- Coward has led IBM's networking biz for nearly five years
- He pioneered a buy rather than build approach and played a key role in five acquisitions
- It's not clear where he's headed next or what his departure means for the company's strategy
Andrew Coward has taken his final bow at IBM, raising critical questions about the future of the company’s networking strategy.
In a LinkedIn post announcing his exit, Coward said only that he was “leaving to pursue new opportunities in networking and AI” but noted he was “undecided” about where he would go next.
Coward served as GM of software networking at IBM since the start of 2021, having been brought on to build up IBM’s networking business in the wake of its Red Hat acquisition.
During his tenure, Coward was a driving force behind its multi-cloud, automation and orchestration efforts. He also pushed a buy rather than build mentality and had a hand in five major acquisitions, including the purchase of automation startup Pliant and the $6.4 billion deal for HashiCorp last year.
AvidThink Founder and Principal Roy Chua told Fierce that behind the scenes, Coward has been working to consolidate and unify IBM’s networking and orchestration product lines – which span everything from Red Hat Ansible and SevOne to NS1, HashiCorp's Terraform/Consul/Vault and IBM Connect Workflows (aka former Pliant assets).
“Andrew's departure could be an indication of increased consolidation and re-organization internally with HashiCorp coming in, or perhaps after five years, he's ready for a change of pace,” Chua speculated. “While IBM has great AI assets like watsonx, it's still a large company, and Andrew is an entrepreneur at heart."
Coward’s decision appears to be a relatively recent one. When meeting with Fierce Network Research's Mitch Wagner at DTW Ignite in Copenhagen last month, Coward passionately described how IBM was seeking to help bridge the gap between how operators think about network issues (in terms of switches and links) and how customers perceive them (“I can’t reach SAP right now”).
The key, he said at the time, was more shifting to a more application-centric view of the network. The aforementioned acquisitions gave IBM all the right pieces to put the puzzle together, but it seems Coward will be leaving before it’s finished.
“I wish we could say that we knew the destination for all this would be AI when it started,” he said in a comment that appears rather wistful in retrospect. “It's more than coincidence but it's less than serendipity."
But IBM’s work around network visibility and automation hasn’t been in vain.
“AI needs to understand the network completely in real time,” Coward said in June. “If it's going to tell you what to go do, or what is going to do it for you, then you better be sure that it's going to do the right thing… We have the omniscience, observability and automation, the powers we need.”
Fierce Network Research's Chief Analyst of Cloud, AI and Data Center, Mitch Wagner contributed to this story.