Qualcomm Shares Extend Decline as Analyst Upgrades Are Undermined by Execution Doubts

The stock has been sliding sharply despite a flurry of analyst upgrades triggered by Qualcomm’s latest investor day. Over the past seven days, shares lost nearly 15%, extending a broader rout that has wiped out more than 23% in the last 30 days. At around €165.6, the equity now sits roughly 26% below its 52-week high of €222.90 reached in May.

The June 24 investor event was meant to rally the bulls. Management unveiled ambitious revenue targets: $5 billion from the data-center business by fiscal 2027, and over $15 billion by 2029. The broader non-smartphone revenue goal was raised to $40 billion for 2029. An automotive pipeline of $65 billion in orders was also touted. The immediate reaction was a near 10% spike, but that gain quickly evaporated as scepticism reasserted itself.

Wall Street’s response has been anything but uniform. Morgan Stanley reversed its stance, upgrading Qualcomm from Underweight to Equal Weight and lifting its price target from $146 to $231. Analyst Joseph Moore conceded the bank had underestimated Qualcomm’s AI ambitions for too long. Yet even this upgrade came with caution: Morgan Stanley labelled the long-term data-center goal of $15 billion as “rather ambitious” and flagged execution risks around accelerators and server CPUs, especially with Nvidia, AMD, Intel, and hyperscaler in-house chips looming.

Other houses remain more bearish. Bernstein kept a Hold rating with a $235 target, arguing Qualcomm’s platform lacks the breadth of Nvidia’s ecosystem. Barclays stayed at Underweight, raising its target from $150 to $245 but calling Qualcomm a “show-me story in a fiercely contested market.” Bank of America nudged its target up to $220 while retaining an Underperform rating, warning that the most important growth initiatives have yet to prove themselves and that the stock already prices in substantial data-center success. UBS moved its target to $235 but held at Neutral. The highest target comes from Benchmark at $300, with RBC Capital at $250, DZ Bank at $265 (upgraded to Buy), and Rosenblatt also at $265, hailing the investor day as a “definitive turning point.” The target range is unusually wide, from $300 down to $100.

Should investors sell immediately? Or is it worth buying Qualcomm?

The analyst community as a whole is deeply divided. One survey showed 12 buy ratings, 22 holds, and 5 sells, while another count from a different period reflected 8 buys, 29 holds, and 7 sells — underscoring the lack of a clear consensus.

At the heart of the debate is Qualcomm’s technology roadmap and the timeline to revenue. The company unveiled its High Bandwidth Compute architecture, which stacks memory and processing units vertically to accelerate data flow. Executive Vice President Durga Malladi confirmed that the first generation will debut in data centers in 2027, with commercial availability in 2028, and then be ported to smartphones, PCs, and cars. The chips driving the push — the Dragonfly C1000 server CPU and the Dragonfly AI300 inference accelerator — will not enter production until 2028 at the earliest. That means investors are being asked to fund a roadmap that generates no meaningful data-center revenue for years.

Meanwhile, Qualcomm’s core smartphone business, which still accounts for roughly two-thirds of product sales, faces structural headwinds. Apple is expected to switch to its own modems as early as 2027, a move widely seen as the biggest structural risk. Morgan Stanley noted two custom-silicon deals with hyperscalers worth over $1 billion each, as well as agreements with Meta and Modular, but Bernstein warned that a weaker handset business could weigh on near-term earnings while the data-center segment remains a promise.

The next key milestones are a virtual investor meeting hosted by Benchmark on June 30, followed by Qualcomm’s quarterly results on July 29. That earnings release will give management its first real opportunity to quantify concrete progress in the data-center ramp. Until then, the market appears to be betting that execution risk outweighs the potential of an ambitious — but distant — AI story.