NEW YORK, June 24, 2026, 17:22 (EDT)
- Qualcomm finished the session down 3.3% at $197.41, slipping under the $197.50 average analyst target from MarketWatch. Shares then jumped more than 12% after hours after Reuters reported on its data-center outlook.
- Qualcomm now aims for over $15 billion in data-center revenue by fiscal 2029 and raised its non-handset chip revenue goal to $40 billion by the same year, nearly twice its earlier target.
- Qualcomm’s planned Modular acquisition would take up to 19.2 million shares—around $3.92 billion based on Tuesday’s close, and roughly 1.8% of its implied share count.
Qualcomm closed Wednesday regular trading near the Street’s average price target. Shares moved higher after the bell when the company released specific revenue figures tied to its AI data center efforts.
The stock closed down $6.71 at $197.41, wiping out about $7.2 billion in market value for the company, based on its share price and valuation. That’s close to double the stock value Qualcomm plans to pay for its Modular deal, an AI software firm. Shares reversed after the bell, jumping more than 12% after Qualcomm said it targets $15 billion in data-center sales by fiscal 2029, according to .
That swing is important because Qualcomm finished trading nine cents under the $197.50 average analyst target from MarketWatch. With a 12% jump after hours, the stock would trade above $221, well past that average target and closer to what bullish analysts expect than to the consensus. Qualcomm’s consensus rating is Hold, according to MarketWatch, which tracks 39 analyst calls.
Qualcomm told investors it expects non-handset chip revenue to hit $40 billion by fiscal 2029. The company is also aiming for over $15 billion from data centers, $10 billion from automotive, and non-GAAP EPS above $18 by fiscal 2029.
Qualcomm CFO Akash Palkhiwala told a presentation the company has plans to be “truly diversified,” Reuters reported. Palkhiwala said he expects the data-center segment to hit $5 billion in fiscal 2027, with $1 billion of that coming from new custom-chip clients. Reuters
Qualcomm shares slipped after the company put out an 8-K, planning to issue as many as 19.2 million shares to buy Modular. The private placement is part of the acquisition, Reuters said. The news service put the value of the deal at $3.92 billion using Qualcomm’s last closing price.
Investors snapped up shares on signs of big-name buyers. Qualcomm said Microsoft is set to use its High Bandwidth Compute chips, which are designed to handle more data for less power. Meta will go with Qualcomm’s Dragonfly C1000 CPU at its AI data centers. CPUs run the servers. The company also reported it has two unnamed hyperscalers lined up for custom chips, with those deals expected to deliver revenue before the end of the year.
Qualcomm’s data-center lead Tony Pialis told Reuters, “I have not had to push my way into hyperscale customers; they’ve been pulling us in.” Meta, in a separate statement, said Qualcomm’s Dragonfly C1000 CPU is set for its next-generation server fleet, with production kicking off in the second half of 2028. Reuters
Qualcomm’s Modular buy hands it inference software for running AI models on various chips. “Qualcomm is betting that by owning software that squeezes more inference more efficiently out of hardware, it can stake a claim in the data center market,” Emarketer analyst Jacob Bourne told Reuters. Reuters
Qualcomm is taking aim at Nvidia’s grip on AI software and chips, but isn’t claiming it will win on all fronts. Qualcomm said Modular’s platform can run models on CPU, GPU, NPU, and custom ASIC architectures. Custom ASICs are chips made for a customer or a certain workload. “AI needs a more open and efficient software foundation,” said Modular CEO Chris Lattner. Business Wire
Nvidia still runs the table on CUDA software and holds the main AI accelerator lead. Broadcom is now a big player in custom AI chips. Qualcomm, for its part, is pushing an angle on cheaper power via HBC and custom silicon instead of only chasing accelerator plans. Qualcomm claims HBC will give six times the bandwidth per watt over high-bandwidth memory, using competitor specs at the card level.
Timing and proof are the issues. Qualcomm’s HBC Gen 1 will sample in mid-2027, Dragonfly AI300 in 2028. Meta’s CPU program is set for second-half 2028 production. Bank of America analysts had cautioned before investor day that Qualcomm is stepping back into a busy AI sector where Nvidia, Cerebras, and custom chips from Amazon and Google are already active, Reuters said.
