The Number That Explains Everything
On April 29, 2004, a company that had existed for barely six years filed a document with the Securities and Exchange Commission that contained, buried in the boilerplate legalese of a Form S-1 registration statement, a number so peculiar that it could only have been planted there as a joke — or a confession. The proposed maximum aggregate offering price for Google Inc.'s initial public offering was listed as $2,718,281,828. That figure is not arbitrary. It is e — Euler's number, the base of the natural logarithm — multiplied by one billion, carried to nine decimal places. The mathematical constant that governs continuous exponential growth, embedded in the founding financial document of a company that would go on to become the most consequential growth engine in the history of the internet. A nerd's Easter egg. But also a statement of intent so brazen that Wall Street, which prefers its ambitions stated in basis points and comparable transactions, largely missed the joke.
Two decades later, the joke has aged into prophecy. Alphabet Inc. — the holding company that Google became in October 2015 — reported full-year 2025 revenue of $402.8 billion, an 15% increase over the prior year. Net income reached $132.2 billion. The company's market capitalization crossed $4 trillion for the first time in January 2026, making it the fourth entity in history to achieve that valuation, joining Nvidia, Microsoft, and Apple. Its stock surged 65% in 2025 alone — the sharpest annual rally since the company doubled coming out of the 2009 financial crisis. More than 750 million people use the Gemini AI app monthly. YouTube's annual revenues surpassed $60 billion across ads and subscriptions. Google Cloud ended the year at a $70 billion annual run rate. Every second of every day, Google processes an estimated 100,000 web searches, each one a tiny auction conducted and settled before the user's finger lifts from the key.
And yet the paradox that makes this company worth studying — the tension that has defined its trajectory across three decades — is that the organization responsible for the single largest profit pool in the history of advertising was founded by two people who actively disdained advertising, built a homepage so aggressively empty of commercial intent that it was nearly blank, and wrote an academic paper in 1998 arguing that "advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers."
The story of Google is, in this sense, the story of a company that became the thing it feared. But it is also the story of how that becoming — reluctant, conflicted, brilliantly engineered — produced one of the most durable competitive advantages in the history of capitalism.
By the Numbers
The Alphabet Empire
$402.8BFull-year 2025 revenue
$132.2BFull-year 2025 net income
~$4TMarket capitalization (Jan 2026)
32%Operating margin (FY2025)
750M+Gemini App monthly active users
$60B+YouTube annual revenue (ads + subscriptions)
$70BGoogle Cloud annual run rate
65%Stock price increase in 2025
Two Graduate Students and a Garage Full of Implications
The founding mythology of Google has been told so many times that it has calcified into something between Silicon Valley scripture and corporate hagiography — the Stanford dorm room, the misspelled mathematical term, the garage on Santa Margarita Avenue in Menlo Park. But the details that matter most are the ones that reveal not just who Larry Page and
Sergey Brin were, but what
kind of company they were constitutionally incapable of
not building.
Larry Page grew up in East Lansing, Michigan, the son of two computer science professors at Michigan State — one of the first families in America where the dinner-table conversation assumed that computers would eventually mediate all human knowledge. He was, by temperament, an engineer who thought in systems: obsessed not with individual products but with the architecture of how things connect. As a boy, he devoured the biography of
Nikola Tesla and came away haunted — not by Tesla's genius, but by the fact that Tesla died impoverished, his inventions commercialized by others. The lesson Page extracted was not about the romance of invention but about the necessity of control. "He could have been more," Page would later say of Tesla. The implication: invention without institutional power is tragedy.
Sergey Brin arrived at Stanford from a different vector entirely. Born in Moscow in 1973, he emigrated to the United States at age six with his family — his father a mathematician, his mother a researcher at NASA's Goddard Space Flight Center. Brin had the particular restlessness of the immigrant polymath: brilliant at everything, bored by most of it, drawn to problems that were both mathematically elegant and practically enormous. Where Page thought in systems, Brin thought in data. Where Page was deliberate and controlling, Brin was fast, intuitive, occasionally reckless.
They met in the summer of 1995, when Page visited Stanford as a prospective PhD student and Brin was assigned to show him around. The apocryphal version has them arguing about everything. The more instructive detail is what they argued about: the structure of the nascent World Wide Web, and specifically whether the link structure of the internet — the pattern of which pages pointed to which other pages — contained information that no one was yet extracting.
