The Company That Forgot What It Was
It was January 2007. Rubin, the father of Android, was watching Steve Jobs unveil the iPhone. On stage, Jobs was swiping through photos with his finger, pinching to zoom, scrolling with impossible smoothness. The crowd gasped.
Rubin sat in silence. His team had been building Android for two years. They’d designed it around physical keyboards and styluses. Everything—the entire operating system architecture—was wrong.
Google had bought Rubin’s tiny startup in 2005 for $50 million, back when the iPhone didn’t exist. Their target wasn’t Apple. It was Microsoft.
Then Jobs changed everything in one keynote.
Sixteen years later, history would repeat. Another shock. Another moment of reckoning. This time, the enemy wasn’t a device. It was a chatbot.
The Question No One Asked
But here’s what almost no one understood about Google, even as it grew into one of the world’s most powerful companies: What kind of company is Google?
Most people would say: The world’s best search engine. The world’s biggest advertising company. The leader in AI and machine learning. The most engineer-obsessed tech company on Earth.
All true. All missing the point.
Understanding Google requires understanding three moments when the company forgot its own identity—and what happened each time it remembered.
Two Different Games
Think about Apple. Everyone knows Apple makes money from hardware—iPhones, Macs, premium devices with premium prices. Apple serves a specific customer segment. It’s a zero-sum game: every person who buys a Samsung phone is a customer Apple lost.
For Apple, profit margin matters more than market share. They chase the high end, abandon the low end, maximize returns from people willing to pay for the Apple brand.
Google plays an entirely different game.
Google is a services company. Everything—search, Gmail, YouTube, Maps, Chrome, Android—serves one purpose: get as many people as possible using Google services. More users mean lower fixed costs per user and more data to power better products.
Look at the numbers. Apple’s costs scale with device sales—materials, supply chain, manufacturing. Google’s costs are largely fixed—engineers, data centers, servers. Apple can adjust spending based on sales. Google needs maximum users to spread fixed costs across the largest possible base.
This is Google’s essence: A company that provides universal services.
Which means Google’s products should work everywhere, for everyone, on every device.
But three times, Google forgot this. Each time nearly killed a major initiative.
First Mistake: The Fatal Feature (2009-2012)
In 2009, Google engineers built something revolutionary for Google Maps: turn-by-turn navigation. The feature seems obvious now—your phone tells you exactly when to turn, which lane to use, how to navigate complex intersections. Standard stuff.
But Google invented it first.
And they made it Android-exclusive.
If you used an iPhone, you got an inferior version of Google Maps. Missing features. Less functionality. Google was competing with Apple, after all. Why give the enemy your best weapons?
For years, Google thought this made sense.
Then in 2012, Apple struck back. iOS 6 launched, and YouTube disappeared from the pre-installed apps. Apple also replaced Google Maps with its own mapping software as the default.
Apple Maps was terrible. Everyone knew Google Maps was superior. But within months, Apple Maps’ user numbers caught up to Google Maps on iOS.
That’s when Google realized their mistake.
By making features Android-exclusive, Google was rejecting potential users. They were betraying their own business model. Google doesn’t make money selling HTC phones. Google makes money when people use Google services—regardless of device.
The lesson was brutal but clear: When your business depends on universal access, platform loyalty is suicide.
After 2012, Google changed everything. New products launched everywhere simultaneously. Features optimized for iOS users. Google even made experiences better on Apple devices in some cases.
Because Google finally understood what it was.
The Secret Weapon They Hid Too Long
While Google was fixing Android, they were sitting on another mistake—one that had been festering for over a decade.
Back in 1999, Urs Hölzle became Google’s eighth employee. A Swiss-American computer scientist, Hölzle faced an impossible challenge: “We were growing at something like ten percent week after week,” he recalled. “If you just barely made it or if you melted down last week and next week is going to be ten percent bigger, that’s going to be what you focus on.”
His solution revolutionized computing. Rather than buying expensive enterprise hardware, Hölzle pioneered building clusters from commodity servers—creating Google’s internal cloud infrastructure called Borg.
This system became Google’s most closely guarded secret. It powered Search, Gmail, and YouTube while competitors struggled to scale.
But here’s the mistake: Google was too protective.
Amazon Web Services launched in 2006. By 2013, AWS was generating billions in revenue, building the future of enterprise computing. Google had the better technology—by far—but treated it like a state secret.
That summer, three Google engineers—Craig McLuckie, Joe Beda, and Brendan Burns—pitched Hölzle on a radical idea: open-source an external version of Borg.
Hölzle’s response: “So let me get this straight. You want to build an external version of the Borg task scheduler. One of our most important competitive advantages. The one we don’t even talk about externally. And, on top of that, you want to open source it?”
