Grab selects AWS as preferred cloud provider

AWS is the preferred cloud provider for Grab with 80% of the workloads running on its cloud.

As a leading technology company in Southeast Asia, Grab’s journey from a ride hailing service provider to a superapp has been an incredible one. Today, Grab’s services are available in over 700 cities in eight countries in Southeast Asia, serving millions of customers through a single app.

At the core of Grab’s success will be its technology providers. As a company that has been leveraging the cloud since its inception in 2012, the superapp continues to innovate its services as emerging technologies are made available. Making the leap to offering financial services via the app was definitely one of the biggest challenges the company faced but it still managed to get the entire process completed and is now providing digital banking services in Singapore and Malaysia.

Enabling Grab to offer all these services is none other than its cloud partner AWS. At AWS’s re-Invent summit in Las Vegas, the leading superapp announced that AWS is its preferred cloud provider. With the partnership, Grab has pursued a technology led strategy to accelerate growth across its mobility, deliveries and financial services verticals, while continuing to improve its operational efficiencies and reduce IT infrastructure costs.

According to Suthen Thomas Paradatheth, CTO, Grab, Grab’s strategy for growth is anchored on constant innovation to outserve the needs of our users and partners.

“This requires rapid experimentation, while ensuring security and stability, along with the ability to fully harness the potential of the latest tech like GenAI. We're pleased to extend our partnership with AWS as our preferred cloud partner to continue to support us on this journey,” he said.

Echoing Paradatheth’s comments is Gunish Chawla, Managing Director, Commercial Enterprise, Digital & SMB, ASEAN for AWS. For Chawla, Grab and AWS have had a long-standing relationship since 2012.

“When Grab first started out, it was critical for us to provide the foundational infrastructure that is so often needed. If you think about Grab’s business, they were launching new services, but they were also scaling in a multi-country footprint. Hence, the first part of essentially working with Grab was providing our infrastructure, and ensuring that we provided the required reliability, uptime and security for them to be able to expand,” said Chawla (pictured below).

Meanwhile, Mohan Krishan, Head of Engineering, Technical Infrastructure for Grab commented that Grab is always a tech company first and they believe in constant innovation with a focus on using the innovation to drive better outcomes for customers, partners, drivers and merchants.

“To be able to innovate, a key component of that is the ability to experiment and to try new capabilities while still ensuring that the cost and the stability of our services continue to be upheld,” said Krishnan.

For Krishan, there are three dimensions to Grab’s tech strategy. The first is optimization, with the second being able to scale securely and the third is the focus on innovation. And in all these three strategies, Grab believes AWS is the right partner to help them deliver.

Optimizing costs and scaling securely

All organizations will want to minimize their expenditure costs and get the best out of their tech investments. For Grab, it's no different as well. While the superapp looks to offer more products and services, it will also want to do so by managing and balancing its costs.

By adopting AWS’s suite of cloud-based solutions, Grab has been able to gain agility and reduce operational costs. For instance, Grab uses analytics service AWS Clean Rooms, which enables secure, privacy-preserving data collaboration between different entities and organizations. Grab also leverages AWS’s purpose-built databases and has migrated more than 400 backend application services from traditional virtual servers to AWS Graviton2 processors to drive high performance, as well as cost and energy efficiency.

At the same time, the superapp will also want to scale its service offerings and ensure customers have a seamless experience. According to Grab, it's on demand transactions were up 22% in the third quarter of 2024. To cate to this surge, Grab relied on Amazon Relational Database Service (Amazon RDS) as its transactional database coupled with Amazon DynamoDB. This ensures high availability, scalability, and adaptability of its platform to drive exceptional customer experience fueled by more accurate searchable data.

As Grab services 41.9 million monthly transacting users and over 13 million driver and delivery partners registered on its platform, it does over a hundred transactions, receives over 500k GPS pings and services over 50,000 ETA requests every second.

“That's a lot of load, and we need to ensure that our infrastructure can not only serve that load but serve that load well across the disparate countries that our customers are based. Now, doing that without involving too much of our teams is key to the value that we get from. It's essentially the key leverage point of our relationship with AWS,” added Krishnan (pictured below).

With the year-end holiday season in Southeast Asia approaching, Grab can seamlessly adapt to evolving customer needs by easily adjusting resources dynamically based on user demand. Conversely, during off-peak periods, resources can be scaled down to save costs.

