Why Alphabet will win in the age of AI
How Google created the technology now threatening its monopoly.
It all began in 2017. A small team of eight scientists working at Google Brain published a paper titled Attention Is All You Need.
That research introduced the Transformer, a new type of machine learning model that would go on to change everything. The very foundation that made ChatGPT, Claude, and every other large language model possible.
Ironically, Google invented the very technology that’s now being used to challenge it. The story of AI is, in many ways, the story of Google’s own creation escaping into the wild.
For those that want to learn even more, Secret Sauce Investing and The Dutch Investors talked about Google, Alphabet and AI for over 90 minutes.
The foundation
Before we dig in, we need to understand what AI is and means.
At the core, artificial intelligence is about creating computer systems that can learn, reason, and make decisions somewhat like humans do. Instead of following only fixed rules, AI systems learn from data and improve over time. They can recognize patterns, solve problems, and handle tasks such as understanding language or identifying objects. In simple terms, it’s teaching machines to think and act intelligently.
When companies say they’re investing in AI, they’re usually putting money into two main areas:
1) Building or buying the technology
They invest in data, software, and computing power to create systems that automate work, analyze information faster, or improve products. For example, a bank might invest in AI to detect fraud automatically, or a car company may invest in self-driving technology.
2) Hiring the people who can make AI useful
Engineers, researchers, and data specialists are needed to design smart tools and make sure they actually help the business.
Investing in AI (usually) means spending money to make processes faster, products smarter, and, ideally, profits higher (in the future). It’s mostly about efficiency and staying ahead of competitors.
While startups race to build anything AI, Alphabet has spent years integrating AI into products used by billions every single day, Search, YouTube, Gmail, Maps, Chrome, Android.
Seven products with over two billion users each, all now powered by Gemini, Google’s in-house AI model.
Alphabet doesn’t need to build a user base for AI, it already owns the distribution. Every Gmail autocomplete, YouTube recommendation, Google Ad, and Workspace draft suggestion is another example of AI at scale, embedded into daily life of users.
And behind it all is Google’s technical stack: massive data centers, fiber networks, and custom-built Tensor Processing Units (TPUs) that make training and serving AI models faster and cheaper than almost anyone else.
Alphabet is the only company today that, besides semiconductor fabrication, has fully mastered both sides of the frontier AI battle: it designs world-class AI chips and builds the state-of-the-art frontier models that run on them. Others may excel at one side or the other, NVIDIA dominates hardware, OpenAI leads in models, Meta has momentum, but only Alphabet controls the entire stack from silicon to the most advanced multimodal AI systems.
That combination, proprietary models, infrastructure, and global reach, is an advantage that no startup can replicate.
The innovator’s dilemma
Critics love to say Google faces an innovator’s dilemma1: that the rise of generative AI threatens its biggest profit engine: Search advertising.
Yes, generative answers might replace some traditional clicks. But the broader effect of AI is to make Search more useful, more personalized, and more efficient for advertisers. Alphabet is adapting its business model rather than fighting it. AI is transforming search into an interactive, knowledge-synthesizing agent, which encourages users to engage more frequently and deeply, ultimately driving more profitable connections between users and the commercial web.
AI-generated ad creatives, smarter audience targeting, and Performance Max campaigns are already improving returns for advertisers, and keeping them firmly inside Google’s ecosystem. Studies from market analysis firms2 indicate that visitors referred from AI search experiences (like a Large Language Model) can be 4.4 times more valuable than those from traditional organic search, based on conversion rate. This suggests that AI is pre-qualifying users, meaning the traffic that does click through is much closer to a purchase decision.
So instead of AI cannibalizing Search, it’s enhancing it. And through it all, Google still owns the text box or, as we like to call it; the front door to the internet.
Gemini is winning
“We’re taking another big step on the path toward AGI… It’s the best model in the world for multimodal understanding and our most powerful agentic and vibe coding model yet.” - DeepMind CEO, D. Hassabis
While many have argued that Gemini is behind or losing to other LLM’s, such as Claude or ChatGPT, it’s recent version of Gemini is beating them across all benchmarks. Just a couple months ago, Google’s Gemini was seen as the AI-loser. That narrative has shifted heavily, with GOOGL 0.00%↑ stock being up +78% in the past 12 months.
