Silicon Labs, an Austin-based company that makes chips for networked devices like appliances and security systems, is preparing for the growing role of artificial intelligence in everyday life.
At its annual Works With conference in downtown Austin on Thursday, the company introduced a new software platform built to support both human developers and AI coding assistants.
“We have been optimising our tools, documentation, and workflows for humans for years,” Silicon Labs software development executive Manish Kothari said. But now a new component has surfaced. AI is evolving into a developer. Agents and co-pilots are starting to integrate, debug, and code alongside humans.
It is now our job to create a platform that works as well with these agents as it does with you. This is the core of “Simplicity AI,” a developer experience designed for people and the AI agents that work alongside them.
CEO Matt Johnson asserts that companies must maintain their lead in the rapidly evolving field of artificial intelligence. To prepare for the future, Silicon Labs is focusing on AI integrations.
The company specialises in embedded chips, which are hidden inside devices like smart lightbulbs, fitness trackers, thermostats, and even industrial machinery. Johnson also sees an opportunity to directly incorporate AI into these devices that are connected to the internet.
Although most people associate AI with chatbots or massive data centres full of servers, Johnson envisions AI-enabled operations that are carried out locally on chips within devices. By processing data locally, these chips can reduce the amount of data sent to distant data centres, boost productivity, and get rid of lag.
“I don’t have to wait for the data to travel all the way to the data centre and back,” Johnson said. You don’t have to cope with the latency of local processing, which makes some things more efficient. As a result, small processors, such as the ones we produce, will be able to carry out specific functions locally on the chip and then send only the necessary data—not the large or unclean amount of data—all the way back to the data centre.
Silicon Labs is preparing for this shift with this week’s conference announcement of its new AI development kit and Series 3 product line. As part of a six-product line, the new chips are designed to increase processing power and security, laying the groundwork for what Johnson calls the “AI at the edge” era.
AI will have a huge impact on the edge. “Take a look at the hype around data centres,” Johnson said. In time, that excitement will drive things to the edge. We also don’t do anything else. We are the biggest company in the world that focuses on it. You can see and feel both.
Semiconductor on tariffs
Johnson has been spending a lot more time in Washington than he ever expected since former President Donald Trump threatened to impose tariffs on chips earlier this year.
“I like most days, but not every day,” Johnson said. I never thought I would have to spend so much time in Washington talking about tariffs and engaging with lawmakers. Not in a bad way, but I just didn’t think this would have an impact on our industry.
Trump has since stepped up his trade policies, proposing a 100% tariff on semiconductor companies that do not manufacture all of their chips in the US instead of China. Some large companies, like Apple and Samsung, have already begun reshoring some production in response to market shifts and political pressure.
It is challenging for a company like Silicon Labs, which primarily relies on manufacturing partners in Taiwan and Singapore, to adjust to such a policy. Although the company has a “path” towards production in the United States, Johnson said, the technology needed to manufacture Silicon Labs’ products is not yet available in the nation, and the “one size fits all” tariff strategy is ineffective in the semiconductor industry.
“They haven’t developed the technologies we need, or they’re not yet production ready (in the U.S.),” Johnson said. We can say with confidence, ‘Well, we have a way,’ and it’s great to see those being established. They want to start manufacturing here as soon as feasible, but that presents a challenge because it takes time. Semiconductors have notoriously long lead times for developing a product, building a fab, and ramping it up. Convincing Washington to recognise the basic physics of our space instead of the urgency of results is the challenging part.