He approved it anyway.
The project became Kubernetes. When Niantic launched Pokémon Go in July 2016 running on Kubernetes, the game saw 50 times more users than expected. Most systems would have crashed immediately. Kubernetes handled it beautifully.
But Google Cloud Platform itself? Still struggling. By 2015, analyst John Dinsdale noted: “Until recently Google never gave the impression (through words or deeds) that cloud services were really important to it.”
Second Mistake: The Enterprise Blind Spot (2008-2018)
Google acquired a startup called Bebop in November 2015 for $380 million. The real prize wasn’t the product. It was the founder: Diane Greene, the legendary VMware co-founder.
Her mission: transform Google Cloud from “having only two significant customers and a collection of startups to having major Fortune 1000 enterprises betting their future on Google Cloud.”
Greene built an enterprise sales team from scratch, grew the Google Cloud Next conference from 2,300 attendees to 23,000+ in just two years, and won marquee customers including Spotify ($450 million over three years) and Snap ($2 billion over five years).
But Google Cloud remained deeply unprofitable and stuck below 10% market share. Greene departed in November 2018.
Her replacement would transform everything: Thomas Kurian, Oracle’s President of Product Development with 22 years at the company.
Born in Kerala, India, Kurian had dropped out of IIT Madras after six months before going to Princeton and Stanford. At Oracle, he led a 35,000-person development team with a $4 billion R&D budget.
At Google, Kurian brought enterprise DNA to a consumer-focused culture. His blunt assessment to the Wall Street Journal: Google’s cloud sales team was “one-tenth to one-fifteenth the size” of AWS and Microsoft Azure’s sales forces.
“We shifted the organization from thinking, we’re building technology to we’re building products and solutions,” Kurian explained.
The results: Revenue grew 45% in his first year. Market share rose by 10 percentage points. Deals over $250 million doubled year-over-year.
Most importantly: After cumulative losses exceeding $14 billion from 2020-2022, Google Cloud reported its first-ever profit in April 2023. By Q3 2025, operating income hit $3.6 billion on $15.2 billion quarterly revenue.
Kurian had fixed Google’s second identity crisis: Google could compete in enterprise, but only if they built for enterprise needs instead of assuming consumer playbooks would work.
By 2025, Google Cloud represents 13-15% of Alphabet’s revenue—down from over 80% advertising dependence. The business now exceeds $50 billion annual revenue, growing 34% year-over-year, faster than Azure (30%) and AWS (18%).
The company that hid Borg for a decade finally learned: universal access means building for enterprises too.
Third Mistake: The $100 Billion Wake-Up Call (2023)
February 8, 2023. Google’s stock plummeted, erasing $100 billion in market value in a single day.
The trigger? A demo video where Google’s AI chatbot Bard confidently stated a wrong answer about the James Webb Space Telescope.
But the real problem wasn’t one error. It was that Google had been caught flat-footed by OpenAI’s ChatGPT, which had become the fastest-growing app in history just months earlier.
The company that invented the transformer architecture—the foundation of modern AI—was now playing catch-up.
Internally, Google declared a “code red.” Co-founder Sergey Brin returned from semi-retirement to write code. Jeff Dean, Google’s most senior AI executive, did the same.
And CEO Sundar Pichai made a bet that would define his tenure.
The Irony of Competition
Here’s what made the situation surreal: Microsoft’s massive investment in OpenAI was driven by fear of Google.
A 2019 email from Microsoft CTO Kevin Scott to CEO Satya Nadella and Bill Gates, revealed in DOJ antitrust documents, warned: “I got very, very worried when I looked at the AI model-training capability gap between Google and us. We are multiple years behind the competition in terms of ML scale.”
That email, titled “Thoughts on OpenAI,” directly led to Microsoft’s $13+ billion investment.
The competitive dynamic: Microsoft feared Google’s AI lead, which paradoxically created the ChatGPT threat that forced Google to accelerate.
“Contrary to belief, I was pleased when ChatGPT launched,” Pichai would later admit. “We had been building this tech for so long, we were so AI native. We hadn’t quite gotten it to the level where you could put it out and people would’ve been okay with Google putting out that product.”
The Bold Response: Merging the Rivals
In April 2023, Pichai made a historic move: merging Google Brain with DeepMind (acquired for $400 million in 2014) into Google DeepMind, led by Demis Hassabis.
The two labs had been rivals. But building a GPT-4 competitor required combining DeepMind’s reinforcement learning expertise with Google Brain’s massive-scale infrastructure.
The project attracted Google’s elite. Brin became a “core contributor.” Dean wrote code himself. The pressure was immense—Google’s $60 billion search business hung in the balance.