The fear was, and somewhat still is, that AI will decrease Search volume. We have argued from the very beginning that this will not be as bad as people anticipate, if not increase volume.
Alphabet CEO, Sundar Pichai, recently said on the Q2 earnings call:
Overall queries and commercial queries on Search continue to grow YoY. […] AI experiences significantly contributed to this increase in usage. […] AI features can meet more of their (users) needs. […] AI Overviews are now driving 10% more queries globally. AI Overviews are now powered by Gemini 2.5, delivering the fastest AI responses in the industry.
In addition to calming investor concerns about the viability of Google's search business, this surge in search volume helped Google accelerate its growth in paid clicks. It looks like the stock market realized that Google is essentially the only vertically integrated company in the AI space. And doing it profitable.
The prisoner’s dilemma
Imagine a new pair of high-tech football boots hits the market. They're legal and benefit the athlete, but they're expensive. They don’t turn your players into superheroes, but they give a small, noticeable edge. But that edge is just what a professional sports team needs to decide a tight match.
Suddenly, every club faces the same two choices:
Buy the boots
Don’t buy the boots
There are three possible outcomes:
Everyone waits: no one spends extra money, and the league stays in balance.
You wait, others buy: your team gets outrun. You lose points, fans and (potential) income.
Everyone buys: all clubs spend millions on getting everyone the best possible footwear. But if everyone buys the same boots, the advantage disappears and play on the pitch stays exactly the same. As a group, you end up worse off.
If you assume that everyone will behave selfishly, you end up with outcomes that are bad for everyone. The pursuit of individual advantage leads to a poorer outcome for everyone. Economists call this a prisoner’s dilemma. The world is full of prisoner’s dilemmas. They arise whenever you’re tempted by a short-term gain that wrecks the longer-term future.
The best outcome for the group is waiting. But individually, you can’t take that risk. So in the end, everyone buys the boots. Sure, clubs could agree not to buy them. But no one fully trusts each other, and it would look a lot like collusion. So that deal would never hold. No club wants to be the one remembered for missing the next revolution in football.
So yes, everyone overspends. Someone might win the title because of it, but as a group they’re mostly burning money, just to avoid falling behind.
Does this sound familiar? We think it sounds eerily similar to the AI revolution.
The AI dilemma
It’s eat or get eaten. Companies feel forced to pour billions into AI, because if the technology fulfills its promise, falling behind could be fatal. And if it flops, that money is gone forever. Either way, you pay.
AI is compute-hungry, capital-intensive, and brutally competitive. The winners will be the ones with the deepest wells of silicon, energy, and data. And this is where Alphabet outperforms the competition. Search, Maps, YouTube, Android, Gmail, are the scaffolding for a future in which intelligence is the interface. AI simply gives that backbone new purpose.
So who comes out on top? Those that:
build and operate the infrastructure profitably;
monetize AI across multiple platforms at scale;
do it without lighting shareholder capital on fire.
It’s not a long list. Alphabet is on that list if you ask us. The beauty of Alphabet’s AI strategy is that it’s not about creating something entirely new, it’s about improving everything Alphabet already has.
Final thoughts
AI will reshape how people search, create, and work. That’s for certain. Where the uncertainty lies is if Alphabet can transition fast-enough. So far, they’ve shown us just that, but the marathon isn’t over yet.
Alphabet built the foundation of the AI era, and now it’s integrating that foundation into everything it does. The company that once organized the world’s information is now teaching the world how to interact with it.
To be continued…
If you wish to download the Google AI-research paper, you can download the PDF below.
If you want the full premium deep dive, check out TDI Premium for the complete report and podcast on Alphabet Inc.
The paradox where successful, well-managed companies fail because they focus on pleasing current, high-paying customers and improving existing products (sustaining innovations), which makes them overlook disruptive innovations that are initially inferior but cheaper, serve new markets, and eventually overthrow established leaders
This article comes at the perfect time. Your point about 'Google’s own creation escaping into the wild' really stuck with me. Its so ironic and makes you think about the future.
Yup. Well said