It did. The insight that became Google was not, at its core, a search insight. It was an insight about authority. Page and Brin recognized that a link from one webpage to another was not merely a navigation aid but a vote — a declaration of relevance, weighted by the authority of the page casting the vote. A link from the New York Times carried more weight than a link from a personal homepage, because the Times itself had accumulated more inbound links. The algorithm was recursive: authority begat authority. They called it PageRank, a pun on Larry's name and on the thing it ranked, and it was the single most consequential idea in the history of information retrieval.
Our fundamental goal is to help users find the information they are looking for as quickly as possible. That's why the first page is the way it is.
What made PageRank revolutionary was not just its accuracy — though it was dramatically more accurate than existing search engines, which ranked pages by crude keyword frequency — but its scalability. The algorithm improved as the web grew. Every new page, every new link, every new website made Google's index richer and its rankings more precise. The web was its training data, and the web was growing exponentially. In October 1999, when Google was receiving four million search requests per day, Brin told a reporter that the company's plan was to "build a better search engine and generate revenue by licensing its technology to other companies." The company would not run banner ads. It would not add chat, email, or stock quotes. "We're not a media company," Brin said. "We're focusing on areas where we can make a big contribution."
This was sincere. It was also, within five years, entirely wrong.
The Reluctant Ad Machine
The transformation of Google from a search-technology licensor into the largest advertising business in history was not a pivot. It was an evolution so gradual, so logically inevitable given the company's assets, that by the time anyone — including the founders — fully appreciated what had happened, the machine was already running at scale.
The key intermediary was a company Google did not build: Overture Services, originally called GoTo.com, founded by Bill Gross in 1998. GoTo.com pioneered the model of auctioning search keywords to advertisers — the highest bidder for a given search term got the top placement in results. The model was crude, easily gamed, and philosophically repugnant to Page and Brin, who had written in their 1998 Stanford paper, "The Anatomy of a Large-
Scale Hypertextual Web Search Engine," that advertising-funded search was inherently corrupting. But GoTo.com proved something that no amount of academic squeamishness could refute: people searching for information had
commercial intent, and that intent was extraordinarily valuable to advertisers.
Google's version of this insight — AdWords, launched in October 2000 — was characteristically more elegant. Rather than simply auctioning placement to the highest bidder, Google incorporated relevance into the auction mechanism. An ad's position was determined not just by how much the advertiser was willing to pay per click, but by how often users actually clicked on it — a proxy for relevance. This meant that a more relevant ad from a lower bidder could outrank a less relevant ad from a higher bidder. The system was self-improving: advertisers were incentivized to write better, more relevant ads, which made the ad experience better for users, which kept users searching on Google, which attracted more advertisers. A virtuous cycle that was also, not incidentally, a printing press for cash.
The genius was structural: by tying ad quality to ad economics, Google had aligned the incentives of advertisers with the interests of users — or at least created a sufficiently convincing simulation of alignment. The ads were, as Brin had promised in 1999, "highly targeted text ads that are fast-loading and not a distraction." They were relevant to searches. And they generated revenue at a scale that made licensing deals look like lemonade stands.
By the time Google filed its S-1 in April 2004, the company was generating $961.9 million in annual revenue, virtually all of it from advertising. The filing itself was as unconventional as the offering price. Page and Brin included a founders' letter modeled explicitly on
Warren Buffett's communications to Berkshire Hathaway shareholders — a stylistic choice that communicated, simultaneously, their admiration for long-term capital allocation and their contempt for the short-termism of public markets.
Google is not a conventional company. We do not intend to become one.
The IPO, conducted via a Dutch auction to democratize access and minimize Wall Street's cut, priced at $85 per share on August 19, 2004 — below the originally anticipated range, after a rocky roadshow marked by a Playboy interview that nearly derailed the SEC registration. The market capitalization at listing: approximately $23 billion. Within a year, the stock had more than tripled.
The founders' letter established principles that would govern — and sometimes haunt — the company for two decades. A dual-class share structure gave Page and Brin effective voting control regardless of how much stock they sold. The letter explicitly warned investors not to expect quarterly earnings guidance, smooth financial results, or conventional corporate behavior. "We will have the fortitude to do this," they wrote, with the confidence of two thirty-year-olds sitting on an exponential growth curve and a governance structure that made them unaccountable to anyone but each other.