December 6, 2023. Google unveiled Gemini. The benchmarks were impressive: Gemini Ultra beat GPT-4 on 30 of 32 benchmarks, including becoming the first model to exceed human expert performance on MMLU (90% vs. 89% human baseline).
“This is a significant milestone in the development of AI, and the start of a new era for us at Google,” Hassabis announced.
But alongside the announcement, Google released a demo video called “Hands-on with Gemini” showing fluid, real-time voice interactions and instant gesture recognition.
Bloomberg’s Parmy Olson exposed the truth: the video was almost entirely staged. There was no real-time interaction, no voice interaction (text prompts only), and still images were used instead of video.
TechCrunch’s verdict: “Google has poisoned the well. How can anyone trust the company when they claim their model does something now?”
Inside Google, employees reportedly called the video “underwhelming” and worried it made Gemini “look more advanced than it actually is.”
The controversy damaged trust precisely when Google needed to rebuild credibility.
The Comeback: From Controversy to Competitive Parity
Despite the rocky launch, Gemini evolved rapidly.
By August 2024, Gemini 1.5 Pro experimental scored 1,300 on the LMSYS Chatbot Arena leaderboard, beating GPT-4o (1,286) and Claude-3 (1,271).
By March 2025, Gemini 2.5 Pro debuted at #1 on LM Arena and has maintained top rankings for 6+ months. The model offers the largest context window in production (1-2 million tokens—equivalent to 700,000 words, 11 hours of audio, or 1 hour of video).
Google’s integration strategy proved powerful. Gemini now runs across Search, Gmail, Docs, Sheets, Android, and Chrome. The Gemini app reached 650 million monthly active users by Q3 2025.
Google reports that nearly half of all new code at Google is now generated by AI, reviewed and accepted by engineers.
Over 70% of existing Google Cloud customers now use AI products. 4.4 million developers are building with Gemini models. The company signed more deals over $1 billion through Q3 2024 than the previous two years combined.
Google had learned the third lesson: When you create breakthrough technology, ship it everywhere immediately—or someone else will.
The Technical Advantage: What Google Does Differently
While competitors depend on Nvidia GPUs, Google deploys 9 custom TPUs for every 1 Nvidia GPU—the highest ratio among hyperscalers.
Constellation Research analyst Holger Mueller notes: “Google has a 3- to 4-year lead when it comes to putting custom algorithm (TensorFlow) on custom hardware (TPUs). It now has also a 1+ year lead for operating multimodal models.”
This infrastructure advantage translates to competitive positioning. Google Cloud is gaining market share, rising from 6% in Q4 2017 to 13% in Q4 2024 and approximately 13-15% by mid-2025. The company now has a $155 billion backlog of contracted future revenue.
The Original Sin (That Made Everything Possible)
Google’s success story actually starts in the late 1990s. Yahoo dominated search. But two Stanford PhD students, Larry Page and Sergey Brin, built something better.
But Google’s real advantage wasn’t just technology. It was accessibility.
The Google homepage: one line of text (“Google”), one search box, nothing else. No download required. No installation. No payment. Just type google.com and get instant access to the world’s best search engine.
This was the PC internet era. The web was the only app store. Browsers were phones. Web pages were apps. Google rode the internet’s openness to ubiquity.
This became Google’s DNA: Great technology + universal accessibility = world-changing success.
Page said it explicitly in Google’s early days: “The ultimate search engine will be smarter than any human. To build that, we need the world’s best engineers working at the deepest level. When we succeed, everyone will use Google. There’s no room for second place.”
This philosophy worked brilliantly. For search. For Gmail. For Chrome. For YouTube.
Until it didn’t.
The Culture That Eats Strategy
When Google Glass failed in 2012-2014, it revealed the company’s deepest cultural blind spot.
The Glass team was filled with brilliant engineers. Astro Teller had previously built fitness-tracking wristbands. Babak Parviz had integrated digital displays into contact lenses. Stars, all of them.
Development went smoothly. Then the team split: Who is our customer?
Half believed Glass should be a consumer device. Half believed it should provide professional tools for specialized industries.
The question reached Brin. His answer: “The debate will resolve itself after launch. We’ll iterate based on user feedback.”
That’s when Google Glass was truly doomed.
Because Brin applied Google’s standard playbook: Hire the best engineers. Build the best technology. Ship to potentially billions of users. Iterate like crazy.
This formula built Google Search. Gmail. Chrome. YouTube.
But hardware doesn’t work this way. Hardware can’t be free. Hardware has high adoption barriers. Hardware iterates slowly.
Software can update overnight. Hardware requires new versions, marketing campaigns, retail distribution. By the time Google Glass could iterate, public opinion had already turned.
This is cultural inertia in action. Google’s DNA—formed through repeated software success—blinded them to hardware realities.
