How Amazon Achieved Rocketing Sales and Growth From AI
Amazon’s latest quarterly results show the results of a company increasingly betting its future on AI.
The giant is reporting strong growth across its cloud computing division whilst rolling out AI tools throughout its vast retail and logistics operations.
The company posts net sales of US$167.7bn for the second quarter of 2025, up 13% from the same period last year.
But it’s the detail behind these numbers that tells the story of Amazon’s AI development.
Amazon Web Services (AWS), Amazon’s cloud division, pulled in US$30.9bn during the quarter – a 17.5% jump that helped drive operating income to US$10.2bn.
For a business that many see as the backbone of the modern internet, these figures suggest corporate customers are increasingly willing to pay for AI-powered services.
Andy Jassy, Amazon’s CEO, seems confident: “Our conviction that AI will change every customer experience is starting to play out,” he says, pointing to millions of customers now using the company’s enhanced Alexa+ service and shopping assistant tools.
Amazon Bedrock’s role in the enterprise’s AI push
Behind Amazon’s revenue growth lies a big effort to embed AI into everything from warehouse robots to customer service.
For example, the company is expanding its Bedrock platform – essentially AI building blocks for businesses – with new tools like AgentCore, which lets companies deploy AI agents securely across their operations.
Then there’s Kiro, Amazon’s new development environment that uses AI to help programmers write code more efficiently.
It’s part of a broader push into what the industry calls “agentic AI” – essentially systems that don’t just answer questions but get them done.
The company has also been signing up blue-chip customers at a rapid clip.
Recent AWS deals include PepsiCo, the American drinks giant, Airbnb, the holiday rental platform and Peloton, the fitness equipment maker. The London Stock Exchange Group and Japanese carmaker Nissan have also signed on.
The impact of robots learning to think for themselves
Perhaps nowhere is Amazon’s AI ambition more visible than in its warehouses, where the company operates more than a million robots.
The introduction of DeepFleet, an AI system that coordinates robot movements, has improved efficiency by 10% – which is no small feat when you’re moving packages at Amazon’s scale.
The company has also unveiled Vulcan, a robot that combines vision and touch sensors to navigate cluttered warehouse spaces.
It’s designed to work alongside humans rather than replace them, though the long-term implications for warehouse employment remain unclear.
Amazon says its AI-powered demand forecasting has improved regional accuracy by 20%, helping the company position inventory closer to customers and speed up deliveries.
Consumer AI getting a voice
On the consumer side, Amazon is betting heavily on voice technology and shopping assistance.
The company has expanded its Alexa+ service to millions of customers and added support for Spanish, French, Italian and German to its Nova Sonic voice AI system.
The shopping agent feature, which helps customers find products using natural language, has reached what Jassy describes as “many millions” of users.
Amazon has also introduced tools like “Hear the highlights,” which converts product reviews into audio clips – useful for customers who prefer listening to reading endless product descriptions.
These consumer-facing features might seem like nice-to-haves, but they’re crucial to Amazon’s broader AI strategy.
Every interaction generates data that helps train the company’s models, creating a feedback loop that could prove difficult for competitors to replicate.
Where Amazon is expanding its global infrastructure
Amazon isn’t just developing AI – it’s building the infrastructure to support it at massive scale.
The company has announced multi-billion dollar investments in North Carolina, Pennsylvania and Australia, all aimed at expanding the computing power needed for AI development.
Amazon has established “AI Zones” in Saudi Arabia through a partnership with local firm HUMAIN and in South Korea with conglomerate SK Group.
These full-scale AI development and deployment centres designed to serve regional markets.
The company has also introduced new hardware specifically for AI workloads, including Amazon EC2 instances powered by Nvidia’s latest Grace Blackwell chips.
Meanwhile, its S3 Vectors storage service promises to cut the cost of storing and querying the mathematical representations that power AI systems by up to 90%.
Additionally, Amazon’s Bedrock platform now includes models from Anthropic and TwelveLabs, which specialises in video understanding – showing Amazon’s strategy of offering customers choice rather than forcing them to use Amazon-developed models exclusively.
“Our AI progress across the board continues to improve our customer experiences, speed of innovation, operational efficiency and business growth and I’m excited for what lies ahead,” Jassy says.
