Nestlé is accelerating its digital push to become an “intelligent enterprise”, one that is connected from farm to fork and empowered by real-time data and artificial intelligence (AI). The food giant wants information to flow seamlessly across procurement, manufacturing and distribution so it can react faster, cut waste, and capture opportunities it might otherwise miss.
To further enable operational efficiency, Nestlé has rolled out “NesGPT”, a secure internal generative AI platform now used by more than 80,000 employees globally. An internal survey revealed that users of NesGPT saved an average of 45 minutes per week on creating content and searching for internal information.
However, scaling AI pipelines — from demand forecasting and shipping predictions to product formulation and the rollout of generative AI tools like NesGPT — takes more than cloud servers. It requires a flexible and resilient backbone.
“The network is at the heart of the business. It may not be very visible, but without it, nothing works — the business simply cannot function,” Ralf Huebenthal, global head of IT Platforms at Nestlé, tells The Edge Singapore. As such, Nestlé turned to Singapore Telecommunications (Singtel) to transform its global network into a dynamic cloud-centric infrastructure capable of providing seamless and consistent connectivity across business applications.
From legacy backbone to edge AI
Nestlé’s old private network, based on multi-protocol label switching (MPLS), served well in the past. However, it was built for stability rather than agility. Provisioning capacity for fast-moving marketing campaigns like Nespresso pop-ups took too long, showing the system was too rigid for today’s needs.
See also: The invisible F1 race
The challenge is magnified by Nestlé’s global footprint of 1,700 sites and more than 350 factories, with many located at remote locations near raw materials rather than major cities. This calls for strong connectivity to support the data and AI needs of warehouses, plantations and production plants at the edge of its operations.
Everything at Nestlé is at scale, and that scale is our advantage. But we also need to be fast at scale. Among the business benefits of the new network design [enabled by Singtel’s help] is higher availability in our factories, which is crucial because if the factory line shuts down, we can’t deliver.Ralf Huebenthal, global head of IT Platforms, Nestlé
He continues: “AI is already integrated in our end-to-end business process, but we cannot run all AI models on-premises as that would require huge data centres [so we’re using a mix of cloud, on-premises, and edge AI]. Having a solid network foundation can help ensure those platforms or services are resilient and reliable.”
See also: Singapore’s digital economy hits $128 bil as AI use among SMEs triples
Singtel Cube as digital conductor
Singtel won the contract to help rebuild Nestlé’s backbone, pitching its Cube solution as the orchestration layer that would act as a digital conductor for the company’s global infrastructure.
Instead of relying on fragmented systems across hundreds of countries, Nestlé can now monitor the performance, security, and traffic flows of its networks (such as private 5G, MPLS, and satellites) through a single platform. Cube is designed to provide enterprises with a consolidated view of their digital estate while reducing the complexity of managing end-to-end connections, from the edge of a plantation to the heart of a data centre.
“What we want to do is simplify connections from the edge to the cloud and reduce complexity for our enterprise clients, including Nestlé. Cube allows Nestlé to have a unified view of its digital/network infrastructure, which was previously lacking. Besides visibility, Cube also delivers observability, automation, and proactive capabilities,” says Ng Tian Chong, CEO of Singtel Singapore, in the same interview.
The promise of observability goes beyond dashboards and monitoring. It means ensuring core business processes would continue to flow uninterrupted, regardless of what was happening in the underlying IT infrastructure.
“Real observability starts with the business process. For example, can I do my production declarations and send shipping notifications on goods received? We look at all of that before filtering it down to the technology stack, all the way to the router and the switch. That way, we ensured Cube was monitoring what truly matters in our network so that when something isn’t working, we can quickly locate the issue and fix it accordingly,” explains Huebenthal.
That approach was highlighted during the recent undersea cable cut in the Red Sea. Nestlé was not affected, but the incident underscored the need for fast re-routing when disruptions occur. With factories depending on 24/7 connectivity, the ability to maintain operations while traffic is redirected has become a critical measure of resilience.
To stay ahead of the latest tech trends, click here for DigitalEdge Section
AI is expected to take that reliability and resilience even further. “Self-healing [networks] today that do not use AI are based on trial and error [as a human still needs to] write a script and see if it works. AI will enable proper self-healing because it will analyse if a processor is overheated, and if it is, it will automatically switch to another. Combining AI with Cube will, therefore, drive faster resolution,” says Huebenthal.
Nestlé, he adds, currently operates at 99.97% availability, but the goal is higher. “Our dream is 99.99% but we want to do it in a reasonable manner with proper observability. A successful enterprise digital transformation starts with the business process, not technology.”
What’s next
According to Huebenthal, more than 300 sites have switched to the dynamic cloud-centric infrastructure enabled by Cube. The cutovers have been smoother than expected, cutting nearly 400 hours of planned downtime, which translates into millions of dollars in avoided losses. The full migration is slated to complete by September next year.
When that transition is completed, Nestlé will look at other areas of collaboration. This includes integrating Cube with other tools, like Cisco’s AI Canvas, to get further network observability and be able to troubleshoot faster, as well as exploring 5G use cases. Cisco AI Canvas combines network, security, cloud, and observability data to deliver actionable insights, dynamic dashboards, and agentic AI capabilities that allow both human experts and autonomous AI agents to diagnose and resolve issues with high accuracy.
“We want to continue innovating with Singtel so that the network is no longer seen as just a piece of cable, but as the nerve centre of what we do,” says Huebenthal, who adds that he considers Singtel as a digital transformation partner instead of just a network provider.
For Singtel, the partnership has been equally strategic. “Both companies are legacy brands in our industries, but we share a common principle of being digital-first and driving innovation around that model,” says Ng.
Nestlé’s scale, he adds, provided an opportunity to ascertain Cube’s capabilities on a global stage. By deploying the orchestration platform across Nestlé’s 1,700 sites, Singtel was able to build more modules and automation with Nestlé engineers, which could later be applied to other industries.
That global reach is underpinned by Singtel’s own investments.
We’ve built a global infrastructure to connect globally. We’ve more than 370 points of presence worldwide and touch over 800 million subscribers since we hold stakes in major operators across Southeast Asia and India. Combining those with the fact that we have Singapore in our name, we’re like the ‘Switzerland of Asia Pacific’ [wherein we’re] neutral and able to bridge the East and West... [a role that has become more valuable as companies navigate] geopolitical bifurcation.Ng Tian Chong, CEO, Singtel Singapore
With this foundation, Singtel sees the Nestlé project not just as a showcase for Cube but as a template for other multinational clients looking to modernise their networks, strengthen their cybersecurity posture and harness AI at scale.