The Power of Mission
Still, Google gets one thing profoundly right: mission.
“Organize the world’s information and make it universally accessible and useful.”
May 2002. Larry Page browsing online, searching for “Kawasaki H1B”—a Japanese motorcycle. Results: Nothing but law firm ads. (H-1B is also a U.S. work visa code.)
He didn’t schedule meetings. Didn’t assign blame. Didn’t form committees.
Page printed the bad results, marked them up, pinned them to the break room bulletin board next to the pool table, and wrote in large letters: “THESE ADS SUCK.”
The following Monday, 5 AM, Page received an email from five engineers. They’d spent their weekend building a new solution—complete with theoretical explanations and testing links.
This system became AdWords. It generated hundreds of billions in revenue.
The critical detail: These five engineers weren’t responsible for ads. They just felt responsible for fixing the problem.
Google’s project evaluation criteria emerged from this culture:
- Does this address a problem affecting a billion people?
- Is the solution radical enough?
- Does it use breakthrough technology?
All three required. Projects meeting these standards attract the best talent automatically.
The Principles That Work
On hiring: “A-players hire A-players. B-players hire C- and D-players.”
Corollary: Better to not hire than to compromise.
The airport delay test: Only hire people you’d happily spend six hours with if your flight got delayed.
Brin’s favorite interview question: “Can you teach me something complicated I don’t know?”
One candidate taught Brin how to bake cakes. Revealed life philosophy and learning mindset.
On speed: “When you make a decision matters more than whether the decision is right.”
Most decisions aren’t life-or-death and irreversible. Information gathering and endless debate waste more value than imperfect decisions.
Google’s CEO Eric Schmidt lived this principle. He set decision deadlines for every major choice. That clarity unlocked speed.
The meta-decision: Before deciding anything, decide when you’ll decide.
The Essential Nature—Evolved
So what is Google?
Strip away the products, the technology, the famous campus, the free food, the brilliant engineers.
Google is now three companies in one:
1. The Advertising Giant: Search and YouTube still generate $250+ billion annually—the cash engine funding everything else.
2. The Cloud Platform: Google Cloud Platform and Workspace serve enterprises worldwide, generating $50+ billion annually with 34% year-over-year growth.
3. The AI Company: Gemini powers products across the entire Google ecosystem while competing directly with ChatGPT and Claude.
But the essence remains: A company that provides universal services.
Everything else follows from this:
- Products must work everywhere, for everyone
- Low barriers to entry matter more than premium positioning
- Market share beats profit margins in the long run
- Openness beats exclusivity
- Universal accessibility is survival
When Google remembers this, they thrive: Search. Gmail. Maps. YouTube. Chrome. Android (after 2012). Google Cloud (after Kurian). Gemini (after the comeback).
When Google forgets this, they fail: Android exclusive features (2009-2012). Cloud’s consumer playbook (2008-2018). Gemini’s staged demo (2023).
The Unfinished Story
“The stakes are high,” Pichai told employees in December 2024. “I think 2025 will be critical. It’s really important we internalize the urgency of this moment, and need to move faster as a company.”
Google Cloud reached $50+ billion in annual revenue and profitability, growing faster than competitors. Gemini achieved technical parity with GPT-4 and Claude, integrated across Google’s entire ecosystem.
But critical questions remain:
Can Google overcome the “second place” perception when ChatGPT owns consumer mindshare?
Will the massive capital investments ($75-93 billion in 2025) deliver sufficient returns?
Can Gemini app reach the 500 million user target Pichai set?
The competition isn’t standing still. OpenAI continues evolving. Microsoft’s enterprise relationships remain formidable. AWS still commands 30% market share.
Three Lessons, Three Transformations
In 2007, Andy Rubin watched Steve Jobs unveil the iPhone and realized Android needed complete reconstruction.
In 2012, Apple Maps caught up to Google Maps overnight—forcing Google to abandon platform exclusivity.
In 2018, Thomas Kurian arrived and transformed a consumer-focused technology showcase into a profitable enterprise business.
In 2023, ChatGPT’s explosive growth shocked Google into merging rival AI labs and shipping Gemini everywhere.
Each time, Google forgot its identity. Each time, the mistake nearly killed a major initiative. Each time, Google recovered by returning to first principles: universal accessibility.
The technology changed. The mission never did.
Organize the world’s information. Make it universally accessible. Make it useful.
Simple. Enduring. Powerful.
As long as Google remembers who they are—and isn’t afraid to prove it through difficult pivots—they’ll continue organizing the world’s information long after we’re gone.
But now they do it three ways: through advertising-funded services, enterprise cloud platforms, and AI-powered experiences.
The company that forgot what it was—three times—now knows better than ever.
