# Onsemi bets $7 billion in stock on a world where AI runs at the edge, not the cloud > Source: > Published: 2026-06-25 22:51:21+00:00 *Onsemi's $7 billion Synaptics deal is a bet that AI won't stay bottled up in data centers. The company is paying in stock because the edge-AI market it wants is still more promise than proof.* Onsemi is trying to buy its way into the next AI hardware fight, and it isn't the one Nvidia has already made famous. The Scottsdale chipmaker said on June 25 that it will acquire Synaptics in an all-stock transaction with a total enterprise value of about $7 billion, giving Synaptics shareholders 1.35 onsemi shares for each share they own. According to The Wall Street Journal's report on the announcement, that exchange ratio represents a roughly 19% premium based on the companies' 10-day volume-weighted average prices. The deal is expected to close in mid-2027, subject to regulatory approval, and Synaptics would become a wholly owned onsemi subsidiary. Former Synaptics shareholders are expected to own about 12% of the combined company. The numbers are clean. The argument is messier, and more interesting. Hassane El-Khoury, onsemi's chief executive, framed the deal around AI moving out of the cloud and into the physical world, including automotive and industrial systems. His line from the announcement is the one you need to watch: the next phase of innovation will depend on systems that can “sense, decide, act and adapt in real time.” That isn't a pitch for training bigger models. It's a pitch for inference in cars, factory equipment, robots, home devices, and machines that can't wait for a round trip to a data center before making a decision. Onsemi already sells into power semiconductors and sensing, especially in automotive and industrial markets. Synaptics brings a different set of parts: edge-AI processors, neural processing units, wireless connectivity, mixed-signal technology, capacitive touch, biometrics, and high-speed interfaces. Put less neatly, onsemi is trying to stitch together the pieces a device needs when it has to see something, process it locally, connect to other systems, and act without wasting power. The company calls that “Physical AI.” Frankly, the phrase is doing a lot of work, but there is a real business argument underneath it. Onsemi says the combination expands its total addressable market to $243 billion by 2030, an increase of $30 billion. It also expects about $200 million in annual cost and revenue synergies, with non-GAAP earnings per share accretion within 18 months after close. Investors didn't cheer. Barron's reported that onsemi shares fell 8.3% to $108.85 in after-hours trading after the announcement, while Synaptics shares rose 11% to $140. The Wall Street Journal put the after-hours drop at 7.3% to $110.12. Either way, the market's message was plain enough: buyers may like the AI story, but they don't love dilution and execution risk on a transaction that won't close until the middle of next year. You can see why. This is an all-stock deal for a company whose own end markets have not been roaring. Automotive and industrial chip demand has been uneven, and onsemi is asking shareholders to accept a long-dated bet on a market that still needs to prove how quickly it will scale. Paying cash for that would have been a harder sell. Paying in stock spreads the risk across both sides. That structure is not cowardice. It is discipline. If edge AI becomes the next serious semiconductor cycle, Synaptics holders ride the upside through onsemi stock. If the category takes longer, onsemi has not emptied the balance sheet to chase a slogan. The timing also tells you something. Qualcomm agreed this week to buy AI software company Modular for about $3.9 billion, according to The Wall Street Journal, in a stock-based deal aimed at helping AI run across different hardware platforms. That is a software-heavy move. Onsemi's Synaptics deal is more physical: sensors, processors, connectivity, control. Same broad AI land grab, different layer of the stack. The harder question is whether onsemi is early or simply hopeful. Synaptics gives it credible technology for local AI processing, but customers still have to buy the premise in volume. A factory operator, automaker, or device maker does not buy “Physical AI” because the phrase sounds useful. They buy chips when local inference lowers latency, cuts power use, improves reliability, or lets a product do something it couldn't do before. That is where this deal will be judged. Not in the press-release language, and not in the first after-hours stock move. The test is whether onsemi can turn Synaptics' compute and connectivity assets into design wins inside vehicles, industrial systems, and smart devices before the market loses patience with every company trying to attach itself to AI. For now, the deal is current, concrete, and large enough to matter. A $7 billion stock transaction with a mid-2027 closing date is not a side project. It is onsemi saying that the data-center AI boom has a second act, and that the second act happens closer to the machine than the server rack. **Also read:** [General Intuition raises $320 million on the thesis that video game footage is the most underrated training data in robotics](https://startupfortune.com/general-intuition-raises-320-million-on-the-thesis-that-video-game-footage-is-the-most-underrated-training-data-in-robotics/) • [OpenAI is leaning toward a 2027 IPO as its CFO warns the company isn't ready for public markets](https://startupfortune.com/openai-is-leaning-toward-a-2027-ipo-as-its-cfo-warns-the-company-isnt-ready-for-public-markets/) • [Most Singaporeans with anxiety or depression never seek professional help but say they would welcome peer support](https://startupfortune.com/most-singaporeans-with-anxiety-or-depression-never-seek-professional-help-but-say-they-would-welcome-peer-support/)