Even though Wall Street is panicking over Google's earnings call, it seems obvious to me what is happening. They are not going to slow down when it comes to AI. Throughout the call they spoke of agents and how we are in the early days of how they are changing Search. They told us that AI Mode, AI overviews and Traditional Search becoming one unified product.
They also said to expect rapid improvement because of RSI. Things are about to get very interesting.
I think new opportunities are arising, but also much of the way we have traditionally been able to make money from Search is disappearing.
Watch my video with thoughts on the Earnings Call
If you prefer video I share my thoughts on the call and also on the new Connected Apps program which will let people use your business agentically right from Search and the Gemini app.
The main thing I took away from this earnings call is the emphasis on the shift to Agentic Search. I think this is still unfolding and no one fully knows what this new era of the web will look like. Google gave us a few hints in this call.
I find it fascinating how Google’s stock plummeted after this call. Alphabet’s revenues were up 25% YOY and Cloud revenue was up 103%. The general concern seems to be that Google is pushing AI Mode more, and so far there are fewer opportunities for ads shown in AI Mode. Therefore, revenue from Search is lower than expected. Still, Search advertising is up 12% YOY. (Wall Street was expecting 15%.)
I have no concerns over Google’s ability to continue to thrive. They mentioned in the call that their first priority is to make AI Mode a quality product and then they will focus more on monetizing it. Yesterday Google announced that ads in AI Mode were expanding to more countries including Australia, Canada and more.
Philipp Schindler said,
“Our approach is deliberate - prioritizing the user experience and engagement first, then scaling the formats.”
Mark my word, Google will make plenty of money from ads in the years to come. I think the new connected apps (see more on this below) could eventually be offered as ads.
They also spoke extensively about how there is incredible demand for Google’s computing power. They talked about TPUs not as chip sales, but as systems. “TPU Systems” were mentioned 35 times in the call. Thomas Kurian said,
“A TPU System is not just a chip, but it’s a specialized system that combines multiple things, including hardware, computer, storage, networking and so we are selling much more than just a chip. Software, training, inference.”
I mention this because Google will likely see incredible growth as more and more customers purchase these TPU systems.
What’s most interesting to me is how Google’s verbiage has shifted when it comes to talking about AI Mode and its relationship to Search.
Google sees AI Search and Traditional Search as one unified product
The idea of traditional Search and AI Search merging into one product was mentioned several times:
Sundar Pichai:
“We are building one seamless experience across AI Overviews, AI Mode, and Traditional Search.”
Philipp Schindler:
“We think of Search holistically, a unified product spanning traditional results and AI powered Search.
Agentic Search is unfolding
I know it feels like Google is behind when it comes to having an agentic product. X has Grokbot. OpenAI has Dots, and Meta’s Muse is taking off. It seems clear to me that the vision of AI being our personal assistant has been in Google’s mind for some time. I think they’re taking their time to fully unpack it all because they need to get it right. Google’s end game here is not just an agent that will help us get things done online, but much bigger!
In May of 2025, Demis Hassabis wrote about Google’s dream of building what they called a “Universal AI Assistant.” Much of this piece discussed Google’s Genie world models acting as training environments for robots and how Project Astra which is now in Gemini Live, will allow us to have insight on the world around us. Google’s mission to organize the world’s information and make it universally accessible and useful.
I see the pieces coming together. It starts with online agents and will eventually expand into every aspect of our lives especially as the technology develops for intelligent eyewear and in-home robots.
Philipp Schindler said:
“We are in the early stages of agentic experiences across our surfaces, and excited about the commercial opportunities ahead.”
He went on to speak about how they started by building agentic experiences into Gemini Enterprise first and, “as a next step [they’re] bringing these features to Search.”
Sundar said,
“It’s clear that the agentic capabilities we all envision are going to be real,”
and also,
“I’m excited about agents, and I would view it as a new frontier, which brings in a new set of use cases to Search.”
Those new use cases likely start with Connected Apps.
btw...if you want to be ahead of the curve in seeing how businesses are adapting to agentic search, I send out an email every week with all of the latest news on Search & AI:
The Connected Apps Platform - people can use your website's functionality right from Search
On the day of the call, Google announced the Connected Apps Platform. There are already a handful of apps that can be used via MCP by Gemini either in the app or from within Search. Here is an example they gave where a user calls @Peloton to bring the Peloton app experience into their agentic flow within Gemini.
You can apply to the Connected Apps Platform here.
It was six months ago that Sundar Pichai said in an interview that “Search will be an agent manager.” I think this will be a new era where we shift from Search being a place to find websites that have information to a place where you use tools to help you (or your agents) get things done.
Gemini use is growing
Pichai said,
“The Gemini app is now at one billion users and is the fastest growing product in Google’s History.”
A billion users is the same number reported by ChatGPT in June of 2026.
Gemini use is growing quickly.
I find this fascinating because my own use of the Gemini app has declined significantly. I use Gemini in Antigravity for almost everything.
Coding improvements are coming, starting with Gemini 4 Carbon
We all know that Google’s models have not been as advanced as those from OpenAI and Anthropic.
Yesterday Google announced a new frontier model, just two weeks after Gemini 4 Argon. Gemini 4 Carbon is currently undergoing voluntary testing as per the US Government’s recommendations and will soon be released. Artificial Analysis currently has Carbon listed as the most intelligent model to date.
On the call, Sundar Pichai said that more models are coming soon, including “more efficient versions to help developers optimize for speed and cost.”
He said they started seeing a step change with Gemini 3.8 Flash when it came to coding internally at Google. And then Gemini 4 Argon was a “huge point of inflection.” He said,
“It is one of the most meaningful internal productivity levers [they] have seen on software engineering, directly tracking high-quality code that’s getting checked in, and the rate at which it’s getting checked in.”
Then he said,
“Carbon is a meaningful step up from Argon.”
Right now I use Gemini 4 Argon across 3 or 4 simultaneous Antigravity conversations most of the day long. On the Ultra plan, I have barely touched my quota. I thought 3.8 Flash was amazing. It helped me start to automate many of the SEO reports I run. When I switched to Gemini 4 Argon a mind blowing thing happened. Gemini 4 Argon + my internal OKF powered knowledge brain and my algo update MCP started producing far better consulting reports than I could alone. I truly believe we are not far from the day where AI will be better at all knowledge work than even the best humans in their field.
On AGI
This week Google announced Genie 4, their latest world model. While it’s cool to see how it can generate interesting scenes and games, the most exciting thing about this model is that it will be used to create training simulations for AI agents and also for robots.
Sundar Pichai:
“Beyond our Gemini 4 series, yesterday we announced Genie 4, our most advanced world model yet that simulates safe, realistic, interactive worlds with native audio in real time. This is an important building block for training robots and agents.”
He then said,
“I would expect the next twelve months to see continued acceleration because of RSI.”
Buckle up. RSI is “Recursive Self Improvement.” AI is being used more and more to help improve AI. We will see it improve rapidly.
Here's the call if you'd like to listen to it:
If you liked this, you'll love my newsletter:
Or, join us in the Search Bar for up to the minute news on Search and AI.
Comments are closed.