Qualcomm unveils Dragonfly C1000, a more power-efficient data center CPU built for agentic AI, and says Meta will use it when Qualcomm starts production in 2028

Qualcomm shares jumped 15% in extended trading on Wednesday after the chipmaker said non-handset revenue in fiscal 2029 will be $40 billion, up from a prior forecast of $22 billion.

CNBC 2026-06-25 Kif Leswing

Context & Ripple Effects

Qualcomm’s Dragonfly C1000 follows its stated plan to re-enter data-center CPUs with a processor designed to link to Nvidia GPUs for AI workloads. It also extends a recent sequence of AI-infrastructure products, including inference chips scheduled for 2026 and 2027, beyond the company’s handset-chip base.

Meta’s planned use of the C1000 at production start gives Qualcomm an identified deployment path, while the company’s higher fiscal-2029 non-handset revenue target makes data-center and other adjacent businesses more central to its growth narrative.

First-order effects

Qualcomm gains a named prospective customer for its data-center CPU and a tangible reference point for the Dragonfly platform ahead of planned 2028 production.
The revised non-handset revenue outlook immediately raises the strategic importance investors and customers may assign to Qualcomm’s non-handset businesses, reflected in the reported after-hours share move.

Second-order effects

Intel and AMD, identified in prior coverage as incumbents in data-center CPUs, face another supplier positioning a CPU around AI systems rather than general-purpose server demand alone.
Because Qualcomm’s earlier plan emphasized connectivity with Nvidia GPUs, adoption of Dragonfly could strengthen demand for more tightly integrated CPU-GPU AI infrastructure rather than isolated accelerators or processors.

Third-order effects

If Qualcomm converts its announced AI chips, CPU roadmap, and Meta deployment into sustained volume, the data-center market could become less defined by standalone CPU competition and more by vendors’ ability to offer efficient components for AI clusters.
The broader business effect would be a further diversification of major mobile-chip suppliers into data-center, industrial, and robotics AI markets; execution and customer uptake remain the key constraints in the coverage provided.

The trend: This is one data point in the shift of AI silicon competition from handset and edge inference toward power-conscious, integrated data-center infrastructure for agentic workloads.

