Will AI lead to Google's downfall?

The Frontier Investor newsletter is a weekly publication with deep dives into building & investing into frontier tech - AI , crypto, exotic markets, and more

Click below to subscribe!

I was talking to a friend of mine earlier in the week who sits in one of the top tech long/short pods at a large hedge fund - he’s not confident on Google’s competitiveness on the AI arms race.

“I think Google’s search business will get disrupted by AI”

Attacking Google’s search business has been an increasingly popular theme amongst AI foundation model builders. In fact, Google’s stock dropped 3% (relatively a lot for a megacap) on news of OpenAI testing their SearchGPT:

How real is the threat of AI to Google? Is this an opportunity for AI investors & builders?

What Matters for Google & Sundar Pichai?

Let’s set the context first - Google’s ads business is by far their first and foremost priority. Yes they have great products outside of it like YouTube & cloud. But the search business is at the center of new product launches both from a capital perspective (cash flow to subsidize growth) and cross-sell perspective.

Their latest 10K shows ~60% of their business is from search still despite great growth from YT and cloud.

From a business perspective, all roads lead back to search here.

Google became dominant in search not because they were the first to launch, but because they were able to deliver results significantly better than any of their competitors. In doing so, they rewired users to default to starting the internet journey with Google queries. The innovations they’ve done on the ad networks etc are precisely because they own the top of the internet funnel.

In fact, Google is by far the biggest source of traffic on the web today. Google does 10x the web visits by their closest non-Google owned competitor (Facebook).

Owning this top of the user funnel is the most important pillar for their business model. In private, Google has shown their hand to this fact if we take a look at court-released documents from their anti-trust lawsuits.

From internal documents, we see that Google discussed internally that Apple’s ownership of traffic through consumer hardware dominance can threaten their position as the top of the funnel to the internet. And even listed a prediction that Apple could launch an ad network of their own by the late 2020s:

This is a reason why Google pays Apple $20B / year to be their default search engine - so that Google can maintain their dominance in top of the funnel for internet journey. For context, this is 1/5th of Apple’s entire annual net income. Not a small number at all.

How Does AI Come Into Play?

So we know that owning the first user touchpoint is the most important piece of building a great ads business.

Unfortunately for Google, one of the clearest outcomes from the GenAI boom is a rewiring of consumer internet behavior.

Previously, the user journey looked something like this:

1. I have a question

2. I search my question or topic on Google (Fastest & most accurate algorithm to answer what I need)

3. I get redirected to another website

But with the launch of ChatGPT (and many other foundation models now), the user journey increasingly looks like this:

1. I have a question

2. I ask on an AI chatbot

3. If answer is not satisfactory, then I go to Google

The speed and convenience of AI chatbots is clearly evident in the speed of ChatGPT’s product market fit. Instead of taking an additional step for a link-based search result, the GPT can push out a concise and relevant direct answer.

As these models become more powerful, the accuracy should increase leading to users to stop at step (2) in the list above. And the product roadmap for these AI players is looking increasingly towards turning on web crawling which becomes directly competitive with Google Search.

We already mentioned OpenAI experimenting with web crawling. But they’re not the only ones.

Perplexity.ai for example is building a dedicated AI based independent search engine - check it out, I highly recommend it. Image, video, and source based text results are all there with Perplexity with faster output.

Of course Google is building their own AI in Gemini - which they’re aggressively integrating with their search experience. But they are at a vulnerable spot given they’re already at 99% search market share and any losses chip away at the network effects they hold over advertisers.

It also doesn’t help Google’s case that they are inserting political bias (some accuse) in addition to brand bias towards their biggest advertising customers. Even if these are just perceived biases, that alone has been a driver for some folks to adopt alternatives.

So to recap - 3 clear consumer behavior trends with this new age of AI chatbots:

1. Google trust and accuracy declining

2. AI search becoming increasingly better, leading to direct Google Search competition

3. AI queries becoming increasingly better

Google still holds the upper hand and can maintain their dominance if they out innovate on AI search experience (or strike deals to route AI based search to Google’s already built out algorithms). But AI founders have Google’s juicy business model in sights and it’s definitely something to keep an eye on.

Lessons for Founders:

A couple lessons for founders here:

1. Focus on user experience: Focus on the user experience is obvious but sometimes overlooked.

2. No moat is too comfortable: IBM and Yahoo were once the internet darlings 20 years ago. Things change fast in the internet age.

3. Move fast: Google was by far the leader in AI for many years - in fact, they were the original publishers of the transformer paper that kicked off this AI boom. But slowness in the last few years has led others to take the lead here.

4. Original content > aggregator models: Owning content gives you a clear business edge. This is evident not just in the example above, but across industries. For example, part of Stripe’s advantage is their direct access to local payment methods (vs aggregating pipes others have built).

That’s all for now.

If you made it this far, forward it to a friend you think may enjoy this post!

Note: This newsletter is intended solely for entertainment purposes and should not be considered an investment recommendation. It does not provide financial advice or serve as a solicitation to buy or sell any assets. Please conduct your own research. All opinions and views expressed are those of the author and do not reflect the institutional views of any company with which the author is affiliated as a partner, collaborator, or investor.