For those curious about the intellectual architecture behind Google's early years, Steven Levy's
In The Plex remains the most granular account of how a search engine became a civilization-scale infrastructure company.
The [Attention](/mental-models/attention) Monopoly
What Google built in the decade between its IPO and the creation of Alphabet was not merely a dominant search engine. It was something more structurally formidable: a monopoly on the moment of intent.
Consider the economics. Every other form of advertising — television, print, billboards, radio, even early banner ads — operates on the principle of interruption. The advertiser pays to insert a message into an experience the consumer is already having. The consumer tolerates the interruption because the content is free or subsidized. The advertiser accepts waste — the vast majority of people who see a TV ad have no intention of buying the product — because there is no better mechanism for reaching those who do.
Google inverted this model. A Google search ad appears precisely at the moment a user has expressed intent — typed a query, revealed a need, signaled willingness to act. There is no interruption because the ad is the answer, or at least adjacent to it. There is minimal waste because the advertiser only pays when the user clicks. And the feedback loop is instantaneous: the advertiser can measure, to the penny, the return on every dollar spent, adjust bids in real time, and scale spending precisely to the point where marginal revenue equals marginal cost.
This is not advertising in the traditional sense. It is a toll booth at the intersection of demand and supply, and Google owns the intersection.
The structural dominance compounded through a series of acquisitions and product launches that extended Google's reach across every surface where intent could be captured or attention could be monetized. YouTube, acquired in October 2006 for $1.65 billion — a price that seemed absurd at the time and now looks like one of the great bargains in corporate history given the platform's $60 billion-plus annual revenue. Android, launched in 2008, which became the operating system for roughly 72% of the world's smartphones and ensured that Google Search was the default entry point for the majority of mobile internet users on Earth. Chrome, released in September 2008, which grew to command roughly 65% of browser market share and served as both a data-collection instrument and a distribution channel for Google services. Gmail. Maps. The acquisition of DoubleClick in 2007 for $3.1 billion, which gave Google control of the infrastructure for display advertising across the web.
Each acquisition followed the same strategic logic: own the surfaces where users express intent or spend attention, and then monetize that intent through the advertising auction. The search engine was the nucleus, but the empire extended outward through every layer of the stack — the browser, the operating system, the video platform, the email client, the mapping service, the ad-serving infrastructure.
🔍
The Acquisition Architecture
Key deals that built Google's attention monopoly
2003Acquires Applied Semantics, which becomes the basis for AdSense
2004IPO at $85/share, raising $1.67 billion
2005Acquires Android Inc. for an estimated $50 million
2006Acquires YouTube for $1.65 billion in stock
2007Acquires DoubleClick for $3.1 billion, consolidating ad infrastructure
2008Launches Chrome browser and Android operating system
2014Acquires DeepMind for approximately $500 million
2015Restructures as Alphabet Inc.; Sundar Pichai becomes Google CEO
By 2015, Google's market share in search exceeded 90% globally. All other search engines combined — Bing, Yahoo, DuckDuckGo, Baidu (outside China) — accounted for barely a tenth of daily searches. The company processed more queries in a single day than the entire search industry had handled in a year during the late 1990s. The competitive moat was not merely technological; it was epistemological. Google had become the default interface through which humanity accessed its own collective knowledge. To "google" something was to search the internet. The verb was the company.
The Adult Supervision and Its Discontents
The question of who actually runs Google — and what it means to run a company whose founders retained permanent voting control — has been the central organizational drama of the company's existence.
The first act was the hiring of Eric Schmidt. In 2001, at the insistence of venture capitalists John Doerr of Kleiner Perkins and
Michael Moritz of Sequoia Capital, who had co-led Google's $25 million Series A in June 1999, Page and Brin agreed to bring in an experienced CEO. Schmidt was 46, a PhD in computer science from Berkeley, former CTO of Sun Microsystems, and CEO of Novell — a résumé that combined genuine technical credibility with the operational experience that two twenty-something founders conspicuously lacked. His role, as he described it, was to provide "adult supervision" while Page and Brin provided the vision.
The Schmidt era — 2001 to 2011 — was the period in which Google became Google: the IPO, the acquisition of YouTube, the launch of Android, the development of AdSense and the broader advertising platform, the expansion from a search company into an internet conglomerate. Schmidt was an effective operator and an even more effective diplomat, navigating Wall Street, Washington, and the increasingly complex internal politics of a company growing from 200 employees to more than 30,000.