Accelerating AI-led growth

According to Krishnan, Grab is one of the pioneers of AI adoption in Southeast Asia. Even before the launch of its superapp, Grab has been committed to being at the forefront of exploring how the latest AI technologies can better serve and respond to the needs of its users and partners.

Krishan also pointed out that Grab’s team of developers are now capable of focusing on more tasks as AI has enabled them to do so.

“I think the whole space around how AI is going to change the role of developers and how we do our work is really interesting, and it's definitely an area that Grab's been very interested in as well. We continue to do lots of experiments in that space and are early adopters in the whole class of gen AI internal productivity tools, including in the software engineering space. We’ve been working closely with AWS and have run lots of trials on some interesting capabilities that AWS has like, for example, Amazon Q. That's been an area that we've been very keen on,” said Krishnan.

He added that Grab has also been learning from AWS on how they have internalized the use of this type of technology to accelerate their own software engineers. He believes this aspect of the partnership has been very rewarding for Grab as they are able to learn how AWS not only builds these capabilities but also uses them.

“Engineers want to be able to get their work done as fast as possible. They're also excited by the ability to be productive and to be able to focus on the creative aspects of their work. And GenAI is actually very good at taking out some of the drudgery of the work,” added Krishnan.

Chawla also commented that Grab has been very proactive and been working very deeply with AWS on their GenAI (7:35) products and pushing them in terms of features and sets.

“Grab is often one of the first people to experiment with a lot of our innovative features. And very often, it's been on the roadmap on account of proactive suggestions from Grab,” said Chawla.

Among some of the innovations from Grab is Catwalk, Grab’s machine learning (ML) model platform. It is built on Amazon Elastic Kubernetes Service (Amazon EKS) and has been used to deploy over 1,000 AI models in production, such as route guidance and pricing. With Catwalk, Grab provides users’ real-time decision-making across its services and delivers personalized experiences like tailored restaurant recommendations, loyalty rewards, and bespoke financial services based on users’ preferences.

Grab also uses AWS’s custom-designed AWS Inferentia chips with specialized ML inference capabilities to cost efficiently power its AI-powered services, including map enhancements and fraud detection in its digital banks.

At the same time, with the superapp storing stores hundreds of petabytes of data and processes over 200 TB of data, Grab continues to develop and implement several AI-powered use cases, particularly to improve driver productivity and support merchant growth. By integrating LLMs with point-of-interest data and historical customer notes, Grab has refined its last-mile guidance system for delivery partners.

“We've already launched products that are customer-facing, that are infused with AI capabilities. So, for example, today merchants can already use AI capabilities that we have in our product to automatically load in their menus without needing to key it out in. We're using our Max AI chatbot to be able to provide our merchants with insights on their business and how they can do better automatically,” said Krishnan.

Krishan added, “We're using AI to better guide our drivers around how to get higher quality jobs or improve their earnings. All of these are capabilities that are already available in the market, and we continue to look at building and innovating on by leveraging technology partners like AWS.”

Grab’s multi-cloud strategy

While AWS is the preferred cloud provider for Grab, Krishan also pointed out that Grab works with other cloud providers as well.

“AWS is definitely our preferred cloud provider. We run the bulk of our workloads on AWS, but we continue to maintain good relationships and partnership with other cloud providers as well. Both Azure and Google Cloud continue to be providers for Grab,” said Krishnan.

For Krishnan, running a multi-cloud is actually a trade-off for Grab. He explained that it takes effort to build the right type of skillset from the teams to be able to work across multiple clouds and to simplify the choices for internal product teams on where they should be deploying their workloads.

“We feel that this strategy of still leveraging multiple clouds, while it takes more effort, has the benefit of actually being able to use the best capabilities from different cloud providers. They are very similar, but they are also different areas where different cloud providers have different advantages. What we've been trying to pursue at Grab is trying to maximize both in terms of simplifying our multi-cloud setup by picking AWS as our preferred choice,” explained Krishnan.

Krishnan also mentioned that about 80% of Grab’s workloads are running in AWS today. When a customer makes a booking on the Grab app, Krishan said it's definitely going to AWS. He added that a lot of transactions that are going through Grab’s systems are running on AWS.