Generative AI is dramatically changing the Internet of Things (IoT) by shifting devices from being simple data collectors to intelligent systems that can understand context, anticipate needs, and respond in real time. Yet many IoT deployments still depend on cloud processing, which introduces latency, consumes bandwidth, and raises privacy concerns.
MediaTek is addressing those challenges by moving AI processing closer to where data is created. Through its Genio platform, the company embeds advanced AI compute and generative capabilities directly into Edge IoT hardware. This edge-first approach reduces reliance on the cloud, improves responsiveness, and gives manufacturers a more straightforward path from design to deployment.
Sameer Sharma, Associate Vice President of AIoT and General Manager for North America and Europe at MediaTek, says the industry’s shift to edge processing is already well underway. “Edge AI shifts decision-making to right where the request happens, and the data is created,” he says. “Instead of sending endless video or sensor data upstream, devices can process information locally and transmit only what’s relevant, solving the concerns about responsiveness, economics, sustainability, and privacy.”
He adds that early adopters are already seeing results. “One partner is using the MediaTek Genio 700 SoC to create a factory monitoring solution capable of AI-driven safety inspections in real time. Another is deploying the more powerful MediaTek Genio 1200 in rugged industrial environments, where local inference reduced latency and ensured reliability even in bandwidth-constrained conditions.” MediaTek’s recently launched Genio 520-720 offer up to 10 TOPS to support a breadth of AI applications.
Bringing AI closer to the data
Edge AI lets IoT devices equipped with integrated NPUs process data locally, avoiding roundtrips to the cloud, reducing latency and improving responsiveness.
This shift is already benefiting IoT applications across various sectors. For example, smart cameras analyze video on the device. Industrial sensors run predictive maintenance models locally, helping equipment remain reliable even with inconsistent connectivity. Retail kiosks personalize interactions while keeping sensitive customer data on the device. In each example, edge AI improves performance while reducing bandwidth, costs and privacy risks.
Accelerating development timelines
A significant barrier for device manufacturers has been the time it takes to build and validate new IoT products. The Genio platform addresses that challenge with flexible frameworks, developer tools and software stacks that help teams move from concept to production in weeks instead of months.
MediaTek’s broad support for different operating systems and AI frameworks such as nVidia TAO also streamlines development across product lines. Sharma says this is intentional. “The future of IoT depends on bringing powerful AI capabilities directly to devices,” he says. “And that future is already taking shape with MediaTek Genio.”
Partnerships and ecosystems that scale
Because a robust partner ecosystem is central to scaling IoT innovation, MediaTek collaborates with cloud providers, AI software vendors, wireless technology companies, and a growing developer community to ensure interoperability and faster time to market.
These partnerships help solution providers validate software on proven hardware, integrate domain-specific models, and bring complete IoT systems to market more quickly.
Sharma says this ecosystem is foundational to MediaTek’s strategy. The company provides low-power Genio computing platforms, connectivity options that include Wi-Fi 6E and 7, 5G and REDCap, and complete software solutions across Android and Linux. The ecosystem also provides evaluation kits, modules, and service partners that support both design and deployment.
Strategic collaborations further accelerate adoption. The MediaTek and Nvidia partnership, for example, gives OEMs access to industry-proven TAO models that can run directly on Genio devices. Sharma says this lets developers focus on higher-level AI applications rather than spending resources building foundational models from scratch.
Real-world momentum across industries
Industries that rely on responsiveness, security, and local decision-making are adopting edge AI quickly, and one of the fastest movers is retail. “The retailers pulling ahead are those using artificial intelligence not as a bolt-on enhancement, but as a strategic capability woven into every customer touchpoint and operational process,” says Sharma.
Edge AI lies at the heart of this transformation, he says. Improvements such as smart shelves, predictive replenishment, adaptive digital signage and conversational kiosks all depend on high-performance, low-latency computing that happens close to the customer. “AI is transforming retail from reactive to anticipatory,” he says.
Industrial automation, healthcare, transportation and smart cities are also expanding their use of edge AI across areas such as predictive maintenance and privacy-preserving diagnostics, as well as intelligent logistics and data-driven infrastructure.
Looking ahead toward agentic AI
Generative AI at the edge is also laying the groundwork for agentic AI, where devices can reason, plan, and take action with minimal human input. In retail, smaller domain-specific language models trained on proprietary data can run locally on edge systems. Sharma explains that these models can understand a retailer’s catalog and brand voice while keeping customer data private.
Agentic systems will build on these capabilities to create more autonomous, context-aware IoT environments.
MediaTek’s vision for the future
“AIoT is now being defined by devices that are not only connected, but also intelligent, responsive, and sustainable,” says Sharma. “For OEMs, developers, and enterprises, the opportunity is clear. Moving AI to the edge opens new capabilities, new business models, and new productivity efficiencies.”
MediaTek’s long-term strategy is to combine silicon innovation with ecosystem depth, so customers can bring AI-enabled devices to market more quickly and efficiently. As edge intelligence advances, the company sees IoT systems becoming increasingly capable, responsive and autonomous.