But Page chafed. The Tesla biography haunted him. He did not want to be the inventor who let someone else build the institution. In January 2011, Google announced that Page would return as CEO, with Schmidt moving to the role of executive chairman. The transition was framed as a natural evolution. It was also a power grab — Page reasserting control over the company he had cofounded, restructuring the management hierarchy to report directly to him, and setting the stage for a period of aggressive, founder-driven expansion that would culminate in the creation of Alphabet.
Page's second tenure as CEO — 2011 to 2015 — was marked by a particular kind of ambition: the conviction that Google's search advertising cash flows should fund not just internet services but civilizational-scale technology bets. Self-driving cars (which became Waymo). Life extension research (Calico). Smart home devices (Nest, acquired for $3.2 billion in January 2014). High-altitude internet balloons (Project Loon). Delivery drones (Project Wing). The X research lab, which Page described as a factory for "moonshots."
The tension was structural. Google's advertising business was a machine of extraordinary precision and efficiency, optimized for quarterly revenue growth and operated by thousands of engineers focused on extracting maximum value from every search query. The moonshots were the opposite: speculative, long-horizon, cash-consuming ventures with uncertain — perhaps nonexistent — commercial payoffs. Running both under a single corporate umbrella created a kind of organizational schizophrenia.
Steve Jobs, in one of his last conversations with Page in 2011, had warned him about exactly this. "The main thing I stressed was focus," Jobs told his biographer Walter Isaacson. "Figure out what Google wants to be when it grows up. It's now all over the map. What are the five products you want to focus on? Get rid of the rest, because they're dragging you down. They're turning you into Microsoft."
I always thought it was kind of stupid if you have this big company, and you can only do, like, five things.
Page's answer was not to focus but to restructure. On August 10, 2015, he announced the creation of Alphabet Inc. — a new holding company under which Google would become a subsidiary, led by Sundar Pichai, while the moonshots and investment arms would operate as independent entities, each with its own CEO reporting to Page. The name was deliberately uncatchy. "I wanted to have a name that people would be proud to work for," Page explained. "But actually didn't want it to be too catchy because the idea really wasn't to have a consumer brand in the way that Google is."
The restructuring served multiple purposes simultaneously. It gave investors transparency into Google's core economics by separating the profitable advertising business from the cash-burning "Other Bets." It gave each moonshot operational independence and — theoretically — the discipline of functioning as a standalone company. And it gave Page what he wanted most: the freedom to think about the next century's technological possibilities without being tethered to quarterly search ad metrics.
In February 2016, the strategy was vindicated by the market. Alphabet reported strong quarterly results that, for the first time, broke out the financials of the non-Google businesses. Investors, seeing the core Google advertising machine stripped of the moonshot drag, pushed the stock up sharply. On February 1, 2016, Alphabet surpassed Apple to become the most valuable company in the United States, with a market capitalization exceeding $554 billion. Apple, which had held the top spot for most of the previous four years, was reeling from the slowest-ever increase in iPhone shipments.
The Pichai Inheritance
Sundar Pichai became CEO of Google on October 2, 2015, and CEO of both Google and Alphabet on December 3, 2019, when Page and Brin announced their retirement from day-to-day management. He was 47. The promotion made him one of the most powerful executives in the world, and also one of the most constrained — a non-founder CEO of a founder-controlled company, operating a machine built by two people who retained the voting power to override any decision he made.
Pichai's biography is one of those Silicon Valley immigrant narratives that doubles as a parable about meritocracy's possibilities and its structural requirements. Born in Madurai, India, in 1972, he grew up in a two-room apartment. His father was an electrical engineer at the British conglomerate GEC; his mother was a stenographer. The family did not own a car until Pichai was twelve. He studied metallurgical engineering at IIT Kharagpur — one of the ferociously competitive Indian Institutes of Technology — then earned a master's in materials science from Stanford and an MBA from Wharton. He joined Google in 2004, the year of the IPO, and rose through the ranks by shipping products that mattered: the Chrome browser, Chrome OS, and eventually overseeing Android and the entire Google apps ecosystem.
