Arm Shares Down 43% Face Sub-$100 Risk as Physical AI Division Launches

ARMARM

Arm shares have plunged about 43% from record highs and risk falling below $100 per share due to SoftBank’s margin loan exposure. The company has launched a Physical AI division targeting humanoid robots and autonomous vehicles to expand its growth footprint and justify its multiple.

1. Susquehanna Upgrade Sparks Rally

Susquehanna analyst Christopher Rolland raised the firm’s view on Arm Holdings to Positive with a 12-month price target implying more than 25% upside, citing recent changes in the company’s go-to-market initiatives. The upgrade followed Arm’s announcement of new licensing options tailored for hyperscalers and edge device makers, which Rolland believes could add incremental royalty revenue of $150 million to $200 million in fiscal year 2027. Shares jumped over 8% on the day of the upgrade, marking the largest one-day gain since April.

2. Shares Down Nearly 40% from 52-Week Highs

Arm’s stock has slumped roughly 38% from its peak reached last December, when optimism around artificial intelligence and the pending physical AI division drove valuations to multi-year highs. Concerns around valuation and competition from in-house chip development by major customers have weighed heavily on sentiment, while SoftBank’s margin loan against its Arm stake remains a potential overhang. Broader chip sector momentum has left Arm trailing peers that focus on memory and equipment, even as overall semiconductor indices gained more than 15% in the past quarter.

3. Physical AI Division Could Drive Next Growth Wave

At CES 2026, Arm formally launched its Physical AI division, consolidating internal teams working on robotics, autonomous machines and smart sensor applications. The new unit aims to address a market projected to exceed $45 billion by 2030, according to Arm’s latest white paper. Management has set an internal goal of securing at least two design wins with major robotics OEMs by the end of the fiscal year, which could translate into annual royalties of $50 million to $75 million if achieved.

4. CEO Outlines Energy Scaling Challenges for AI

CEO René Haas told CNBC at the World Economic Forum in Davos that while the U.S. power grid has sufficient capacity for projected AI workloads, the real bottleneck lies in transmission infrastructure and permitting delays. He noted that data center operators can face lead times of 18 to 24 months to secure grid interconnections, potentially slowing deployment of next-generation AI clusters. Haas indicated that Arm is collaborating with utilities and industry groups to develop more efficient power management architectures and streamline approvals.

Sources

YY2