Uber has decided to expand its use of Amazon Web Services (AWS) for critical parts of its ride-sharing platform, signalling a shift in the ongoing battle among cloud providers. The ride-hailing company will increasingly rely on AWS’s proprietary chips, including the Graviton ARM-based processors and the new Trainium3 AI chip, to power its operations.

Uber’s Cloud Strategy Takes a New Turn

Uber’s shift towards AWS marks a notable change in its cloud computing strategy. Historically, the company operated its own data centres, but in 2023 it began moving large chunks of its IT infrastructure to the cloud, signing major multi-year contracts with Oracle and Google. At the time, Uber emphasised the use of ARM-powered compute instances, particularly those developed by Ampere, which Oracle’s cloud services used. The move to AWS now introduces Amazon’s Graviton chips and the AI-focused Trainium3 processors into Uber’s technology stack.

Graviton chips are designed by AWS to offer efficient, low-power ARM-based server CPUs that can handle demanding workloads while reducing energy costs. Trainium3, meanwhile, is Amazon’s answer to Nvidia’s dominance in AI chipsets, aimed at accelerating machine learning tasks with a focus on high performance and cost-effectiveness.

More Than Just a Chip Deal

The increased use of AWS’s AI chips isn’t just a simple upgrade; it sends a clear strategic message from Amazon in the competitive cloud market. AWS’s rivals, Google Cloud and Oracle Cloud Infrastructure (OCI), have been vying for Uber’s business since the company announced it would pivot away from its own data centres. Amazon’s move to deepen its relationship with Uber can be seen as a challenge to those competitors — a demonstration that AWS remains a force to be reckoned with, especially in the AI and cloud infrastructure race.

Uber’s embrace of AWS chips also reveals the complex web of relationships in Silicon Valley. Ampere’s story is a case in point.

Founded by former Intel executive Renee James after she was passed over for the CEO role at Intel, Ampere received significant backing from Oracle, which once held approximately a third of the company. This investment reflected Oracle’s own ambitions in cloud hardware and ARM-based computing.

However, the landscape shifted dramatically in late 2024 when SoftBank acquired Ampere, prompting Oracle to sell its stake and record a substantial pre-tax gain of $2.7 billion. Around that time, James left Oracle’s board and Ampere’s day-to-day operations. Oracle’s decision to exit the chip business underlines a broader trend: companies increasingly prefer to buy specialised hardware from chipmakers like Nvidia rather than develop their own designs in-house.

Amazon’s AI Chips Take Centre Stage

Amazon’s investment in AI chip development is part of a broader trend in cloud computing providers seeking to control more of the technology stack. As AI workloads grow in complexity and scale, having custom-designed chips can offer performance advantages and cost savings. AWS’s Graviton chips have already gained traction for their efficiency, while Trainium3 targets the burgeoning AI training market, which Nvidia has largely dominated.

Mastheads

Uber’s decision to trial Trainium3 shows they’re gaining confidence in Amazon’s chip capabilities. The ride-hailing giant processes vast amounts of data daily, from routing and logistics to customer experience and fraud detection. Leveraging AI chips tailored for these tasks can improve speed and reduce operational costs.

These implications extend beyond just Uber. Amazon’s push into AI chips signals a broader industry shift where cloud providers are no longer just resellers of generic hardware but are becoming chip designers and manufacturers. Vertical integration might shift how tech companies pick cloud vendors, favoring those offering optimized hardware and software.

The Broader Cloud Computing Battle

Enterprise cloud contracts are fiercely contested. Google and Oracle have been aggressive in courting companies like Uber, often touting specialised hardware such as Ampere’s ARM chips as a differentiator. Yet Amazon’s renewed success with Uber suggests AWS’s strategy of combining proprietary chips with a massive global cloud infrastructure remains compelling.

Uber’s cloud journey reflects a bigger industry trend: how tech firms juggle performance, cost, and vendor ties in a fast-changing cloud market. By embracing AWS’s latest chips, Uber is betting on a future where custom silicon and AI acceleration play a key role in its growth and efficiency.

Meanwhile, Oracle’s pivot towards building data centres for AI workloads, including partnerships with OpenAI, shows the cloud market isn't static. Larry Ellison, Oracle’s chairman, has publicly stated that designing chips internally no longer offers a competitive edge, preferring to strike deals with established chip manufacturers like Nvidia. The approach contrasts with Amazon’s more vertically integrated model.

Uber’s expanded AWS contract also hints at the increasing importance of AI across industries. AI is more than a buzzword; it’s reshaping areas like logistics and customer service. Companies investing in specialised AI hardware aim to harness this potential while managing costs.

Historical Context and Future Prospects

Looking back, Uber’s cloud strategy mirrors the wider shifts in the industry. Early on, many tech giants managed their own data centres, but the shift to cloud providers has accelerated in recent years, driven by cost savings, scalability, and innovation. AWS, Google Cloud, and Oracle have emerged as major players, each with different strengths and approaches.

Amazon’s development of AI chips like Trainium3 is part of a trend where cloud providers seek to differentiate through hardware. Nvidia’s dominance in AI accelerators has been challenged by these new entrants, sparking a chip arms race. For companies like Uber, which rely on large-scale AI workloads, the choice of cloud provider increasingly hinges on the quality and cost of AI hardware.

Uber’s increasing use of AWS chips could also have ripple effects. If Trainium3 proves successful, more customers might follow suit, accelerating a shift away from Nvidia-centric AI computing. That said, Nvidia’s ecosystem remains strong, meaning the competition will be intense.

Uber’s expanding contract with AWS shows the ride-hailing giant’s need for flexible, high-performance computing resources as it navigates a competitive market with rising fuel costs, regulatory pressures, and evolving consumer demand.

The cloud wars are far from over. But Uber’s latest move shows that Amazon’s bet on AI chips is gaining traction — and that the company is willing to back its own silicon to keep rivals at bay.

Related Articles

Uber’s growing reliance on Amazon’s custom AI chips signals a new chapter in cloud computing competition, where chip design and AI performance become key battlegrounds. As the company trials AWS’s latest technologies, the broader industry watches closely to see which cloud giants will lead the next wave of innovation.

This article was created with AI assistance.