His management style was the inverse of Page's. Where Page was visionary and impatient, Pichai was deliberate and consensus-driven. Where Page thought in decades, Pichai thought in product cycles. Where Page was willing to burn cash on civilizational bets, Pichai was focused on execution and operational discipline. He was, in the framing that tech companies use with either admiration or condescension depending on the speaker, a "product guy."
The inheritance was vast and complicated. Pichai took over a company with 55,000 employees (now significantly larger), dominant market positions in search, mobile operating systems, browsers, video, email, and cloud infrastructure, and a set of organizational challenges that had been accumulating for years. The culture that had once encouraged radical openness — the weekly TGIF all-hands meetings where any employee could question any executive, the mailing lists where engineers debated everything from product direction to geopolitics — was straining under the weight of the company's size, its political visibility, and an increasingly fractious workforce.
The Google Walkout of November 2018, in which more than 20,000 employees protested the company's handling of sexual harassment allegations against senior executives, was a turning point. The subsequent period saw a gradual but unmistakable shift in corporate culture: fewer open forums, stricter workplace guidelines, and — in the view of some employees — a suppression of the internal dissent that had once been a point of corporate pride. The firing of Timnit Gebru in December 2020, the co-lead of Google's Ethical AI team, after a dispute over a research paper that examined biases in large language models, crystallized the tension. More than 1,200 Google employees and 1,500 academic researchers signed letters of protest. The incident revealed a company caught between its self-image as an open intellectual community and its institutional imperatives as a corporation with $150 billion in revenue and profound exposure to regulatory and reputational risk.
Pichai's response was characteristically measured. He acknowledged mistakes, expressed regret, committed to process improvements. He did not reverse course.
The Paper That Invented and Then Nearly Killed the Company
There is a deep irony — one that borders on the cosmically absurd — in the fact that the technology underpinning ChatGPT, the product that posed the first genuine existential threat to Google Search in two decades, was invented at Google.
The 2017 paper "Attention Is All You Need," authored by a team of eight Google Brain researchers, introduced the transformer architecture — the 'T' in GPT. Transformers revolutionized machine learning by enabling models to process entire sequences of text simultaneously rather than sequentially, dramatically improving both speed and the ability to capture long-range relationships in language. The paper was one of the most cited in the history of computer science. Six of the eight authors subsequently left Google to found or join AI startups.
Google's AI research pedigree was, by any reasonable measure, unmatched. The company had acquired DeepMind in 2014 for approximately $500 million, gaining what was arguably the world's premier AI research lab. Google Brain, the internal deep learning team, had been doing foundational work since 2011. The company had built custom AI chips — Tensor Processing Units, or TPUs — since 2015, giving it hardware-level advantages that no other company except perhaps Nvidia could match. Pichai had declared Google to be an "AI-first company" in 2016, seven years before the rest of the industry caught up.
And yet when OpenAI released ChatGPT on November 30, 2022, it was Google that scrambled. The product demonstrated, in a format that any consumer could grasp, that a large language model could synthesize information from across the internet and deliver it in conversational prose — not as a list of blue links, but as a direct answer. For a company that had built a $160 billion search-and-advertising business on the premise that the best way to organize information was to rank web pages by relevance, this was not merely a competitive threat. It was a philosophical challenge to the product paradigm that had generated essentially all of Google's revenue for a quarter century.
We've taken a full, deep, full-stack approach to AI. And that really plays out.
The internal response, by multiple accounts, was frantic. A "code red" was issued. Product timelines were accelerated. Bard, Google's initial ChatGPT competitor, was rushed to market and debuted in February 2023 with an embarrassing factual error in its demo that wiped $100 billion from Alphabet's market capitalization in a single day. The incident crystallized the innovator's dilemma facing Google: move too slowly and lose the AI race; move too fast and undermine the search advertising business that generated the cash to fund everything else.
The deeper problem was structural. Every AI-generated answer that appeared at the top of a Google search result — what the company eventually called "AI Overviews" — potentially reduced the number of links a user clicked, and therefore the number of opportunities to serve ads. Google's entire economic model was built on the click. Summarizing the answer eliminated the click. It was, as one analyst put it, a business model that was being asked to eat itself.
The Sleeping Giant Stirs
What happened next — across 2023, 2024, and especially 2025 — was one of the most remarkable competitive recoveries in modern technology history. Google did not panic its way to victory. It ground its way there, leveraging the full-stack advantages that no other company on Earth could replicate.
