Chris Caso

Steve said December should be better than the usual down 3-5%, but declined to define new long-term seasonality.

Steve reaffirmed debt paydown as the priority, with no buyback or dividend increase planned for the foreseeable future.

Steve said Microchip has not raised prices, prioritizing repaired relationships, and will adjust selectively only where input costs rose, especially for non-partners.

Steve said both distribution and OEM inventory is broadly corrected, with thousands of customers buying at run rate, though some SKUs remain high.

Eric accepted mid-70s incremental fall-through as a modeling approach but warned it will be lumpy. Steve emphasized that high-margin external foundry products (FPGA, data center, networking) growing faster than internal products means mix improvement will supplement utilization gains.

Steve confirmed debt reduction prioritized over buybacks for an extended period. Net leverage at 4.18x versus original 1.5x target. Dividend to remain flat. December quarter was first in a long time where adjusted free cash flow exceeded dividend payments.

Steve acknowledged softer tone versus 90 days ago when he expected low single-digit December growth. Attributed to overall business environment softness and tariff impact on customer psyche, plus customers scheduling bookings for March to dress up year-end balance sheets.

Steve said current write-offs are driven by 2-year-old product with shifted mix, not excess production. Expects meaningful improvement starting March quarter when YoY sales growth turns positive, changing the 18-month trailing sales calculation that drives write-off thresholds.

Steve noted direct customers likely have higher remaining inventory than distributors since they were prioritized during the upcycle.

Steve acknowledged undershipping versus end demand but said with 110,000+ customers and broad SKU mix, precise quantification is not possible. Anecdotally, many line items have run rate but customers are not buying due to remaining inventory.

Steve detailed minimal direct tariff exposure (<4% made in China), proactive production shifts to Taiwan, and a worst-case scenario showing inventory still declining.

Eric said underutilization charges (~$54M) won't change meaningfully this quarter but inventory reserves will drop dramatically as inventory declines by $350M+ in FY26.

Steve firmly committed to maintaining the dividend, noting the cash shortfall is temporary and driven by semi-annual bond payment timing.

Steve clarified Microchip uses 300mm at foundries and argued total cost of ownership is comparable to internal 300mm. Noted China-for-China strategy exists, details on March 3.

Ganesh attributed European weakness to high industrial/auto exposure and noted Greater China was relatively stable without great strength.

Ganesh clarified the temporary borrowing was quarter-to-quarter timing, not a structural debt increase, with intent to repay in stronger quarters.

Answered with a concrete counterpoint that the non-enterprise-data businesses grew double digits, alongside Michael Hsing's picks and shovels framing.

Confirmed CPU as a continuing growth driver and that MPS is broadly indexed across x86 and ARM, with Michael Hsing indicating share is comfortably past 30%.

Tony confirmed H2 auto ramp expectation based on previously won designs coming to market. Pipeline expanding. Michael noted continued socket wins and market share expansion.

Management declined to parse merchant vs ASIC or volume vs content. Tony emphasized the floor raise reflects extended backlog visibility, not a fundamental shift in growth approach.

Tony highlighted strong Q4 ordering with book-to-bill well above one, backlog extending to Q2 2026. Enterprise data, communications, and automotive trends all expected to continue.

Bernie raised enterprise data 2026 growth floor from 30-40% to 50%+. Michael indicated even higher expectations, citing multi-customer design wins across the board.

Bernie gave the most specific enterprise data 2026 outlook yet: 30-40% growth, mostly H2 weighted. Michael was characteristically more bullish but declined to pin a number.

Michael cited module growth and MPS's power density advantage. Bernie's 30-40% estimate prompted an unusually candid reaction from Michael questioning the specificity.

Bernie confirmed Q4 will be up sequentially after Q3's 20-30% growth, while declining to formally narrow the full-year range due to short lead times.

Bernie maintained cautiously optimistic posture with broad-based strong demand. Short-term ordering patterns noted as atypical versus past recoveries.

Bernie cited accumulating positive data points while maintaining a range-of-outcomes framework. Michael noted proximity to H2 gives naturally better visibility.

Bernie noted a broadening customer base for AI power beyond one main customer. Michael indicated dollar content potential of $500-600 million per half year.

Tony pushed back on quarterly precision, noting multiple business factors beyond any single driver make quarter-by-quarter forecasting impractical.

Bernie confirmed the flattish outlook covers the full enterprise data segment and highlighted the triple-down effect in memory, optical, and networking.

Pointed back to Investor Day guidance and named the drivers (Alphawave, Modular, data center) but gave no quantified growth rate for next year.

Explained the royalty cap mechanism directly, making clear the pricing benefit only accrues below the ASP cap.

Akash indicated premium tier holding while lower tiers weaken, noting this is a reasonable way to think about forward trajectory.

Cristiano confirmed it is a custom product for a hyperscaler but did not specify whether it is a CPU or accelerator.

Cristiano confirmed this is a fundamentally different approach: purpose-built for inference with highest compute density at lowest cost and energy, beyond traditional GPU/HBM architecture.

Akash cited durable multi-year content growth driven by OEM competition and consumer demand for more capable devices, plus mix shift to premium tier.

Akash downplayed timing concerns, emphasizing strength across Android, auto, and IoT. December remains seasonally strongest quarter regardless of lower Apple share.

Akash emphasized disciplined OpEx management with reallocation strategy. New hiring focused on required skills for data center. No major OpEx step-up implied.

