David Erickson

Mid-Stride

There is a particular type of founder who is not building for the exit. David Sterling Erickson is building because the problem is real, the math is right, and someone had to do it. That someone turned out to be him and three other Stanford PhD graduates who decided, in 2013, that enterprise networks were flying blind - and that they had the mathematical tools to fix it.

Erickson leads Forward Networks as CEO from the company's Santa Clara headquarters. On a given Tuesday, Goldman Sachs is using his software to reason about a live model of more than 15,000 network devices. PayPal uses it. Telstra uses it. U.S. federal agencies use it. The product they are all running - Forward Enterprise - is not a monitoring tool, not a log aggregator, not a SIEM. It is a mathematically accurate digital twin of the entire network, updated continuously, queryable like a database, verifiable like code.

The distinction sounds subtle. It is not. Most enterprise network operations teams spend enormous time not knowing what their network is doing. Erickson's team built a product that makes that ignorance structurally impossible.

The 2023 Series D - $50 million, led by MSD Partners, with Andreessen Horowitz, Goldman Sachs Asset Management, and others already on the cap table - was not a lifeline. It followed 139% ARR growth. The kind of number that does not require a story around it.

The Lab Where SDN Was Born

Before Forward Networks, there was the Beacon controller. Before Beacon, there was a question on a Stanford computer science exam that turned into a research area that turned into an industry.

Erickson did his master's and PhD at Stanford under Professor Nick McKeown - the research group that helped invent Software Defined Networking. SDN is the idea that the control plane (the intelligence that decides where traffic goes) should be separated from the data plane (the hardware that actually moves it). This separation, which is now foundational to how cloud networks operate, was academic theory in 2008. Erickson was in the lab when it became something else.

His master's thesis won the Christofer Stephenson Memorial Award for best thesis in the program. His demo at ACM SIGCOMM 2008 - virtual machine mobility in an OpenFlow network - won Best Demonstration. He received the Microsoft Research PhD Fellowship in 2010. These are not decorations on a resume. They represent being in the first cohort of people to actually build and demonstrate what OpenFlow could do at scale.

The PhD dissertation, completed in May 2013, was titled "Using Network Knowledge to Improve Workload Performance in Virtualized Data Centers." The operative phrase is "network knowledge." That phrase is still the product thesis at Forward Networks twelve years later, just deployed at a different order of magnitude.

Four PhDs, One Network Problem

The four co-founders of Forward Networks - Erickson, Nikhil Handigol, Brandon Heller, and Peyman Kazemian - all knew each other from McKeown's lab at Stanford. That is the kind of founding team that fundraising decks get written about. It is also the kind of team where the technical arguments inside the company are genuinely hard, because everyone at the table has published peer-reviewed research.

Kazemian's contribution was particularly important. His research introduced header space analysis - a mathematical framework for modeling every possible packet path through a network, verified through formal proofs rather than sampling or heuristics. Forward Networks' core technology is built on this foundation. The product does not approximate. It calculates.

Erickson's role in that company is to translate what the research can do into what customers need solved. That requires fluency in both the mathematics and the operational reality of running a Fortune 500 network, which is a more unusual combination than it sounds.

What a Network Digital Twin Actually Does

Forward Enterprise collects configuration data and operational state from every network device - routers, switches, firewalls, load balancers - across an enterprise's entire infrastructure. Multi-vendor. Multi-cloud. On-premises and cloud, simultaneously. It then builds a mathematically accurate model of that network: every possible packet path, every policy rule, every access control list, every routing decision, as they exist right now.

The result is a network you can query. You can ask it whether any path exists from the internet to a specific server that bypasses the firewall. You can ask it whether last week's configuration change introduced a compliance violation. You can simulate what happens if a link fails. You can verify that your Zero Trust policies are actually enforced, not just configured. The answers are not estimates. They are mathematically provable.

In April 2026, Forward Networks announced general availability of Forward AI - an agentic system built on top of the digital twin. The AI can reason about network state, identify anomalies, and generate insights using the underlying mathematical model as its ground truth. Unlike AI systems that operate on logs or telemetry approximations, Forward AI's answers are bounded by what is provably true about the network.

The enterprise network has always been among the most complex, least visible pieces of infrastructure in a large organization. Erickson's bet - made in 2013 and still running in 2026 - is that visibility that is merely good enough is not good enough at all.

David Erickson on Forward Networks

Awards & Milestones

The Arc

Stack & Tools at Forward Networks

Forward Networks operates a sophisticated enterprise software stack supporting multi-cloud, multi-vendor network environments: