Enterprise
AI Demand Drives 600,000 Component Obsolescence Events in 2025
Chipmakers are retiring profitable legacy parts early to free capacity for high-bandwidth memory and AI accelerators, forcing industrial buyers to rethink lifecycle planning.
Key takeaways
- Over 600,000 electronic components reached end-of-life in 2025, a 33% increase over 2023, driven by semiconductor manufacturers reallocating capacity to AI infrastructure rather than supply chain shocks.
- Micron ended DDR4 and LPDDR4 production and exited its consumer business entirely to focus on advanced memory for AI data centers, while TSMC announced it would streamline mature-node production including 6-inch and 8-inch wafers.
- High-bandwidth memory requires three times the wafer area of conventional DRAM, forcing manufacturers to cut older, lower-margin product lines when AI demand consumes production capacity.
- Defense, aerospace, automotive, and medical device manufacturers face acute risk because their products remain in service for decades while component lifecycles shrink, and redesigns require lengthy requalification in regulated industries.
- Lead-time increases now serve as early warnings of impending obsolescence, and teams must qualify alternate parts during design reviews rather than waiting for EOL notices to arrive.
More than 600,000 electronic components were discontinued or declared obsolete in 2025, representing a 33 percent increase over 2023 and marking the second-highest obsolescence year on record, according to lifecycle data from supply chain risk management platform Z2Data. Unlike the 2022 spike—which stemmed from pandemic-era supply chain disruptions—the current wave reflects a deliberate strategic shift across the semiconductor industry to prioritize artificial intelligence infrastructure.
Why it matters
Component obsolescence has historically followed technology maturity curves: parts were retired after newer alternatives captured demand. Today, healthy parts with strong installed bases are receiving end-of-life notices simply because their production lines are more valuable when reallocated to AI. This structural change forces procurement and engineering teams in aerospace, automotive, medical devices, and industrial automation to abandon legacy lifecycle assumptions and adopt proactive risk management—or face unplanned redesigns, extended lead times, and margin erosion.
Memory manufacturers lead the capacity reallocation
Memory provides the clearest example of AI-driven obsolescence. Micron issued end-of-life notices for DDR4 and LPDDR4 in June 2025, with DDR4 shipments expected to cease within two to three months. The company later announced it would exit its consumer business entirely to focus on advanced memory for AI data centers, ending the 29-year-old Crucial brand.
High-bandwidth memory requires roughly three times the wafer area to produce the same capacity as conventional DRAM, along with complex back-end manufacturing steps including through-silicon vias. When AI demand consumes that much production capacity, manufacturers must decide what to stop making—and older, lower-margin lines are being cut first.
Samsung delayed its planned DDR4 obsolescence only after signing a long-term non-cancellable contract with a major server client. Lifecycle extensions still occur, but they go to the largest buyers with the strongest commitments. Mid-sized OEMs and industrial manufacturers typically lack that leverage.
Mature-node pressure spreads beyond memory
TSMC announced in early 2026 that it would streamline mature-node production, including 6-inch and 8-inch wafers, to free capacity for AI products. The 28 to 90 nm processes used for microcontrollers, analog ICs, power devices, and passives face similar strain as foundries prioritize their most profitable capacity.
AI data centers don't rely solely on GPUs and HBM—they also consume large volumes of conventional components. Power management ICs and MOSFETs are seeing AI-driven demand that makes them harder to source. A medical device manufacturer buying a common voltage regulator may now compete with hyperscalers for the same allocation.
In December 2025, IDC described the memory shortage as "a potentially permanent, strategic reallocation of the world's silicon wafer capacity," signaling analysts view this as a structural rather than cyclical change.
Defense and regulated industries face acute risk
Consumer electronics companies can often design around discontinued chips in their next product cycle. Organizations whose products remain in service for decades lack that flexibility. Defense accounts for less than two percent of the U.S. semiconductor market, giving it minimal leverage over chipmakers, yet an aircraft like the F-35 is expected to stay in service for 20 to 40 years—creating a stark mismatch between product lifecycle and component availability.
Automotive, industrial automation, and medical device manufacturers face similar challenges. Redesigns require requalification, retesting, and in regulated industries, recertification—processes that can take longer than the remaining life of the part being replaced.
Adapting to accelerated obsolescence
Lead-time changes now serve as early warnings. A part whose lead time jumps from eight weeks to twenty-six may be getting pulled toward high-priority markets or quietly wound down ahead of discontinuation. Teams that treat these signals as actionable intelligence gain time to evaluate redesign, last-time-buy, or extended end-of-life production options.
Lifecycle risk must move into design reviews. Before approving a part for new designs, engineers should assess whether it shares a process or fab with products the manufacturer is prioritizing for AI, and whether it sells into multiple end markets rather than depending on a single segment.
Qualifying alternates before they're needed transforms EOL notices from crises into routine events. Teams with pre-qualified alternates for high-risk parts avoid redesigning under tight turnarounds with revenue at stake.
These details were first reported by Z2Data in an analysis of 2025 component obsolescence trends.