Qualcomm courts AI device partners

Qualcomm is trying to make sure it wins the next device shift even if nobody knows what the next device is yet. That is the real story here. Cristiano Amon is openly talking about secret AI hardware projects with OpenAI, Meta, and “pretty much all” the big AI players, while Qualcomm is also rolling out new midrange phone chips for the market it already knows how to serve. The company is basically saying: if AI escapes the smartphone, we want to power that too. If it doesn’t, we still sell the silicon inside the phone. ### What did Amon actually say? On a podcast clip that started circulating this week and in follow-on interviews, Amon said Qualcomm is working with “pretty much all” major AI companies on undisclosed devices. He named OpenAI and Meta as examples and said some of the form factors are still secret. The important part is not one gadget leak. It is that Qualcomm wants to be the chip supplier under a whole category of AI-first hardware before the category even settles. (finance.yahoo.com) ### Why is Qualcomm talking this way now? Because the smartphone market is no longer enough of a growth story by itself. Qualcomm still makes huge money from handset chips, but investors have been pushing every chip company to explain its AI plan in concrete hardware terms. Amon’s answer is that phones will stay important, but wearables, glasses, and other ambient devices could become the next computing layer over the next five years. (benzinga.com) That framing turns Qualcomm from “phone chip vendor” into “default silicon partner for AI devices.” ### What happened on the phone side? Qualcomm did not abandon phones while talking about the post-phone world. On May 7, it announced Snapdragon 6 Gen 5 and Snapdragon 4 Gen 5, two chips aimed at cheaper Android devices. The pitch is familiar but useful — better gaming, smoother everyday performance, and more on-device AI camera features without pushing buyers into flagship prices. That keeps Qualcomm present where most phone volume still lives. (msn.com) ### What is notable about those chips? The details tell you Qualcomm is trying to normalize premium features in lower tiers. Snapdragon 6 Gen 5 adds Wi‑Fi 7 support and Qualcomm’s Adaptive Performance Engine 4.0. Snapdragon 4 Gen 5 is pitched as the first 4-series chip capable of 90fps gaming, with a reported 77% GPU jump over its predecessor. Both are built on 4nm and are expected in devices in the second half of 2026. (qualcomm.com) ### What about the 2nm and 3nm talk? That part is much shakier. There are fresh reports that Qualcomm may offer a more layered high-end lineup, with some flagship parts on 2nm and cheaper premium options on 3nm. But that appears to be rumor and supply-chain chatter, not something Qualcomm has formally announced this week. The broader idea still makes sense — node costs are rising, and vendors want more pricing flexibility — but the specifics should be treated as provisional for now. (qualcomm.com) ### Why does this two-track strategy matter? Because nobody has proved what an AI-native consumer device looks like yet. Humane stumbled. Rabbit did not become a platform. Smart glasses look more plausible, but even there the winning version is not obvious. So Qualcomm is doing the sensible thing — keep feeding the giant existing phone market while seeding chips into every weird, early AI endpoint that might break out. Think of it as selling shovels during a land rush, but also keeping your hardware store open. (wccftech.com) ### Who benefits if Qualcomm is right? OpenAI, Meta, and other AI companies get a partner that already knows radios, power management, and edge AI — the boring hard parts that make a device wearable instead of just demoable. Android phone brands benefit too, because Qualcomm keeps pushing newer features down into cheaper tiers instead of saving everything for $1,000 phones. That makes the company useful whether the future is glasses, pendants, earbuds, or just better midrange handsets. (finance.yahoo.com) ### Bottom line Qualcomm is not betting on one killer AI gadget. It is betting that the next wave will need low-power chips, connectivity, and on-device inference across many shapes — and that being everywhere beats guessing perfectly. (finance.yahoo.com) (qualcomm.com)