Striveworks completes Series B capital raise to scale operational AI tool
Gettyimages.com / Your_photo
By
Ross Wilkers
Striveworks now has Washington Harbour Partners in place as a new investor to shape this phase of the strategy and as Striveworks' CEO tells us, knowing when the tech does not work is just as key to product development as when it works.
In starting up Striveworks, CEO Jim Rebesco and the founding team sought to draw on some lessons learned from the decade he spent at the high-frequency trading company Virtu Financial.
Virtu built a large number of artificial intelligence-centric algorithms and models to follow financial markets, but Rebesco will concede they did not have a single mechanism for monitoring every model that was in production.
“It needs to have confidence, meaning like need to be able to monitor it, understand when it's working, when it's not, and quickly change from not working to working again quickly,” Jim Rebesco, Striveworks’ co-founder and chief executive, told WT. “AI's got to work and AI's got to know when it's not working and if you can put those two things together, you can start to really power a lot of important applications.”
Striveworks then opened for business in 2018 with the idea of deploying operational AI at scale at the needed rigor and visibility required for governments to trust the technology. Striveworks designs its platform to function as a no-code offering that helps data science and analytics teams work on the models before deploying them.
In 2023, the company closed $33 million capital raise after finding its initial footing in the national security and highly-regulated verticals.
Striveworks has now caught the attention of Washington Harbour Partners, the private investment firm whose government market holdings include Raft and Groundswell. WHP also owns equity stakes in several other contractors and led a Series B investment round for Striveworks announced Tuesday (today), of which the terms were not disclosed.
Striveworks is also a member of an Anduril-led team that secured a $99.6 million Other Transaction Authority agreement in 2025 to deliver a core AI platform and model for the Army’s Next Generation Command and Control initiative, also called NGC2.
NGC2 is one of the Army’s flagship efforts at incorporating more software-defined and open architecture applications into the service branch’s environment for sharing information. Striveworks’ role in the team is to provide the catalog for AI platform services and models to use in NGC2.
Rebesco characterized NGC2 as an effort the Army shaped through a “characteristic of needs statement” versus traditional requirements-centric processes. In essence, the government laid out its problems in great specificity and let industry propose solutions for them.
Striveworks will put the bulk of this newfound capital toward its research-and-development and engineering efforts in order to further scale out its offerings, Rebesco said. Hiring will also be a priority for Striveworks from the Series B, he added.
“Mapping from data to decisions or actionable insights” is one phrase Rebesco brought up in the interview as often repeated in the government technology ecosystem, but one he said has a “kernel of truth in there” because it does illustrate the problem set Striveworks and other companies work with.
“Human beings needs some way of processing a lot of data, getting down to what’s really important and helping them do their human job more accurately, faster and better,” he said.
National security enterprise data only makes that problem and its related workflows more complicated as organizations try to connect AI agents to data sources and other tools. Rebesco referenced logistics, personnel and intelligence functions as examples of where this need to work with large amounts of data comes in.
But as he put it and others have too in talking to WT, generic AI chatbots and other similar out-of-the-box tools are mostly not granular enough to accomplish the tasks at hand. This is especially true in edge environments where access to basic compute power and storage is denied or greatly curtailed.
“That's the kind of mission IT question that they have to be a really thoughtful partner in, in terms of how I manage both as network and compute get necessarily restricted by going further and further away from the cloud what's the quality of service I can deliver in all these agentic applications,” Rebesco said.
“Some people will answer like just bring a giant toaster with the LLMs on it, I don't think that's the right answer.”