AI Will Work Even on Powered-Off Phones and Laptops: Google-OpenAI Rivalry Heats Up with Launch of New Gemini Agents
Google no longer wishes to remain merely a search engine company. Its next objective is to establish the same dominance in the world of consumer AI that it once held in internet search. And now, it appears increasingly likely that it could wrest the crown in the AI race away from OpenAI. On May 19, Google unveiled a series of AI agents powered by its new Gemini 3.5 Flash model. These include AI coders designed specifically to challenge tools offered by OpenAI and Anthropic.
Google’s focus has now shifted beyond just chatbots to “working AI”—intelligent agents capable of performing tasks. The company is introducing agents that can continue to operate even after a user has turned off their phone or laptop. An agent named “Gemini Spark” will be capable of autonomously handling various digital tasks, such as scanning emails and organizing group trips. Meanwhile, “Information Agents” being integrated into Google Search will keep a close watch on sports tournaments, retail sales, and the stock market. Some of these AI agents will reside directly within the Gemini app—an application currently utilized by approximately 900 million people every month—while other AI features will be integrated into Google Search. This move could significantly increase the pressure on OpenAI.
In fact, Google’s rapid progress had already put OpenAI on high alert. Following the release of the Gemini 3 model in November, OpenAI CEO Sam Altman issued a “Code Red” emergency alert to his employees. Consequently, the company intensified its focus on developing its own coding agents. However, the arrival of Gemini 3.5 Flash and Google’s new suite of AI agents has once again cast doubt on ChatGPT’s lead in the market. The market capitalization of Google’s parent company, Alphabet Inc., has surged from ₹382 lakh crore in January to reach approximately ₹478 lakh crore today.
Nevertheless, alongside this success, the challenges are also mounting. According to CEO Sundar Pichai, the number of tokens—the metric used to measure AI usage—stood at 48 trillion a year ago; this figure has now surged to 3.2 quadrillion (3.2 followed by 15 zeros) per month. This implies that computing costs are rising rapidly. To mitigate these expenses, Google may introduce advertisements into its AI services. Currently, the Gemini app remains ad-free; however, Google has already begun displaying advertisements within the AI-generated responses provided by its search services. This signals that, in the near future, AI will evolve into not merely a technological battleground, but also the world’s largest advertising and data enterprise.
Capital Expenditure Set to Increase Sixfold: Generating AI tokens requires significant computing power. Consequently, Google’s capital expenditure is projected to increase sixfold compared to four years ago, reaching ₹18.18 lakh crore this year. This expenditure could rise even further due to escalating costs ranging from semiconductor chips to electricity.