Discussion

@danielnewmanuv Daniel Newman on x
$META and $MSFT commit to multi-generation Datacenter commitments with $QCOM. A lot of people said Qualcomm couldn't enter this market successfully. Those people appear to be wrong. 👀😬 [image]
@qualcomm @qualcomm on x
Today at #QCOMInvestorDay: Qualcomm Dragonfly debuts with C1000 CPU, AI300 accelerators, data center DSPs and a new near-memory computing architecture: Qualcomm High Bandwidth Compute. Plus, custom silicon for the data center. Learn more:
@qualcomm @qualcomm on x
This is AI at data center scale. Qualcomm Dragonfly is here, engineered for agentic workloads, with leading power efficiency across integrated platforms, and hyperscaler momentum already building. Where do you want to go? #QCOMInvestorDay [video]
@patrickmoorhead Patrick Moorhead on x
Some more deets on $META adopting datacenter $QCOM CPUs in 2H/28 from the press release: -strategic multi-generation collaboration for Qualcomm Technologies to be a supplier for data center CPUs for Meta. (Note “collaboaration ” which is MOU versus contract) -the Qualcomm Dragon…
@andrewcurran_ Andrew Curran on x
Mark Zuckerberg's future datacenters will be powered by the newly announced Qualcomm Dragonfly C1000. [image]
@patrickmoorhead Patrick Moorhead on x
Microsoft deploying Qualcomm-based HBC XPU solutions inside Azure according to Microsoft CEO @satyanadella. I'd say that is big. $QCOM $MSFT [image]
@mytechmusings Prakash Sangam on x
First major announcement)after @Modular acquisition) is here... #C1000 CPU by @Qualcomm for #AgenticAI is here... @Qualcomm claims 2x performance & $200B TAM, & endorsement of @Meta with multi-gen collaboration (worst kept secret) $qcom $qcominvestorday [image]
@patrickmoorhead Patrick Moorhead on x
Not one, but three different Qualcomm datacenter CPUs. One with HBC. $META CEO @finkd says they have a MOU to use them. $QCOM [image]
@benbajarin Ben Bajarin on x
Qualcomm C1000 purpose built agentic data center CPU. While we have the total CPU TAM going to 180-200B in 2030 we think dedicated agentic CPU is 45-55B in 2030. $qcom [image]
@davealtavilla David Altavilla on x
.@Qualcomm EVP Tony Pialis steps through what the company claims will be the fastest Cloud Native CPU in the data center, $QCOM C100, along with a multi-generational agreement with @Meta for deployment. Big news! [image]
@ryanshrout Ryan Shrout on x
@Qualcomm ... Next surprise: Zuckerberg and Meta! Agreement to use their CPUs across multiple generations. [image]
@clattner_llvm Chris Lattner on x
I'm excited to share that Qualcomm is acquiring Modular: this will accelerate our path to unifying accelerated compute with an open platform. This will also mark a new era in open software development for Qualcomm. 👇
@modular @modular on x
We're excited to announce that Modular has entered an agreement to be acquired by @Qualcomm. The future of unified compute has never been stronger. Read the full announcement: https://www.modular.com/...
@waterloo_intern Ali on x
on one hand i'm stoked for @Qualcomm, they're getting, literally speaking, one of the best inference, compilers, and kernel team and stack in the world but...sad that this might cause it to converge from the world's only hardware-agnostic inference stack... to another hardware
@generalcatalyst @generalcatalyst on x
Through its agreement to be acquired by @Qualcomm, @Modular aims to accelerate its mission and real-world impact by empowering developers everywhere with production-grade AI software. Congratulations to @clattner_llvm, @iamtimdavis, and the extended Modular team on their [image]
@melt_dem Meltem Demirors on x
congrats to @Modular, excited to see Mojo become part of the Qualcomm ecosystem the great wave of vertical integration is beginning - hardware players moving into compilers and orchestration software, lenders moving into co lo operations and cloud the spice must flow 🤑
@patrickmoorhead Patrick Moorhead on x
I'm interested to see what $QCOM does with Modular. Qualcomm is very good at edge enabling software, but that's not the same as datacenter software capability. Strategically, this could help to better answer the datacenter question. I'm wondering if this is more of an aquihire…
@ryanshrout Ryan Shrout on x
This a great move for @Qualcomm. Not only does it get a great software platform that is legitimately one of the best for workload disaggregation, but also gets one of the best minds in AI software as part of the deal. Can't wait to hear what the company says about this today.
@__tinygrad__ @__tinygrad__ on x
RIP, no future in Mojo now. Qualcomm is not a good steward for open source software. Modular raised too much money, should have stayed leaner to have a chance at winning.
@austinsemis Austin Lyons on x
Wow! Congrats @clattner_llvm and team and @Qualcomm.
@qualcomm @qualcomm on x
Hardware plus software defines leadership in the AI era. Today we announced an agreement to acquire @Modular, advancing our evolution as a developer-first AI solutions company delivering generative and agentic AI from edge to cloud. https://www.qualcomm.com/... [image]
@davemuni Dave Munichiello on x
1/ I am honored to congratulate @clattner_llvm, @iamtimdavis, and the team at @Modular on the announcement that they have signed a definitive agreement to be acquired by @Qualcomm.
Dave Munichiello Dave Munichiello on linkedin
I am thrilled to congratulate Chris Lattner, Tim Davis, and the entire team at Modular on the announcement that they have signed a definitive agreement …
Lauren Goode Lauren Goode on linkedin
NEWS: Qualcomm has acquired Modular, one of the most promising chip software startups of the AI era. …
@benbajarin Ben Bajarin on x
Known for a long time Qualcomm has a deep and wide supply chain. This slide captures a lot of what Qualcomm has in supply chain leverage for customers who need it (I.e data center compute customers) 1m+ a year in leading node wafers. 2.5+m total wafers a year. $qcom [image]
@benbajarin Ben Bajarin on x
Finally details of Qualcomm custom asic/silicon business. Multi generational wins with hyperscalers contributing revenue this year. $qcom [image]
@thetranscript_ @thetranscript_ on x
$QCOM Qualcomm has unveiled an expanded diversification strategy at Investor Day, raising its FY29 non-handset revenue target to $40B, nearly 2x its prior target, with new goals of >$15B in data center revenue, $10B in auto revenue & >$14B in IoT revenue by FY29. $QCOM: +13% AH […
@patrickmoorhead Patrick Moorhead on x
Satya's exact words around Qualcomm HBC: - we are excited about your innovation in the data center, especially around high bandwidth compute, and we look forward to building on that together. - HBC implements an innovative architecture with high memory bandwidth and integrated
@danielnewmanuv Daniel Newman on x
$QCOM CFO presentation was the unlock. Up almost 10% on new revenue forecast with $15 Billion in expected datacenter revenue by FY2029. Like I've said. Qualcomm can win big by winning small. Just a 1-5% share of a trillion+ dollar TAM in datacenter infrastructure gets it there. […
@benbajarin Ben Bajarin on x
We have been slowly building our own model of silicon content per robot but this idea of 4-5 SoCs, LOTs of fiber/interconnect, and an insane amount of analog/sensor is inbound. These slides from Qualcomm are helpful. Toward end of decade humanoids start to scale. $QCOM [image]
@patrickmoorhead Patrick Moorhead on x
$QCOM CEO @cristianoamon kicks off its 2026 investor day: • Opened by framing this as a new chapter for Qualcomm and thanked the team and executives. • Noted the timing: this Investor Day marks five years since becoming CEO. • Said Qualcomm moved from a mobile-focused [image]
@benbajarin Ben Bajarin on x
Signs of the times with a lot of the exec QA here at Qualcomm investor day. A company modeled to grow high single to low double digits a year is now signaling their business to double, much in part by AI data center growth. $qcom
@stocksavvyshay Shay Boloor on x
$META plans to use $QCOM C1000 CPU in its data centers giving the company a major hyperscaler design-in for its AI data center push. Meta is ideal customer because it spends aggressively and every reason to diversify compute as inference workloads make CPUs strategic again. [imag…
@firstadopter Tae Kim on x
Both customer video testimonials first mentioned using Qualcomm chips in flat-out, long-run failures in the market: - Microsoft CEO: Windows on Arm laptops using Qualcomm processors - Meta CEO: Metaverse VR headsets using Qualcomm processors But don't worry, guys. This time

Chronicles