Google's AI Advantages
“I’ve always thought of AI as the most profound technology humanity is working on. More profound than fire or electricity or anything that we’ve done in the past."
- Sundar Pichai
Alphabet investors face a new existential crisis every few years. There was the shift to mobile, but putting a computer in everyone's pocket created many more opportunities to Google things. There was fear that vertical-specific apps would replace generalized search, but apps only made sense in a few categories like e-commerce and travel, and many consumers chose to start their searches on Google anyway. There was the rise of Facebook, Instagram, TikTok, and Amazon ads, but the digital advertising market was plenty big enough for multiple successful companies. And let's not forget Amazon's Alexa and Apple's Siri, but voice wasn't a great interface for computing—people read twice as fast as they speak, and you can convey far more information with pictures than words.
The latest fear is that Google will be disrupted by generative artificial intelligence. It's ironic, considering Google was an early leader in AI. It developed much of the essential technology (especially "transformers") that made the current wave of AI innovation possible. Alphabet's stock has been trading toward the low end of its historical valuation range since ChatGPT launched in late 2022.
There are a few ways generative AI could hurt Google's business:
• Consumers might use Google Search less, preferring to get direct answers from gen-AI applications like ChatGPT;
• If Google retains users by including generative AI in its own results, it could mean less room for ads;
• Gen AI in Google Search could increase computing costs and latency.
I don't think any of these risks will end up mattering. Just like the shift to mobile, I'm betting generative AI will make Google Search more useful than ever, and that the market opportunity is large enough to accommodate multiple winners.
Google's AI Advantages
Here are some of the competitive advantages Google brings to this new era of AI:
Distribution. I’ve seen many complaints that Google either hasn't evolved its search experience—that it's stuck on 10 blue hyperlinks—or that Google Search has become unusable because of too many ads or spammy, overly-SEO'd results. I'm always reminded of declarations that "nobody uses Facebook anymore," even as Facebook was reporting more than 2 billion daily active users.
Google.com is the most popular website in the world by a wide margin. According to Similarweb, Google receives more than 84 billion monthly visits. The #2 website—YouTube, also owned by Alphabet—gets 32 billion monthly visits. Microsoft's Bing only receives about 1.3 billion monthly visits, and ChatGPT is averaging about 1.7 billion. Consumer habits and browser defaults partly explain Google's enduring popularity, but the service is also extremely useful. If it weren't, people wouldn't go there so often. I personally Google things dozens of times a day, from the mundane (confirming the definition of a word) to the more in-depth (helping my 9-year-old son research pollination or the Hindenburg disaster).
Anyone who thinks Google is just 10 blue links must not have used the site in many years. If I search for factual information—say, "Mount Fuji elevation"—I not only get the answer right at the top of the results (12,388 feet), but also pictures of the volcano; a map with nearby hotels and attractions; other mountains with their heights; related questions, facts, artwork, and animals; 10,000+ Google reviews; and a long list of links for further reading. If I search for "lamp" I get a scrolling gallery of lamp ads complete with pictures, reviews, delivery options, and purchase links—plus a set of filters to refine my search by style, features, price, retailers, and more. All this customized, relevant information appears in under a second.
It's increasingly common to see generative AI at the top of my search results, too. It doesn't make sense for every query, but for more detailed questions (less than 20% of my searches?), gen AI is superior. Technology enthusiasts often forget that average people aren't paying nearly as much attention to the latest developments at OpenAI, Perplexity, or Anthropic. Their first experience of generative AI will be when it magically appears in their Google Search results; it won't even occur to them to look elsewhere. This is the power of Google's distribution.
Monetization. Google also has an advantage in monetizing generative AI. It's much easier to get consumer adoption if a service is free, which means selling ads. Arguments that Google can't place ads within and alongside gen-AI results seem to fundamentally misunderstand the nature of search. Few queries have a single correct answer, and those usually don't have commercial intent (the Mount Fuji example). When there are multiple options, advertisers are willing to pay for more prominent placement (the lamp example). If I'm looking for a hotel room, or shopping for a new car, or researching financial advisors, or picking a restaurant, I want to see options. There's no reason some of those options can't be sponsored. Trying to recreate Google's ad machine would be an enormous hurdle for any new entrant.
