04 | 2026 Jari Huilla, CTO & partner Kipinä
Artificial intelligence is currently a ubiquitous topic. Recent weeks have seen widespread discussion regarding increasing usage restrictions, rising costs, and the apparent degradation of previously capable models when accessed via cloud services and APIs. This raises legitimate concerns for many: can businesses safely build their operations on AI if pricing and availability are subject to uncontrolled fluctuations?
Fortunately, AI models can also be run locally—on your own hardware and within your own environment. Why does this matter?
1. If you own your AI, no one can take it away from you
When running open models on hardware you own, no external provider can suddenly restrict your access to the service.
When the model runs within your own environment:
usage is independent of third-party services
data remains strictly within the organization
information security is subject to granular control
This is not merely a technical choice, but a strategic decision tied to business risk.
2. Models don't degrade on their own
Models provided via cloud services sometimes initially seem smart and capable, only to appear degraded over the subsequent weeks. At the time of writing, this discussion has primarily centered around Anthropic's Claude Opus versions 4.6 and 4.7.
Locally executed models remain unchanged unless you modify them yourself.
3. Value is generated from a combination—not a single technology
Let's be clear: local AI is not the answer to everything, and both acquiring and maintaining the necessary hardware require resources. Consequently, local AI should be viewed as part of a broader ecosystem. The smartest solutions are built by combining:
open models executed locally on dedicated hardware (ensuring control and data security)
cloud services provided by AI laboratories (offering scale and performance)
managed model platforms from public cloud providers (providing access to frontier models without exposing data to AI laboratories)
open models executed on rented infrastructure (enabling the deployment of even the largest open-source models)
The key is the capability to architect the right combination for the specific use case. For example, most information security and data privacy requirements can be met with the appropriate combination.
What does this mean in practice?
Local AI is not for everyone—nor for every use case—but it signals a broader shift:
toward more decentralized AI
toward stronger data governance
toward more pragmatic use cases
And perhaps most importantly: toward a world where AI is not just an outsourced service, but a core internal capability.
If you want to understand how locally run models actually function in practice (in a Linux environment, step-by-step), we recommend reading Jari's original, more technical article: Running local AI models on Linux
Jari Huilla, CTO & Partner, Kipinä
Jari Huilla is Kipinä's CTO, possessing an exceptionally long and diverse background in technology; his first role in the industry was at Nokia Research Center at the age of 15. Over the years, he has served as a developer, leader, and growth company builder. At Kipinä, Jari bridges deep technical expertise with business-driven thinking. He is particularly passionate about making AI solutions not only technically functional but genuinely purposeful—and ensuring their limitations are fully understood rather than simply bypassed.