Google, the sleeping giant in global AI race, now ‘fully awake’

Not anymore.

The internet giant has released new AI software and struck deals, e.g Connect chip With Anthropic PBC, which reassured investors that the company would not easily lose to ChatGPT creator OpenAI and other competitors. Google’s latest multi-purpose model, Gemini 3It won instant praise for its reasoning and programming abilities, as well as specialized tasks that stymied AI-powered chatbots. Google’s cloud business, which was once also managed, is growing steadily, thanks in part to a global rush to develop artificial intelligence services and demand for computing.

There are signs of rising demand for Google’s specialized AI chips, one of the few viable alternatives to Nvidia’s dominant hardware. A report on Monday that Meta Platforms Inc. Talks are underway to use Google’s chips to soar shares of parent company Alphabet Inc. The stock has added nearly $1 trillion in market value since mid-October, helped by Warren Buffett’s take A share worth $4.9 billion During the third quarter and broader Wall Street enthusiasm for its AI efforts.

Shares of Alphabet Inc, which owns Google, rose as much as 3.22% in New York on Tuesday. The company is on track to reach $4 trillion market cap for the first time.

SoftBank Group, one of the largest backers of OpenAI, It fell to its lowest level in two months on Tuesday due to concerns about competition from Google Gemini. Nvidia shares fell as much as 5.51% on Tuesday, losing $243 billion in market value.

“Google has arguably always been the dark horse in this AI race,” said Neil Shah, analyst and co-founder at Counterpoint Research. He is the “sleeping giant that is now wide awake.”

For many years, Google executives argued that deep, expensive research would help the company fend off competitors, defend its position as the leading search engine, and invent the computing platforms of the future. Then came ChatGPT, posing the first real threat to Google Search in years, even though Google pioneered the technology underpinning OpenAI’s chatbot. However, Google has a lot of resources that OpenAI does not: a pool of data ready to train and improve AI models; Streaming profits and their computing infrastructure.

“We have taken a complete, deep, integrated approach to AI,” Google and Alphabet CEO Sundar Pichai told investors last quarter. “And it really happens.”

Any concerns that Google could be stymied by regulators are fading. company recently Avoid the most serious outcome From an antitrust case in the United States – breaking up its business – due in part to the perceived threat from new entrants in the field of artificial intelligence. The search giant has shown some progress in long-term efforts to diversify beyond its core business. Waymo, Alphabet’s self-driving car unit, is coming to several new cities and has just added highway driving to its taxi service, a feat made possible by the company’s massive research and investments.

Some of Google’s advantages come from its economics. It is one of the few companies that produces what the industry calls the complete suite in computing. Google makes the AI ​​applications that people use, such as the popular Nano Banana image generator, as well as the software models, cloud computing architecture, and chips underneath. The company also has a goldmine of data to create AI models from search index, Android phones, and YouTube – data provided by Google. He often keeps to himself. This means, in theory, that Google has more control over the technical direction of its AI products and does not necessarily have to pay suppliers, unlike OpenAI.

Several technology companies, including Microsoft Corp., have planned and OpenAI, for ways to develop its own semiconductors or establish relationships that would make it less dependent on Nvidia’s best-selling products. For years, Google was effectively its sole customer for its home-grown processors, called tensor processing units, or TPUs, which the company first designed more than a decade ago to speed up the generation of search results and has since adapted to handle complex AI tasks. This is changing. Anthropic AI startup he said in October It said it would use up to 1 million Google thermal processing units in a deal worth tens of billions of dollars.

Monday, Technology Bulletin Information reported Meta plans to use Google chips in its data centers in 2027. Google declined to address specific plans, but said its cloud business is “accelerating demand” for both custom TPUs and Nvidia GPUs. “We are committed to supporting both, as we have for years,” a company spokesperson wrote in a statement.

Meta declined to comment on the report Monday night.

“We are pleased with Google’s success,” an Nvidia spokesperson said in a statement on Tuesday. “They have made significant advances in AI, and we continue to supply Google.” The spokesperson added: “Nvidia is a generation ahead of the industry – it is the only platform that runs every AI model and does it everywhere computing happens.”

Analysts read meta news as a sign of Google’s success. “Many others have failed in their quest to build custom chips, but it is clear that Google can add another string to its bow here,” Ben Barringer, head of technology research at Quilter Cheviot, wrote in an email.

Google took risks to get here. In early 2023, Google consolidated its AI efforts under the leadership of Demis Hassabis, head of its London AI lab DeepMind. The cabinet reshuffle witnessed some obstacles, the most notable of which is: Failed proposal Image generation product. For several years, DeepMind has pursued research in areas such as protein folding, which has led to new commercial strategies (and… Nobel Prize) but it contributed little to Google’s bottom line. Under the reorganization, the AI ​​unit is focused almost directly on foundational models that keep pace with OpenAI, Microsoft, and others.

Hassabis, a renowned computer scientist, has helped retain key AI engineers despite multi-million-dollar offers from competitors. His boss, Pichai, was as well Ready to splurge On talent.

The Gemini 3 Pro has risen to the top of the closely watched AI leaderboards in LMArena and Humanity’s Latest Exam. Andrei Karpathy, one of the founding members of OpenAI, He said It is “obviously a level 1 LLM,” referring to large language models. Google has pitched the model as one that can solve complex problems in science and mathematics, and tackle troublesome problems – e.g Generate images and overlaid text with incorrect spelling – which could prevent enterprise customers from adopting AI services more broadly.

Consumer interest is difficult to measure. Google said last week that 650 million people use its Gemini app. OpenAI recently said that ChatGPT has reached 800 million users weekly. As of October, the Gemini app had 73 million monthly downloads, far fewer than the 93 million monthly downloads of ChatGPT, according to research firm Sensor Tower.

Google is an advertising giant, but it has historically struggled to find other business models. Its cloud business posted third-quarter revenue of $15.2 billion, up 34% from a year earlier. However, this is still third behind Microsoft and Amazon Web Services, which reported more than double Google’s cloud sales last quarter. Counterpoint Research’s Shah said Google’s AI adoption with enterprises lags behind Microsoft and Anthropic.

Meanwhile, OpenAI is targeting profits by selling a premium version of ChatGPT and adjacent software to businesses. that it to cut It deals with chip makers Broadcom Inc. To Advanced Micro Devices Inc. To Nvidia to support its ambitions in the field of artificial intelligence.

Maryam Arik, CEO of AI startup Doubleword, said Google’s TPUs are mostly attractive to a few companies with large computing bills, such as Meta and Anthropic.

The chip industry “is not a zero-sum game with only one winner,” Barringer said.

First, AI developers can only access Google Slides through the company’s cloud service. They can use Nvidia graphics processing units, or GPUs, more flexibly. “Once you use TPUs, you’re locked into” Google’s cloud ecosystem, Arik said.

Being tied to a single supplier may be something companies avoid. This is no longer the case for Google, thanks to the progress it has made in the field of artificial intelligence.

“It’s certainly fair to say that Google is back in the game with Gemini 3,” Forrester analyst Thomas Howson said. “Indeed, to paraphrase a quote attributed to Mark Twain, reports of Google’s death have been widely exaggerated, not to mention irrelevant.”