not really currently looking for a business model as it charges towards Artifical General Intelligence. Whether or not this goal will ever be reached is really a debate for philosophers.
Returning to the economic reality, subscriptions can only get OpenAI so far if it is to burn such huge amounts of cash. This while Google makes huge profits from the most widely used advertising product on the Internet.
Physical and mental availability applied to AI
The real contest, I feel, lies in some of Byron Sharp’s most important traits for How Brands Grow – mental and physical availability.
Right now, ChatGPT is winning in mental availability. A company that few outside of Silicon Valley had heard of 3 years ago has a product of nearly 200m monthly users and 80% chatbot market share. It helps that it has ‘chat’ in its brand name, but also it was early and it’s a very good product. ChatGPT is to the LLM market as Google is to search.
The chatbot market remains relatively small
The problem is it’s only a market of c.250m (tops) right now. That sounds massive, but in contrast to social media, which is a market of 5bn, it’s about 5%. So OpenAI has won mental availability in a market with huge growth potential, but a market that is nascent nonetheless. It may take a superior LLM product to take back market share, but what about an LLM which is just physically more available?
Via the various strategic partnerships forming around ChatGPT, it is gaining more physical and mental availability. Bing, Co Pilot, Apple Intelligence, and probably the entire Microsoft Office suite will integrate it soon. This is a massive amount of physical availability, granted, but it’s not nearly enough to kill off Google.
Google’s dominant strengths
Why? Because Google is huge. The sheer vastness of its mental and physical availability is told through it having 131 billion monthly visits – 20 for everyone in the world. YouTube (a Google property) then receives 70bn visits. Facebook, a distant third, receives ‘just’ 13bn. Open AI is around 2.5bn – and we shouldn’t mistake website visits for users.
Google Workspace is also the largest office productivity suite, with 3bn monthly users, and 1bn for Google Docs. Microsoft Office 365 is 400m by contrast. Google’s Gemini AI/LLM is accessible directly in Google Docs, which is a potential game changer if it can consistently perform on a similar level to ChatGPT.
This is quite possibly where the long term LLM game will be won and lost. Large Language Models are really productivity applications, thus they should be physically available in other productivity applications. Google has the biggest physical availability here.
Google has undoubtedly seen a challenge to its almost complete stranglehold over search in the last two years, with Bing increasing overall share to over 10% for the first time, which correlates with its ChatGPT integration. But Google still holds 80% of the total desktop search market, and an incredible 96% of the mobile search market.
ChatGPT is a challenger, but it cannot currently be considered a usurper. The scale of user base and revenues is just too vast. Other novel search engines, like Perplexity AI, may suit a certain type of AI adopter, but they are a long way from becoming mass market products, never mind Google killers.
The biggest challenge to Google actually lies with itself. Its position is now so dominant, so unsurpassable, that the possibility of its search monopoly being broken up has moved to a probability. If that happens, then perhaps the new generative AI wave will be there to take the spoils. Until such time, its difficult to see too much changing.
