Google downplays DeepSeek impact amid capital expenditure plans
Google CEO Sundar Pichai addressed investors on Tuesday, minimizing the implications of DeepSeek on Alphabet's business. He stated that the AI models from Google are comparable in cost efficiency to those of the Chinese startup.
During the earnings report, Google announced a plan for $75 billion in capital expenditure this year, exceeding the analysts' anticipated $57.9 billion, which resulted in an 8% decline in shares during postmarket trading.
Pichai acknowledged DeepSeek's impressive team and expressed confidence in the continuous improvement of frontier models. He declared that Google’s Gemini models excel on the Pareto frontier, representing optimal goals in costs, performance, and latency.
"If you look at all three attributes, I think we lead this Pareto frontier," said Pichai, referencing Google's latest 2.0 Flash thinking models. He emphasized their efficiency compared to DeepSeek's offerings.
The investments in AI by Google and other tech giants are being examined more closely following DeepSeek’s recent launch of a competitive AI model that claims lower training costs.
Pichai mentioned that reasoning models are propelling increased spending towards inference, which involves an AI making conclusions post-training.
He expressed enthusiasm regarding AI opportunities, noting that as the cost of usage decreases, more applications will become viable, creating significant opportunities for growth.