Akash attributed growth to premium tier mix expansion from 21% to 30%, customers gaining share, and China subsidies positively impacting premium mix.

Akash provided design and app metrics but did not directly address margin profile of lower-priced PC platforms. Cristiano noted 12% long-term share target supports confidence.

Akash cited consumer demand for more capable phones as the primary ASP driver, with TSMC 3/4nm cost increases kicking in January to be reflected in ASPs over time.

Akash confirmed Samsung higher share is reflected across all quarters and does not fundamentally change the seasonal Q1-high/Q3-low pattern.

Palkhiwala said the growth is premium-tier driven, with OEMs accelerating launches on the new flagship chipset, rather than a mid-tier broadening.

Explained the pricing philosophy and said the annual update will likely be better than the low single digit decline guided, but did not quantify.

Walked through content, system complexity, connectivity, security and functional safety as the layered value capture mechanism.

Bill explained VSMC provides more controlled supply. For external capacity beyond existing agreements, additional charges apply as foundry capacity tightens.

Rafael explained conviction comes from traction in accelerated growth drivers, data center ramps, and industrial & IoT momentum. Views 2027 as a milestone, not a destination.

Rafael explained H1 2025 inventory digestion paused some accelerated drivers (radar, electrification, SDV new model ramps). H2 acceleration restored thesis, and 2026 expected within or above model.

Bill outlined H1 OpEx above 23% model from seasonal benefits reset, Q2 merits/promotions, and annual IP license. H2 below 23%. Full year at or below 23% target. R&D kept at 16% for SDV and physical AI capture.

Rafael confirmed Tier 1 inventory is way below NXP's manufacturing cycle time, an unsustainably low level. Bill added NXP undershipped for 8 quarters and the headwind is now subsiding.

Bill confirmed no change to capital allocation. Last 12 months returned 106% of non-GAAP free cash flow to shareholders. Buybacks restarted after cash position secured for investments.

Kurt confirmed 30% of China revenue already sourced from China wafer manufacturing, with significant additional volume from non-US foundries. Strategy predates tariff concerns by 2+ years.

Bill confirmed restructuring ahead of plan to absorb 1,250 employees from three pending acquisitions. Confident in hitting 23% OpEx model in H2 2025.

Kurt confirmed the 30% digital networking portion of comms infra will continue to decline materially over the next several quarters, partially offset by RFID growth.

Kurt firmly denied any China inventory build-up, citing structural growth from Chinese OEM market share gains and 50% EV penetration. Changed org structure with dedicated China business leader.

Haviv explained analog is in Phase III with real-time start modulation. Inventory days at 209, expected to drift toward lower end of 150-250 range during upturn. AT capacity investments ongoing.

Confirmed TI follows market pricing, wants to see demand sustainability before acting. 2025 price agreements made in weaker demand environment being revisited as demand strengthens.

Haviv noted Q1 China auto typically dips for Chinese New Year but backlog is pretty even across geographies. Mike said Q4 China was up ~16% YoY and down ~7% sequentially, in line with overall top-level results.

Rafael said no significant change in loadings vs Q4 and declined to disclose direction. Mike added that TI has the manufacturing flexibility to adjust as conditions evolve.

Rafael explained the mechanical logic: lower Q4 revenue at same loadings would grow inventory, so loadings must moderate to keep inventory flat to down. Recovery pace determines when loadings increase.

Haviv attributed the slower pace to customer hesitancy on CapEx due to tariff uncertainty, particularly in industrial. Called out data center as the sole area with no investment slowdown.

Rafael described the ideal model of stable loadings that drain inventory in upturns and build in downturns, acknowledging guardrails needed given cycle uncertainty.

Haviv said auto customers are cautious with tariffs, running real-time on consignment with no inventory replenishment. Noted industrial's 7-quarter peak-to-recovery suggests auto could start H2.

Management explained both the cyclical timing difference (Embedded peaked ~1 year after Analog) and LFAB underutilization impact. Clear and complete.

Haviv provided detail on China auto EV content growth, PE recovery from prior supply shortfalls, and noted industrial hasn't yet started cyclical growth in China.

Roger cited strong intrinsic Chinese demand for mainstream logic and memory but gave no specific share for next year.

Christophe said only tactical timing shifts remain and ASML checks site readiness before allocating tools; Roger flagged customer construction execution to monitor.

Roger described freeing up cabins, reducing cycle time, and leveraging prior long-lead-time supply investments, with the 30% plan as the timing proxy.

Roger declined to give a 2026 China projection but said mainstream logic and memory demand in China remains healthy and shows no signs of dropping off a cliff.

Christophe said Memory 2025 revenue is roughly flat vs 2024 with HBM driving continued strength, while Logic has stepped up after a lower 2024. Roger noted Q2 memory bookings were only 16% but Q1 was 40%, classic lumpiness.

Roger agreed the geographic-dispersion thesis should drive incremental capacity over time but called it a story of puts and takes given the near-term tariff uncertainty.

Roger refused to put a number on full-year tariff impact citing ongoing US review and value-chain absorption uncertainty, and said ASML should not bear the burden, it should fall on the next element in the value chain.

Roger characterized 2026 as a potential growth year but declined to give any magnitude or direction, citing early visibility.

Christophe pointed to two sources of downside embedded in the low end: residual customer pushout risk (much smaller than 2024 but not excluded) and newly released export-control rules whose impact is hard to quantify.