Amazon As we increase our AI spending, we aim to extract savings from our robotics investments.
As the Financial Times (FT) It has been reported On Wednesday (February 26th), the Tech giant is expected to spend $35 billion on a retail network, including robot-powered warehouses. temu.
Most of the $100 billion will be spent on artificial intelligence (AI) initiatives, but about a quarter will be allocated for automating Amazon’s e-commerce business, citing analyst estimates, the report adds.
“Today we see how fruitful this technology can change our everyday lives.” Ty BradyThe Amazon Robotics chief technologist told FT, adding that the company plans to “continue to invest” in automation.
The report points out that Amazon’s Shreveport at LA Fulfillment Centre has already shown the type of savings automation brings.
The robots are involved at all stages in this six-month 3 million square foot facility Filledhelping the company cut 25% compared to the last generation of warehouses after a 10-fold increase in robotics.
As Pymnts wrote earlier this month, Amazon plans Spend $26 billion This quarter, which develops AI capabilities for Amazon Web Services (AWS), is an spending level that is expected to remain consistent throughout the year.
That type of spending is consistent with Amazon’s leading technicians. Spending $320 billion As they begin in 2025 Microsoft president Brad Smith Recent Blog Posts – Publish in the New Industrial Revolution.
“But ai A large investment is required,” Pymnts wrote recently. “Training large language models uses thousands of GPUs (each NVIDIA GPU costs over $10,000) or specialized AI chips in total for hundreds of millions or hundreds of millions of dollars. To run these AI models at scale, you also need more servers and high-performance data centers that require more cooling and maintenance.”
Other robot/AI news has been reviewed by Pymnts this week Possibility of household robots After the debut of a new robot from AI startup figures.
Jenny ShaneGeneral Manager of Robot Builders Nexcobottold Pymnts that humanoid robots face more complex challenges than their industrial cousins.
“Traditional industrial robotic arms with vision systems rely primarily on pre-programmed instructions to perform tasks, which work well in a repetitive, goal-oriented factory environment where applications are repetitive, goal-oriented,” she said.
However, “Unlike factories, the home environment is very dynamic and the tasks vary widely from home, so implementing humanoid robots in a home environment is a more complicated advance,” added Shane.
