DriveNets New Funding Round Helps Meet Customers Demands for AI Fabrics
Introduction
DriveNets just closed a $410 million funding round at an $8 billion valuation. The company enters this new phase with cash flow-positive operations, validated AI reference designs, including one recently published with AMD, and production deployments already running at large-scale AI data centers and the largest service providers. Operators now gain a stronger partner committed to heterogeneous, multivendor AI fabrics that deliver performance without lock-in as the overall AI networking TAM climbs above $200 billion by the end of the decade.
Cash Flow Positive Performance Powers Inventory Build for AI Deployments
DriveNets has been cash flow positive for the past 18 months. The additional capital will help DriveNets build inventory and guarantee supplies for customers ahead of major AI cluster ramps. Current industry-wide supply constraints will remain well into CY2027, so allowing customers the ability to avoid the supply constraints and instead allow them to receive hardware and software on the aggressive timelines GPU roadmaps demand.
AMD Reference Designs and Dell AI Factory Availability Bring Validated Fabrics to Broader Markets
These blueprints integrate high-performance Ethernet fabrics, host configurations, and orchestration into repeatable deployments that maximize utilization. DriveNets supplies detailed reference architectures optimized for AMD Instinct GPU clusters to help with GPU diversity. Dell Technologies now offers the DriveNets software stack through its AI Factory program. Enterprises and AI cloud providers access a full solution via a trusted channel while retaining hardware flexibility and multivendor choice.
Hardware Development Plus Proven Reliability at Tier One SPs Creates a Distinct Advantage
DriveNets advances its own hardware initiatives to complement merchant silicon platforms. At the same time, the largest service providers already run DriveNets in production core networks that handle global scale and strict reliability requirements. This combination gives AI builders both forward-leaning designs and the operational confidence that only extensive field experience provides.
New Capital Accelerates Multivendor Full Stack Growth
The funding round supplies resources to increase investment in inventory to support customers in the currently supply-constrained market. We expect DriveNets to deepen its heterogeneous, multivendor model across more AI accelerator vendors, and larger AI fabrics and customers. The full-stack approach integrates seamlessly across diverse compute environments to drive performance optimization that will maximize the utilization of any AI accelerator. In turn, helping operators build a better cost model, and faster deployment cycles as AI infrastructure scales.
Ready for Heterogenous AI
Heterogeneous AI is emerging as workload economics force operators to mix accelerators, GPUs, custom silicon, and specialized chips within a single cluster to optimize cost-per-token and tokens-per-watt. DriveNets has two structural advantages here. First, its full-stack networking optimization maximizes utilization of any individual AI accelerator. Second, the same architecture coordinates performance across mixed accelerator environments, ensuring the entire cluster operates as one integrated system rather than the sum of independently optimized parts. As the AI market moves toward heterogeneous architectures, DriveNets is well-positioned to capture this opportunity.
Conclusion
DriveNets is combining investment with AMD and Dell momentum, ongoing hardware development, and unmatched reliability at major service providers, with multi-vendor AI stack optimization and heterogeneous AI readiness, at the same time the networking market is inflecting to hyper-growth mode. The company stands ready to help Hyperscalers, NeoClouds, and enterprises build the next wave of AI networks in a market about to surpass $200 billion.
Customers who use AMD or looking for multi-vendor GPU options, and prioritize performance, flexibility, and proven execution at scale now have an even more capable partner as the industry shifts toward standard Ethernet fabrics.
By Alan Weckel, Founder and Technology Analyst at 650 Group.