Google is betting that artificial intelligence can secure its future without destroying the search advertising business that made it one of the world’s most valuable companies.
Alphabet, Google’s parent company, is pouring vast sums into AI while redesigning its search engine to provide direct answers rather than a list of links. The move could help it fend off rivals, but it also risks reducing the value of each search query to advertisers.
The scale of the challenge was underlined when Alphabet’s shares fell by more than 12% in a single day after its latest earnings report in late October – the company’s sharpest one-day decline since it went public in 2004.
Executives acknowledged that the growing use of the chatbot-style AI Mode was contributing to slower growth in search advertising revenue because it generated less money per query than traditional search. At the same time, Alphabet said it expected to spend about 200 billion dollars this year on infrastructure, including data centres, largely to support its AI ambitions.
Google’s AI dilemma
The predicament reflects a long-standing business problem: an established company can be reluctant to embrace a technology that threatens its most profitable existing product, even as rivals use that technology to challenge its position.
Google built its fortune by placing adverts alongside search results. Its dominance became so extensive that “Google” entered everyday language as a verb, while its share of the US search market rose from 80% in 2009 to nearly 90% in 2020.
Former Google executive Marissa Mayer joined the company in 1999, when it had only a handful of employees. She recalled that she became its 20th employee after interviewing at a business with just seven people and said the company was generating slightly more than 200,000 dollars in annual revenue that year.
By 2022, Alphabet was reporting annual revenue of more than 280 billion dollars. Google had transformed its search success into a powerful financial engine, using the proceeds to fund research into technologies that would eventually threaten the engine itself.
In 2017, eight Google employees developed the transformer technology that made modern chatbots possible. The company had also worked on language translation, speech and image recognition, autonomous vehicles and an internal chatbot called Meena.
But Google did not release Meena, while OpenAI’s ChatGPT was launched publicly on 30 November 2022. Its rapid popularity forced Google to respond, reportedly triggering a “code red” inside the company.
Google chief scientist Jeff Dean later said the company had judged its chatbots against the standards of search, where users expect answers to be accurate every time. The systems could make up information or produce offensive responses, he said, and were not close enough to that standard.
ChatGPT had similar shortcomings, but users embraced it for tasks beyond conventional searching, including drafting messages and summarising documents. Google had underestimated the wider range of uses for conversational systems, Dean acknowledged.
Mayer has questioned whether being first would necessarily have benefited Google. An established company faces different expectations about quality and reliability from a new entrant, she said.
Google has since launched its own chatbot, Gemini, and begun placing AI-generated answers directly on its search page. That approach could protect its relevance as people turn to rival systems, but it may also reduce the number of links users click and the opportunities to display adverts.
Could AI loosen Google’s grip?
Technology analyst Benedict Evans believes AI represents the latest major platform shift, following the personal computer, the internet and the smartphone. Such changes can weaken the advantages that allowed an incumbent to dominate.
Google’s position was strengthened by a feedback loop: more users generated more data, helping the company improve its results and attract still more users. Its reach was also reinforced through products including email, documents, maps, video and the operating system used on many smartphones.
The company paid to make its search engine the default on browsers and devices, deals that a federal judge ruled in 2024 had illegally helped preserve its search monopoly. But a rival AI assistant could provide an alternative route to information, bypassing Google’s main search page altogether.
It is not yet clear whether AI will produce one dominant winner or support several large companies. The cost of computing power and energy creates a significant barrier to entry, but Mayer expects “multiple big players” rather than a single controlling company.
She believes Google could remain one of those major players even if its share of search declines. A smaller share of a much larger AI-driven market, she argues, could still deliver greater revenue.
Evans points to Microsoft as an example of a company that lost dominance after a technological upheaval but continued to grow. The internet weakened Windows’ strategic advantages, yet the expansion of the computer market helped Microsoft sell more Windows and Office software.
That outcome, however, required patience. Microsoft’s share price did not overtake its dotcom-era peak for almost 17 years.
Who will profit from the AI race?
Google’s challenge extends beyond protecting search. Investors are also questioning whether the businesses developing and operating advanced AI models will capture enough value to justify their enormous costs.
Evans warns that competing models could become so similar that their creators struggle to charge high prices. The biggest profits might instead go to companies that build products and services on top of the models.
He compares the risk with mobile networks, where telecommunications companies spent heavily on infrastructure while smartphone makers and app developers captured much of the value created by the mobile internet.
Google is better placed than many rivals to benefit from different versions of the AI economy. It can sell access to its models and computing capacity, while also using the technology in search, assistants and other products. Its businesses include cloud computing, YouTube and the driverless-car company Waymo.
The company also supplies computing power to other businesses, including Anthropic, in which it is a major investor. But it must still show that those opportunities can compensate for the cost of developing AI and any decline in the profits generated by traditional search.
Google chief executive Sundar Pichai has rejected the idea that the company must choose between protecting its existing business and pursuing innovation. “I think the dilemma only exists if you treat it as a dilemma,” he said, arguing that technology companies must embrace periods of major change as forcefully as possible.
Mayer believes the company is using AI in a way that is consistent with its historic purpose: answering questions. The harder problem, she said, is persuading advertisers to move their spending into formats that work within an AI-led service.
Google said its AI features were already contributing to increased activity on its platform, with search queries growing at their fastest year-on-year rate in five years. It said it was taking a deliberate approach by prioritising the user experience before expanding monetisation.
Whether advertising in AI can become as profitable as search remains unresolved. Google helped create the technology driving the latest transformation; it is now wagering that the same technology can produce new sources of revenue before its old cash cow is exhausted.