The key advantage was integration — the fact that Google controlled every layer of the AI stack, from the silicon to the model to the application. Its TPU chips, now in their seventh generation (Ironwood, unveiled in November 2025), provided an alternative to Nvidia's dominant GPUs. Its data centers — purpose-built over two decades and now absorbing approximately $75 billion in capital expenditure for 2025, with an anticipated $175 to $185 billion planned for 2026 — gave it computing infrastructure at a scale that only Microsoft and Amazon could approach. Its search index, Android phone telemetry, YouTube video corpus, and Gmail communications provided training data of a breadth and depth that OpenAI, for all its fundraising prowess, simply could not match. And its 750 million-plus monthly active Gemini users gave it a real-time feedback loop for improving models at consumer scale.
The Gemini model family — Google's answer to GPT — evolved rapidly. Gemini 2.0 was integrated into AI Overviews across more than 100 countries. Then, in December 2025, Google released Gemini 3 to what analysts described as "rave reviews," praising its capabilities in reasoning, coding, and niche tasks that had tripped up rival chatbots. OpenAI reportedly declared a "code red" — its highest alert level — to improve ChatGPT in response. SoftBank, one of OpenAI's biggest backers, fell to a two-month low on the news.
Sergey Brin, who had stepped away from active management in 2019 expecting to spend his retirement studying physics in cafés, found himself drawn back. When COVID shuttered the cafés in early 2020, he returned to the Google campus and began working on what would become Gemini. By 2023, he was visiting the office three to four times a week, holding weekly discussions with AI researchers, and reportedly influencing personnel decisions. In February 2025, he made waves with an internal memo recommending that Google employees go into the office every weekday and suggesting that sixty hours a week was the "sweet spot" of productivity — a dramatic departure from the flexible work culture that Google had pioneered.
To be able to have that technical creative outlet, I think that's very rewarding. If I'd stayed retired, I think that would've been a big mistake.
The market responded. Alphabet's stock added nearly $1 trillion in market capitalization between mid-October and late November 2025, aided by Warren Buffett's Berkshire Hathaway taking a $4.9 billion stake during the third quarter — a remarkable endorsement from an investor who had historically avoided technology companies. In January 2026, Alphabet's market cap surpassed Apple's for the first time since 2019, closing at $3.88 trillion versus Apple's $3.84 trillion. Days later, on January 12, 2026, Apple announced that Google's Gemini would be the foundation for its artificial intelligence models and the next generation of Siri. Alphabet's stock popped to $331.86, putting its market capitalization just over $4 trillion.
The symbolism was exquisite. The company that Steve Jobs had warned was "all over the map" had become the platform on which Apple would build its AI future.
The Cloud That Came In Third
For years, Google Cloud was the punchline of the hyperscaler wars — a distant third behind Amazon Web Services and Microsoft Azure, unable to translate Google's engineering prowess into enterprise sales relationships. The problem was cultural. Google had built its infrastructure for its own engineers, optimized for its own workloads, governed by its own aesthetic of mathematical elegance. Enterprise customers, who cared about support contracts and compliance certifications and salespeople who returned phone calls, found the experience baffling.
The turnaround was slow, expensive, and ultimately powered by AI. Google Cloud grew its Q4 2025 revenue by 48% year-over-year to $17.7 billion, ending the year at an over-$70 billion annual run rate. The growth was driven by demand for AI infrastructure — Google Cloud Platform, the subset with the strongest AI component, grew faster than the overall cloud business. On Alphabet's October 2025 earnings call, Pichai reported that Google's cloud business had signed more deals over $1 billion through the first three quarters of 2025 than in the two prior years combined.
The AI demand was so intense that Google couldn't build capacity fast enough. CFO Anat Ashkenazi, who had joined from Eli Lilly in July 2024, explained on the Q4 2024 earnings call that the cloud growth slowdown (from 35% to 30% that quarter) was partially attributable to supply constraints: "Google doesn't have enough capacity in its data centers to meet all the demand." Hence the $75 billion capex budget for 2025, including eleven new data centers — and the staggering $175 to $185 billion planned for 2026.
Citi analysts named Google a top internet pick for 2026, noting that 70% of Google Cloud customers were already using its AI products. "Google has the chip, the infrastructure capacity, and the model amid growing demand." The cloud business, once an also-ran, had become the second growth engine — and the vehicle through which Google's AI research was being monetized beyond search.
