AI That’s Devouring the World, and Google at Its Center

In August 2011, Marc Andreessen, co‑founder and general partner of the prominent venture capital (VC) firm Andreessen Horowitz (a16z), published his now‑legendary essay "Software is eating the world", arguing that the enormous impact Silicon Valley tech companies would have on the world was being vastly underestimated. He pointed out that software had already penetrated our daily lives so deeply that it had become essential—from supermarket logistics systems to media and entertainment platforms like Netflix—and predicted that the gap between companies that excel at software and those that do not would widen exponentially.

Fifteen years on, if you look at the world’s top 10 companies by market capitalization, nine of them—excluding Saudi Aramco, the state‑owned oil giant of Saudi Arabia—are American, and eight of those (Microsoft, Nvidia, Apple, Amazon, Alphabet, Meta, Broadcom, and Tesla), excluding Berkshire Hathaway, are software or technology‑related companies. Marc Andreessen’s claim that software is eating the world was anything but an exaggeration; today, the companies that wield overwhelming global influence are, without exception, software companies. Some might argue that Nvidia is a hardware company because it designs semiconductors, but it is more accurate to describe it as a chip designer with exceptionally strong software capabilities. Over the past 15 years, a U.S. economy led by software has created far more added value than Europe or Asia and has driven major paradigm shifts across society and industry.

The last 15 years, dominated by software, have now reached an inflection point, and the paradigm is shifting toward AI (Artificial Intelligence). Here too, U.S. startups and Big Tech firms such as OpenAI, Anthropic, Google (Alphabet), and Meta are investing tens of billions of dollars a year in capital expenditures, leading paradigm shifts in both models and services. Among them, Google stands out for the frightening speed at which its technology is improving and its ability to integrate those advances into services. When it first announced Bard, the predecessor to Gemini, in March 2023, the prevailing view was that "Google has completely lost the upper hand in the AI model race". There is even a website called "Killed by Google" that lists services Google launched with great fanfare only to kill off after poor user response—symbolic of the criticism that, among major Big Tech companies, Google has world‑class infrastructure and engineering but lacks the ability to translate that into compelling products. As one industry insider jokingly put it, "Google Search makes too much money for them to care."

After Bard’s launch and the wave of ridicule that followed, Google focused intensely on improving the performance of its Gemini models. Sergey Brin, the co‑founder who had left the company years ago, is now reportedly coming into the office almost every day to work on boosting Gemini’s capabilities—a sign that Google sees AI as both an existential threat and a historic opportunity, and is mobilizing the entire company around it. The fruits of the past two years of effort were unveiled a few days ago at Google I/O 2025, and the recap video by tech outlet The Verge alone ran 32 minutes, packed with substantial announcements.

The most notable part is the fundamental performance improvement of the Gemini AI models. With its in‑house‑designed 7th‑generation TPU "Ironwood," Google has maximized computational efficiency, and the Gemini Flash 2.5 update has enhanced inference, coding, and long‑context processing capabilities. It also introduced a range of specialized models—such as the image generation model "Imagine 4," text‑to‑speech models, and security‑focused models—broadening the scope of AI applications. Tools like "Thought Summaries" and "Thinking Budgets," designed to help users understand and control how AI works, can be seen as part of an effort to increase transparency and accountability in AI. These moves are crucial steps toward strengthening the core competitiveness of Google’s AI technology and laying the groundwork for more reliable AI services.

Google is aggressively integrating Gemini’s advanced capabilities across its vast portfolio of services. The AI‑powered video communication tool "Google Beam," real‑time translation in Meet, and "Gemini Live," the successor to Project Astra, all demonstrate the potential to transform how users communicate. The Gemini app’s "Agent Mode," Gmail’s "Personalized Smart Replies," and the code assistant "Jules" are expected to boost productivity. In particular, AI search mode, Deep Search, and Search Live have the potential to fundamentally change how we search for and explore information.

Virtual try‑ons for clothing, AI shopping agents, and Gemini’s integration into the Chrome browser offer a glimpse of how AI could reshape everyday life and consumer behavior. The video generation model "VO3," the music generation model "LIA 2," and the AI content detection technology "Synth ID" also signal major changes ahead for content creation and distribution. All of this clearly illustrates Google’s strategy to use AI not as a single feature, but as a foundational technology that underpins the entire platform.

Google announced its AI Pro and Ultra subscription plans, outlining a monetization model for advanced AI capabilities. This shows how the company intends to translate the value of AI technology into business. At the same time, its collaboration with Samsung Electronics on Android XR development and related device launches signals Google’s renewed push into the extended‑reality market, which is drawing attention as a next‑generation platform beyond the smartphone.

Google is also using AI to tackle difficult problems in both basic and applied sciences, including mathematics, biology, and medicine. Projects such as Alpha Proof, Co‑scientist, AlphaFold 3, and Isomorphic Labs suggest that AI can play a key role in expanding the boundaries of human knowledge and accelerating scientific discovery.

No matter how well something is described in writing, it can’t match the shock and freshness of seeing it for yourself, so I recommend watching the video.

This phrase is one I coined myself, a line that compresses into a single sentence just how fast and powerfully AI is coming to dominate every industry.

In English, there is an expression: “You can't have your cake and eat it too.” In other words, you can’t enjoy two different benefits at the same time. But AI is doing the opposite. It is devouring the world while also seizing its power and control.

AI is rapidly reshaping manufacturing, finance, media, education, and healthcare. What matters is not that it merely supports existing industries, but that it is taking the seat at the very center of each industry’s ecosystem. Nvidia has gone beyond semiconductors to become the standard for computing, Microsoft is turning AI into something akin to an operating system, and OpenAI and Google are rewriting the rules of search and information access.

In the end, AI is the platform, the player, and even the self-appointed referee.

In the past, technology supported industries; now, technology is becoming the industry itself. At the center of that shift is no longer “software” but “AI.”

And at the very center of this, Google is not just turning AI into a tool; it is building the future Google on top of AI. Gemini is not just a language model; it is the new infrastructure of the Google empire, and this infrastructure is permeating every area where users interact—search, email, video, commerce, research, and education.