The Ecosystem Architect
In mid-2023, NVIDIA did something unusual: it split its global channel chief role in two. Alvin Da Costa took the consulting giants - Deloitte, Infosys, Wipro, Capgemini. Darrin Chen got everything else. Which, in NVIDIA's case, meant the world.
Chen now runs go-to-market strategy and operations for the NVIDIA Partner Network, the company's global partner ecosystem that connects NVIDIA's accelerated computing and AI platforms to solution providers like World Wide Technology, systems integrators like Mark III Systems, and thousands of partners stretching across verticals and geographies. The NPN is the connective tissue between what NVIDIA builds and what actually reaches customers. Chen is the one maintaining that tissue.
The restructuring wasn't housekeeping. It was a structural response to explosive growth. As NVIDIA transformed from a GPU designer into what CEO Jensen Huang calls a "full-stack computing company," the partner layer had to keep pace. More partners. More specializations. More complexity. The company needed dedicated leadership for its program infrastructure - and it reached for someone with three decades of evidence.
That someone had been at this since before NVIDIA made GPUs famous.
How You Build a 30-Year Career in Channels
Chen joined NVIDIA not by applying - but by acquisition. When NVIDIA closed its $7 billion purchase of Mellanox Technologies in April 2020, Chen came with it. He had spent more than 11 years at Mellanox, building the company's worldwide channel operation from the ground up and eventually becoming VP of Worldwide Channels. When the deal closed, he continued doing the same job - now under an NVIDIA banner, handling networking channels specifically.
Mellanox was the kind of company only infrastructure people talked about at the time - high-speed InfiniBand and Ethernet interconnects powering data centers and HPC clusters, the plumbing nobody saw but everybody depended on. When NVIDIA bought it, the rationale was clear: to own the compute and the connection. Chen was already the person who had built the commercial relationships around that connection layer.
His path to Mellanox ran through QLogic - nine years building product marketing, channel marketing, and channel sales programs for a company that made Fibre Channel host bus adapters, the storage networking hardware of its era. Before QLogic, he was at Seagate as a product marketing manager during the peak of the storage wars. Before that, AST Computer, a PC manufacturer that once challenged Compaq before the market consolidated around it.
What's consistent across all of it: channels, programs, partners. The same craft, applied to successive generations of infrastructure technology.
I am very pleased to appoint Ido, Darrin and Tziporet to VP and Hillel to senior principal engineer. They have been tremendous assets and their dedication and contributions have been significant to making Mellanox successful.
Eyal Waldman, CEO, Mellanox Technologies — March 2014The Company Inside the Company
The NVIDIA Partner Network is, in functional terms, a company inside a company. It has tiers - Registered, Preferred, Elite. It has specialization tracks: Agentic AI, AI Factory, Embedded AI. It has a portal, training programs, sales toolkits, and marketing resources. Chen's team maintains all of it - and makes it work coherently across partners that range from boutique AI consultancies to global system integrators.
The NPN's evolution mirrors NVIDIA's own. Five years ago, an NVIDIA partner was likely someone reselling GPUs or building workstations. Today, the partnership designations include "AI Factory" - a term that didn't exist in the industry vocabulary until Jensen Huang started using it. The partners Chen's program supports are now building full AI infrastructure stacks: compute clusters, software platforms, inference pipelines. The program has to match the ambition of what's being deployed.
Running this at scale requires something that doesn't appear in most executive bios: genuine fluency in channel economics. Distribution margins, partner program mechanics, deal registration systems, MDF allocation, tier qualification criteria - the operational machinery that determines whether a partner ecosystem actually grows or quietly stagnates. Chen has been inside that machinery for three decades.
NVIDIA Partner Network (NPN) - Program Structure
What Economics and Psychology Have to Do With Channels
Chen studied both Economics and Psychology at UC Irvine - the kind of combination that, in retrospect, looks like it was designed for channel sales. Economics gives you the market-level logic: why partners join programs, how incentive structures shape behavior, where distribution margins squeeze. Psychology gives you the human layer: what makes a partner feel invested, how trust is built through consistency, why the best reseller relationships function less like vendor-customer dynamics and more like long-term alliances.
He followed the UCI degree with an MBA at Pepperdine - practical, business-focused, the kind of credential that signals someone who wants to run operations, not theorize about them. By the time he was into his career at QLogic, he was already running programs rather than just executing within them.
The dual background shows in how channel programs actually work when they work: they're not just incentive tables and discount ladders. They're relationship architectures. Someone has to design the structure in a way that makes partners feel like the program is built for them, not just for the vendor's revenue goals. That's a psychological problem as much as an economic one.
Through Every Wave
There's a version of a tech career that chases every shiny thing - new company, new role, new industry vertical every two or three years. Chen's career ran the opposite direction. He stayed. At QLogic for nine years. At Mellanox for eleven. Not because the opportunities weren't there, but because channel programs compound. You don't build a world-class partner ecosystem in eighteen months and then hand it off. You build it, run it through a down cycle, rebuild when the market shifts, and eventually you understand it at a level that can't be shortcut.
The result is a career that reads like a map of infrastructure computing's evolution. Storage networking (Seagate, QLogic). High-speed interconnects (Mellanox). GPU-accelerated computing, networking, and now full-stack AI (NVIDIA). Each wave was bigger than the last. Each one required the same core skill - connecting vendors to partners to customers through programs that actually work in practice.
At NVIDIA in 2024, with the company posting revenues of over $60 billion in a single quarter and partners scrambling to build AI Factory capabilities, that skill is in acute demand. Chen is running the apparatus that makes NVIDIA's partner growth tractable - the program infrastructure, the GTM frameworks, the operational machinery that turns partner enthusiasm into deployed AI systems at scale.
It's not the role that gets the keynote at GTC. But the keynote doesn't exist without it.