The Antitrust Shadow
The U.S. Department of Justice's antitrust case against Google, filed in October 2020, was the most significant competition enforcement action against a technology company since the Microsoft case of 1998. The core allegation: that Google had maintained its monopoly in search through a network of exclusionary agreements — paying Apple an estimated $26 billion annually to be the default search engine on iPhones and Safari, and leveraging its ownership of Android and Chrome to ensure its search engine was the default on virtually every digital surface.
In August 2024, Judge Amit Mehta ruled that Google had indeed violated antitrust law, finding that the company had maintained an illegal monopoly in general search and search text advertising. The remedy phase — what, if anything, the government could force Google to do about it — became the central legal drama of 2025.
The most severe scenario, feared by investors and desired by competitors, was a breakup: the forced divestiture of Chrome, Android, or even the advertising technology stack. By late 2025, however, Google had avoided the most extreme outcomes, in part because the rapid emergence of AI competition — ChatGPT, Perplexity, Anthropic's Claude — made the argument that Google's search monopoly was impervious to challenge significantly less convincing to the court.
The irony was rich. The competitive threat that Google most feared — the one that had triggered code reds and pulled Sergey Brin out of retirement — may have been the thing that saved it from being dismembered by its own government.
Waymo and the Long Bet
Among the "Other Bets" that justified the Alphabet structure, Waymo stands alone — both as the most capital-intensive long-duration investment in the company's history and as the one that, by late 2025, had most clearly crossed from speculative moonshot to commercially viable business.
Originally Google's self-driving car project, launched within X in 2009 and spun out as Waymo in 2016, the unit had consumed billions in R&D over more than a decade with essentially zero revenue. By 2025, Waymo was operating autonomous taxi services in multiple cities and had added freeway driving to its capabilities — a technical milestone enabled by the same AI research infrastructure that powered Gemini.
The unit was not yet profitable. But it was operational at a scale that no competitor — not Tesla's Full Self-Driving, not Cruise (which GM shuttered after a 2023 safety incident), not any Chinese rival — had matched in the United States. When Pichai described Alphabet as a company with "different businesses with different time horizons," Waymo was the business he was talking about. The question was whether a decade of patience and billions in investment would produce a transportation platform of sufficient scale to justify the cost — or whether it would remain, as skeptics suggested, the world's most expensive science fair project.
The Cathedral and the Cash Register
The tension that has defined Google from its founding — between the altruistic mission to organize the world's information and the commercial reality of monetizing attention — has never resolved. It has simply grown more complex, more consequential, and more expensive.
When Larry Page told the Financial Times in October 2014 that "the societal goal is our primary goal" and that even Google's mission statement was no longer big enough for what he had in mind, he was expressing something genuine about the founders' ambition. The company was, and in some sense still is, an institution that believes it can solve problems — climate change, transportation, disease, the structure of human knowledge itself — through the application of engineering talent and exponential computing power.
But the cash register never sleeps. In Q4 2025, "Google Search & other" — the direct descendant of the PageRank algorithm, the product of two graduate students who thought advertising was corrupting — generated 56% of Alphabet's $96.5 billion in quarterly revenue, which itself grew to $113.8 billion in the following quarter. The advertising auction runs continuously, billions of times per day, extracting value from human curiosity at a scale that makes the founding paper's concerns about bias seem almost quaint.
Alphabet's operating margin held at 32% in fiscal 2025. The company earned $132.2 billion in net income, generated over $325 million paid subscriptions across its consumer services, and deployed capital at a rate — $75 billion in 2025 capex, $175 to $185 billion planned for 2026 — that exceeded the
GDP of most nations. When Pichai told investors that "search saw more usage than ever before, with AI continuing to drive an expansionary moment," he was describing not just a product metric but the continued vitality of the most profitable attention monopoly ever constructed.
The company that was not a media company now owns the world's largest video platform. The company that would not run banner ads now operates the most sophisticated advertising infrastructure on Earth. The company founded by two people who believed search should be free of commercial influence now pays Apple $26 billion a year to remain the default — a payment that is simultaneously the price of maintaining the monopoly and evidence that the monopoly requires maintenance.
On the S-1 that started it all, the registration fee for $2,718,281,828 in proposed securities was $344,406.31. The company is now worth approximately 1.5 million times that offering price. e raised to the power of ambition, compounding continuously.
