Sundar Pichai Claims Gemini Is as Cheap as DeepSeek, Will Push More AI Into Google Search in 2025

• Alphabet’s Gemini AI models are as cheap to run as China’s DeepSeek, CEO Sundar Pichai has claimed.
• Artificial intelligence will soon play a much larger role in Google search.
• Pichai said AI in search will allow users to ask follow-up questions, making Google something more akin to OpenAI’s ChatGPT.

While the U.S. continues to lead in commercialization, the release of China’s DeepSeek has sparked concerns that overseas competition is gaining ground with highly cost-efficient AI models.

Like other leading executives, Alphabet’s Sundar Pichai has played down the threat of new competition and emphasized how much the tech giant focuses on AI.

Gemini as Cheap as DeepSeek

During Alphabet’s latest earnings call, CEO Pichai bolstered the capabilities of Gemini alongside the offerings from leading companies such as OpenAI and Meta, as well as emerging challengers like DeepSeek.

Since its launch, the new challenger has sent shockwaves through the technology industry, rocking the foundations of U.S. dominance with its cheaper and equally powerful model.

Discussing DeepSeek, Pichai said he believed they were a “tremendous team” who have done “very, very good work.”

However, the CEO played down any concerns that it was more efficient than Gemini.

Pichai said Gemini’s models were “some of the most efficient models out there, including comparing to DeepSeek’s V3 and R1.”

The CEO said that the increase in cheaper AI models worldwide would only help increase their adoption, which would, in turn, help Google due to its billions of users.

AI and Search

Pichai also believes that 2025 will be a pinnacle year for AI search. The search engine giant is currently on a “journey” of shifting from a traditional list of links to a more interactive, assistant-like experience.

Last year, the company implemented AI overviews to its search engine, marking a controversial change to how its billions of users receive information.

Following its launch in May 2024, users quickly reported errors with the overviews, including Google stating that humans should eat at least one small rock daily.

However, Pichai told staff and investors that the company was only beginning its AI search journey.

“As AI continues to expand the universe of queries that people can ask, 2025 is going to be one of the biggest years for search innovation yet,” said Pichai in Alphabet’s earnings call.

Since the release of OpenAI’s ChatGPT in 2022, Google has joined other leading tech companies in playing catch-up with AI.

Pichai said Google plans to implement more AI features from its research lab, DeepMind, into search throughout 2025.

“I think the [Search] product will evolve even more,” Pichai said. “As you make it more easy for people to interact and ask follow-up questions, etc., I think we have an opportunity to drive further growth.”

The Challenges of AI

So far, Google has had a rough time while rolling out its AI products.

The initial release of AI Overview was rife with questionable responses from users , including strange answers such as “sticking cheese to pizza is better with “non-toxic glue.”

This wasn’t the first time Google had come under fire for AI misinformation.

Following the release of Bard, now Gemini, the AI model made a live blunder during its on-stage demo by providing a faulty tidbit about NASA’s James Webb Space Telescope.

These missteps underscore a broader challenge seen across AI development. Unlike traditional search engines, which pull directly from sources, AI models generate responses based on probabilistic reasoning.

As Google and other tech giants push further into AI-driven search and assistants, ensuring the accuracy and trustworthiness of these tools remains a significant hurdle.