To The Moon

This week, $Alphabet(GOOG)$ $Alphabet(GOOGL)$ announced Gemini 2.0 and a flurry of new AI features, including AI agents that can take control of Chrome and surf the web for you and a deep research AI.

As I’ve explored the world of AI, I keep coming back to 2 simple questions:

1. How is any breakthrough differentiated from the competition?

2. How can an AI startup (even OpenAI) disrupt or overcome an incumbent offering similar capabilities with vastly more distribution?

Google doesn’t seem to be falling into the disruption trap.

Google has arguably the best models today and, BY FAR, the best infrastructure and distribution of the major AI players. And it’s not close.

If Google can continue this momentum in AI, it could mean this is a sustaining innovation, not a disruptive innovation, terms coined by Clayton Christensen in The Innovator’s Dilemma. And it could be a sustaining innovation on steroids.

Google’s Infrastructure Advantage

Google has spent two and a half decades taking microseconds out of the time it takes to perform a search. That performance improvement has cost billions of dollars in data centers and fiber that Google can leverage today in AI.

• Data Centers: There’s no official count of Google’s data centers, but the company has at least 43 data centers either running or under construction, according to Data Center Map.

• Fiber: Google has a global network of fiber to connect data centers.

• TPUs: Google has its custom chips called tensor processing units that run Google’s AI efforts and are core to Google Cloud’s growth. This week, the company announced Trillium, the 6th generation chip, and said, “Trillium is 4x more performant and 67% more energy efficient than its predecessor.”

• Data: Training AI is all about data, and Google has arguably the best data set with search data, email, and (most importantly today) YouTube.

There are only a few companies in the world that could compete with Google on any one of these pieces of infrastructure, and no company has all four.

Google’s Distribution Stranglehold

As AI develops, I think it’s becoming clear models are becoming a commodity, and applications/use cases will be easy to replicate. Distribution may be the most important point of leverage, and Google is sitting pretty with:

• Chrome - #1 browser

• YouTube - #1 streaming service (bigger than Netflix)

• Gmail - #1 email provider

• Android - #1 mobile OS

• Google Maps - #1 mapping service

Not a bad starting point for getting new AI applications into people’s hands.

Bringing It All Together

Why is all of this relevant today?

Google’s introduction of Gemini 2.0 and some of the AI Agents (this will be a theme in tech for the next year) was one of the more impressive demos I’ve seen from a normie perspective.

At 1:30 of the video below it shows Gemini taking control of the Chrome browser and performing multiple actions. Prompting something like “Find the shoes Steph Curry wore tonight and add them to my Amazon cart” would do research and take action for you.

We are early in finding use cases for AI Agents, but to me, this is much more compelling than using an LLM to write an email or write an article with multiple hallucinations.

With AI Agents, technology can do work for us, and Google is in a unique position to both build the models and deploy them at scale with Chrome and Android.

I don’t see how ChatGPT tops these inherent advantages.

Sustaining Innovation vs Disruptive Innovation

It doesn’t look to me like Google is being disrupted by AI. It looks to me like Google is going to get AI into the hands of more people than any company in the world. ChatGPT may be the go-to for techies who are early adopters but remember that normies will just use what’s available on their phones or in their browsers. More normies have Chrome and Android than any other browser or mobile OS.

Does AI translate to more money for Google than search? I don’t know. But if AI Agents are saving customers time and money, there’s value to be extracted. Like monetizing search, Google will figure it out.

AI looks like a sustaining innovation to me, and in this case, it could supercharge Google’s growth in the cloud, mobile software, hardware, chips, and more.

Quantum Disruption?

Speaking of disruption. Did you see this?

Meet Willow, our state-of-the-art quantum chip

I’m not going to pretend to know anything about quantum computing other than what I’ve learned in the last 72 hours, but it seems incredibly impressive.

I didn’t even know about Alphabet’s quantum computing efforts.

I like to say, “Buy optionality you don’t have to pay for.” Alphabet seems to have that in spades.