Industrial IoT (IIoT)
Why a Software Supply Chain Company Became an IoT Device Management Leader
Published by IOT San Jose Research & Editorial Team

- Why Siloed Device Management Stopped Making Sense
- What Unified Device Management Actually Looks Like
- A Concrete Illustration of the Siloing Problem
- Business Value for South Bay IoT Companies
- Implementation Advice for IoT Device Fleet Management
- Common Challenges in IoT Device Management
- What's Next for IoT Device Management
JFrog, the Sunnyvale software supply chain company best known for the Artifactory binary repository most software developers interact with daily, paid roughly 33 million dollars in 2021 for an Israeli startup called Upswift and folded it into the platform as JFrog Connect. The logic behind that acquisition, treating IoT device fleet management as an extension of the same software supply chain discipline already governing how code moves from a developer's laptop into production, captures a genuine shift in how device management is being rethought in 2026. Rather than a separate, siloed tool bolted onto the side of an organization's development pipeline, device management is increasingly being absorbed into the same DevSecOps infrastructure already managing everything else a company ships.
Why Siloed Device Management Stopped Making Sense
For most of the IoT industry's history, updating a fleet of deployed devices in the field was handled through tools and processes entirely separate from how a company managed its cloud or server software. A development team might use sophisticated continuous integration and deployment pipelines for its backend services while pushing firmware updates to IoT devices through a much more manual, ad hoc process, often built around custom scripts with little of the automated testing, rollback protection, or security scanning that had become standard for other software. JFrog's own leadership has described this gap directly: updating edge devices is often siloed from the modern software supply chain, even though the massively distributed nature of edge and IoT deployment expands an organization's attack surface more than almost any other category of software delivery.
That siloing created real operational risk. A vulnerability discovered in a company's cloud services might get patched within hours through an automated pipeline, while an equivalent vulnerability in deployed IoT device firmware could linger for months because the update process required manual intervention, lacked rollback protection, or simply was not treated with the same urgency as server-side software. JFrog's acquisition of Vdoo in 2022, a connected device cybersecurity company, for roughly 300 million dollars extended this unification further, bringing device-specific security scanning into the same platform already handling artifact management and distribution.
What Unified Device Management Actually Looks Like
A modern, consolidated approach to IoT device management treats a fleet of deployed devices the same way a software team treats a fleet of servers: with automated over-the-air updates, staged rollouts that limit a bad update's blast radius before it reaches the entire fleet, automatic rollback when an update fails, and continuous security scanning of the firmware itself for known vulnerabilities before it ever ships. This represents a meaningful shift from treating a device update as a discrete, manually triggered event toward treating it as one more stage in a continuous software delivery pipeline that happens to terminate on physical hardware instead of a cloud server.
The security dimension deserves particular emphasis given how device management and vulnerability exposure are directly connected. A fleet of thousands or millions of deployed IoT devices represents a meaningfully larger attack surface than a company's cloud infrastructure alone, and unlike a cloud service that an organization fully controls, field devices are physically distributed, often in locations with limited connectivity, and harder to immediately patch than a centrally hosted service.
Industry data on IoT device management platforms consistently identifies the same five core capabilities as essential regardless of vendor: device onboarding and identity management, remote configuration and control, health monitoring, firmware updates, and security enforcement. What has changed in 2026 is less the list of required capabilities itself and more the expectation that these capabilities integrate directly with an organization's broader software delivery infrastructure rather than existing as a standalone point solution disconnected from how the rest of the company ships software.
A Concrete Illustration of the Siloing Problem
Consider a mid-size connected device manufacturer that has invested heavily in modern CI/CD practices for its cloud backend, automated testing, staged deployments, instant rollback, while its firmware team still ships updates to deployed hardware through a bespoke internal tool built years earlier, with manual approval steps and no automated rollback if an update bricks a device in the field. When a critical security vulnerability is discovered in a shared library used by both the cloud backend and the device firmware, the cloud team can patch and redeploy within a day using its existing pipeline. The firmware team, lacking equivalent tooling, might take weeks to validate, package, and begin rolling out an equivalent fix to tens of thousands of field devices, during which the vulnerability remains exploitable on every device still running old firmware. That gap, not a difference in engineering talent or organizational priority, but a genuine tooling and process gap between how cloud and device software get shipped, is precisely what unified device management platforms are built to close.
Business Value for South Bay IoT Companies
For hardware and IoT companies across the South Bay, this consolidation trend changes how device management should be evaluated during platform selection. A company treating device management as a narrow, standalone tool risks ending up with exactly the kind of siloed, manually-intensive update process that creates both operational risk and security exposure. Evaluating device management capability as an extension of an organization's existing software delivery pipeline, rather than a separate procurement decision handled by a different team, tends to produce more secure, more reliable outcomes, since the same testing, staging, and rollback discipline a company already applies to its other software gets extended naturally to its device fleet rather than reinvented from scratch.
Implementation Advice for IoT Device Fleet Management
- Treat firmware updates as part of the existing software delivery pipeline, not a separate process, applying the same staging, testing, and rollback discipline already used for other software deployments.
- Prioritize staged rollouts over fleet-wide updates, limiting a problematic update's exposure to a small subset of devices before expanding to the full fleet, a standard practice in modern software deployment that remains inconsistently applied to IoT firmware specifically.
- Build automatic rollback capability from the start, since a failed update that leaves a device in an unrecoverable state, requiring a physical site visit to fix, carries a much higher cost than an equivalent failure in cloud software that can simply be redeployed remotely.
- Integrate security scanning into the firmware pipeline itself, catching known vulnerabilities before a build ships rather than discovering them after devices are already deployed in the field.
Common Challenges in IoT Device Management
| Challenge | Description |
|---|---|
| Connectivity gaps in the field | Devices in remote or industrial locations may have intermittent connectivity, complicating reliable update delivery and confirmation |
| Diverse hardware and firmware generations | Large fleets often include devices spanning multiple hardware revisions, each requiring compatible but distinct update packages |
| Bricking risk from failed updates | A failed firmware update without robust rollback protection can leave a physical device unusable without a site visit |
| Scale of the attack surface | A fleet of thousands or millions of distributed devices represents meaningfully more attack surface than centralized cloud infrastructure alone |
These challenges connect directly to the embedded AI deployment pipeline covered elsewhere on this site, since models and firmware increasingly need to be managed, updated, and secured through the same kind of unified pipeline discipline, whether the payload being deployed is a traditional firmware update or a retrained machine learning model.
What's Next for IoT Device Management
As agentic AIoT systems increasingly make autonomous decisions based on deployed models, the stakes around reliable, secure device and firmware management continue rising, since an autonomous system acting on outdated or compromised firmware carries meaningfully higher risk than a passive monitoring device with the same vulnerability. For South Bay IoT companies, the practical lesson from the consolidation already underway in the device management space is that the discipline and infrastructure a company already applies to its cloud software delivery is no longer optional for its device fleet either, it is becoming the baseline expectation for any organization managing IoT devices at meaningful scale.
