Pensando Systems, the long-awaited, software-defined brainchild of former Cisco execs, officially launched today with customers including Goldman Sachs, NetApp, and Equinix, and $278 million funding from the likes of Hewlett Packard Enterprise (HPE), Oracle, and Lightspeed Venture Partners.

It’s led by Silicon Valley legends Mario Mazzola, Prem Jain, Luca Cafiero, and Soni Jiandani — known collectively as “MPLS” — who have worked together for more than 25 years and left Cisco to start Pensando, where they went to work building a custom processor and SDN platform for the edge.

“It’s basically a software-defined, distributed service platform, which is edge accelerated, always secure, and for cloud, enterprise, as well as service provider customers,” CEO Prem Jain told SDxCentral.

Rumors started surfacing about the San Jose, California-based startup back in 2017. The team that sold three “spin-in” startups to Cisco had a new venture called Pensando. Those rumors quickly proved true. But the company — and exactly what it was up to — remained highly secretive for another two years.

“We said we will launch the company when we have a customer who has already deployed and will come on stage and tell their story,” Jain said.

Pensando has one customer (Goldman Sachs, also an investor) in production, upwards of 10 other major organizations including Equinix and two cloud providers in trials, “and a whole slew of enterprise customers from our partnership with HPE,” Jain said. NetApp will also start shipping Pensando’s technology in the next 30 days, he added.

John Chambers, a former Cisco CEO and current CEO of JC2 Ventures, is also an investor, and today the startup announced Chambers is chairman of the Pensando board of directors.

Additionally, Mark Potter, CTO at HPE, and Barry Eggers, a partner of Lightspeed Venture Partners, joined the board of directors. And both of their respective companies led Pensando’s Series C round, which raked in $145 million in funding. This brings its total raised to $278 million after an earlier founder-led series A round of $71 million and a customer-led series B round of $62 million.

The startup says its platform gives public cloud providers, service providers, and enterprises the ability to compete against Amazon — and win. “We can democratize the cloud,” Jain said.

About three years ago Amazon paid $350 million for an Israeli chip startup called Annapurna Labs. The cloud giant used that company’s technology as the basis for its AWS Nitro platform, which offloads storage, networking, management, monitoring, and security from servers to dedicated silicon. And it’s only used internally at Amazon.

Pensando’s distributed platform also provides programmable, software-defined cloud, compute, networking, storage, and security services. However, it’s cloud and infrastructure agnostic, and the startup claims it outperforms AWS Nitro, with 500% to 900% improvements in productivity, performance, and scale.

“Our chip enables other cloud vendors to compete against Amazon in a much better way,” Jain said. “It has lower latency, better performance, and it can provide them with more capabilities and function features than Amazon can provide.”

Pensando’s platform includes hardware and a software stack. The hardware, Naples 100 and Naples 25, is a programmable, distributed services card (DSC) that can be installed in any server and scales linearly. The DSC allows cloud providers and other customers to offer a number of software-defined services simultaneously, while freeing up more expensive hardware resources like general-purpose CPU cycles. Jain said it operates at 100G wire-speed with high-performance and low latency. It’s compatible with virtual machines (VMs) or bare-metal servers as well as containerized workloads.

The platform also includes a lifecycle management tool called Venice that applies policies from a central location across all active Naples nodes running in clouds and on-premises infrastructure. It also performs tasks such as in-service software upgrades, always-on telemetry and visibility across the entire infrastructure.

The distributed architecture means customers “can provide services like load balancing, firewalls, encryption, visibility, networking services, SDN” wherever the data is located, Jain said. And this becomes especially relevant as more data — Gartner puts it at 75% by 2025 — is created and processed at the edge.

Another benefit for enterprise customers is that the software-defined platform eliminates the need for legacy appliances such as firewalls, load balancers, and encryption devices.

“Goldman Sachs is significantly simplifying our data center architecture and reducing expenses with Pensando,” said Joshua Matheus, managing director at Goldman Sachs in a statement. “Pensando allows us to secure our east-west data center traffic while ensuring compliance, and at the same time, with improved telemetry, we reduce our time-to-problem resolution from hours to minutes.”

Jain said the startup is working with service providers as well. He won’t name names, but adds: “I used to run the service provider business at Cisco working with Verizon, AT&T, Comcast, and others. We have a very good relationship to them.”

The platform offers network slicing, and its distributed firewall, load balancing and SDN services with centralized policy automation can help carriers better deploy 5G services at scale, he added. “By making it into a distributed model, they can make it very cost effective and they can scale” to accommodate the magnitude of data 5G networks will enable, Jain said.