Sunday, September 27, 2026
AI Infrastructure · News & Analysis
Home › Chips & Hardware › Report
Chips & Hardware · Report

Infineon is divesting from memory to fund a strategic pivot toward AI-related power and efficiency technology.

A major semiconductor vendor's deliberate shift to AI power-supply and thermal management validates growing infrastructure demand and repositions suppliers around power-constrained deployments.
Trade pressSlicast · September 26, 2026 at 09:10 UTC · US · Source: AD HOC NEWS
importance 50

Infineon is undertaking a rare portfolio maneuver during a cyclical downturn: divesting a legacy business to fund expansion into the power electronics essential to AI server infrastructure. In early September, the Munich-based chipmaker agreed to sell its NOR-Flash and F-RAM memory operations to Taiwan's Winbond Electronics for USD 1.12 billion in cash. The transaction is scheduled to close in the second half of 2027, subject to regulatory approval.

The memory unit generates roughly EUR 350 million in annual revenue and employs approximately 350 people across ten countries, according to Handelsblatt. Management frames the exit as overdue pruning. The memory business, serving automotive, industrial and infrastructure customers, offers limited differentiation, and the capital locked into it can yield higher returns elsewhere.

That reallocation is already underway. More than a month before the Winbond agreement, Infineon acquired India's C2i Semiconductors, a developer of software-defined multiphase controllers and smart power stages designed specifically for energy-intensive AI data centers. While the purchase price remained undisclosed, the strategic intent was transparent: accumulate the technologies that capture greater value in server power distribution. The deal is expected to close as early as the third quarter of 2026, underscoring management urgency.

Recent product announcements reinforce this direction. This week, Infineon introduced its PSOC Control C3 Performance Line of microcontrollers for real-time power and motor control applications, incorporating post-quantum cryptography compliance under the CNSA Suite 2.0 standard. The company previously launched a two-channel 120 V EiceDRIVER gate driver for data-center power supplies. Additional moves include a collaboration with SolarEdge on protective switch technology for 800 VDC networks and a scheduled webinar on battery-backup and capacitor-bank systems for AI data centers.

Investor sentiment, however, remains divided. Warburg Research upgraded the stock to "Buy" on September 7, maintaining an EUR 84 price target and citing AI data-center demand acceleration offsetting the recent correction. Morgan Stanley disagreed the following day, downgrading to "Equalweight" from "Overweight" and cutting its target to EUR 65 from EUR 81, citing persistent weakness in automotive and industrial demand. Oddo BHF sided with the bulls a week later, upgrading to "Outperform" with an EUR 80 target, while UBS held at "Neutral" with an EUR 64 target.

The stock closed Friday at EUR 57.22, up 52 percent year-to-date, suggesting much of the AI narrative is already reflected in valuation. Momentum has stalled: the shares dipped below their 50-day moving average of EUR 58.96 on Wednesday, and briefly hit the DAX floor on Thursday amid concerns about industrial demand and auto sector weakness before recovering to Friday's close. At 36 percent below its 52-week high, the stock still trades at a significant discount.

The core tension is one of timing. Infineon's automotive and industrial chip franchises remain its earnings foundation, and when customer orders soften there, balance-sheet strength offers limited insulation. The Winbond cash settlement will not arrive until 2027, and the data-center power business will require years to reach meaningful profit scale. Portfolio restructuring of this complexity diverts management attention and tests shareholder patience through inevitable volatility.

The underlying strategy, however, remains sound. Infineon is exiting a low-margin periphery while establishing positions in a structural growth market — the power architecture of AI infrastructure — where its technology already demonstrates traction. For now, near-term returns still depend on the traditional factory floor and vehicle production.

Read the original
Infineon is divesting from memory to fund a… · Slicast