Financial Resources. Bigger isn't always better—especially with new technologies, where smaller companies are nimble and innovative—but training AI models requires huge upfront costs. Alphabet spent $12 billion on capital expenditures just in the first quarter, and nearly as much on research & development. OpenAI has only raised about $12 billion since its founding in 2015 (almost entirely from Microsoft). And while OpenAI is growing rapidly, reportedly on track for about $2 billion of annualized revenue, Alphabet generates that much revenue every 2.5 days.
Technology. It's not enough to train a generative AI model—you also need fast and cost-effective inference. Google has been investing in its datacenters and network for more than 20 years. Its proprietary AI accelerator chips, Tensor Processing Units, are on their 5th generation and have been in development since 2016. TPUs train and serve Google's Gemini model, creating a meaningful cost advantage versus companies that rely on Nvidia's GPUs. According to management, the cost to deliver gen-AI responses in Search has fallen by 80% over the past year. Technology advances can be shared across other Alphabet products like Google Cloud and YouTube's recommendation engine.
Data. Alphabet has unmatched data on its users, from search history to location history in Google Maps, pictures in Google Photos, and the content of emails in Gmail. Some of this data is used to refine search results and display relevant ads, but it can also train more helpful and personalized AI assistants.
Risks
I think Google is well positioned for the rise of generative AI. Still, any time there's a paradigm shift in technology, there's a possibility incumbents will be left behind. Management execution is more important than ever, and Google's management hasn't always been up to the task. A few gen-AI releases have gone poorly, such as when Gemini's image generation feature insisted on racial diversity in response to historically specific prompts (the Founding Fathers, Nazi soldiers). The company is trying to become more efficient (with good reason), but layoffs have dragged on for more than a year, weighing on employee morale. While I can appreciate Google's desire to experiment, it's developed a reputation for haphazardly discontinuing products, which makes customers hesitant to try new offerings. Management should also share more information with investors, such as a complete view of YouTube's finances and disclosures about the revenue breakdown by advertiser vertical.
Management aside, the three biggest vulnerabilities I see are regulation, Meta, and Apple. Google faces regulatory scrutiny on many different fronts and has already lost a few high-profile cases. Despite billions in fines, the business has adapted with few lasting impacts, but that may not always be the case.
Meta is making huge investments in artificial intelligence. I see this as a bigger threat than startups like OpenAI—which is rumored to be working on a search engine—because Meta has all the same advantages as Google: Distribution, monetization, financial resources, technology, and data. With an ambitious, 39-year-old founder as CEO, Meta has acted more decisively and aggressively than Google.
Lastly, Google's relationship with Apple could be its Achilles' heel. There's no better way for a competitor to get traction than to outbid Google to become the default search engine on iOS. It wouldn't be easy or cheap. Court documents from a recent antitrust case revealed that Google paid Apple $20 billion for this privilege in 2022—36% of the associated advertising revenue. Microsoft offered Apple a 90% revenue share to replace Google with Bing, even volunteering to hide the Bing brand, but Apple declined. I see this as strong evidence for Google's moat, but it would be a bad sign if Apple ever changed its mind.
Disclosures
Moatiful is an independent publication of Trajan Wealth, L.L.C., an SEC registered investment advisor. The views expressed are solely those of the author, and may not reflect the views of Trajan Wealth. Nothing in this blog is intended as investment advice, nor is it an offer to buy or sell any security. Posts are for entertainment purposes only and should not be relied on when making investment decisions. Please consult your financial advisor for questions about your personal financial situation. All investments involve risk, including the potential for loss. Historical results may not be indicative of future performance. Data from third-party sources is not guaranteed to be accurate, timely, or complete. Links to external sources are provided for convenience only, and do not constitute an endorsement by Trajan Wealth. Clients and employees of Trajan Wealth may have a position in any of the securities mentioned. Data and opinions are subject to change at any time without notice.