Written by: Rita

Trend Navigation Guide

On July 1st, Morgan Stanley released a semiconductor inventory tracking report that clarified the current state of the North American semiconductor supply chain: inventory in the first quarter of 2026 increased by only 9 days month-on-month, whereas normally it should increase by 19 days. Distributors are actively reducing inventory, chip manufacturers are restraining production expansion, and overall customer demand is stable, which indeed tightens supply.

However, a widespread inventory replenishment has not yet occurred. The total inventory across the whole chain is still 33 days higher than the historical median. Since the peak inventory in 2023, the industry has spent nearly two years reducing inventory from over 100 days, but it has struggled to compress these final 33 days.

The core judgment of the report points to a diverging situation. The industry as a whole has not yet reached the time for inventory replenishment, but the supply-demand relationship in the three segments—analog chips, computing networks, and memory chips—has already tightened ahead of schedule. Morgan Stanley has provided a clear list of targets: ADI, NXP, NVDA, AVGO, CRDO, MU, SNDK, plus equipment companies MKSI, KLAC, LRCX, ONTO.

Inventory is only 10 days short of normal, but the direction is right

In the first quarter, the total inventory in the supply chain increased by 9 days month-on-month, whereas the historical average increase during the same period is 19 days. The reduction of 10 days is the result of cooperation among the three segments. The direction is correct, but the entire chain still exceeds the historical median by 33 days, equivalent to over a month's excess inventory in the entire industry.

Reducing inventory from over 100 days to 33 days took nearly two years. The final 33 days represent the boundary of pricing power; having an extra month of inventory gives customers an additional month to pressure prices. The initial inventory reduction was driven by industry consensus, with everyone synchronously cutting orders, reducing production, and controlling shipments. The last 33 days require demand cooperation for digestion, but demand has not yet improved.

Distributors are proactively cutting inventory against the trend

Among the three segments, the only one that recorded a month-on-month decline in Days of Inventory (DOI) in this traditional inventory accumulation season in the first quarter is distributors.

Distributors' DOI dropped to 61 days, a decrease of 2 days, while the historical average for the same period should have increased by 4 days. WPG's inventory increased by 18%, with costs rising by 21%; Avnet's inventory increased by 3%, with costs up by 13%. Revenue growth rate exceeded inventory growth rate, indicating that goods are moving normally. All three distributors have absolute inventory amounts rising, but DOI is trending down, which means business is being conducted and goods are flowing, rather than sitting idle in warehouses. Distributors are the healthiest segment in the entire supply chain.

Chip manufacturers have already divided into two worlds

Chip manufacturers' DOI increased by 2 days to 114 days, whereas the historical average for the same period increased by 5 days. On the surface, this looks below seasonal trends, but when broken down, two opposing forces can be seen.

Smartphones increased by 21 days, and foundries increased by 8 days, indicating manufacturers in the AI computing supply chain are actively stocking up, with demand expectations still high. On the other hand, analog chips decreased by 2 days, and semiconductor equipment decreased by 6 days, suggesting that these two fields have already cleared up supply.

Within the same industry, some are stocking up for growth while others are contracting supply for pricing power. These two narratives together accurately depict the true state of chip manufacturers.

The demand signals from customers have not yet materialized

Customers' DOI increased by 9 days, while the historical average for the same period increased by 8 days, appearing to be relatively stable. However, telecom equipment manufacturers surged by 17 days, ODMs increased by 16 days, and automotive suppliers decreased by 1 day.

The telecom segment is 23 days above the historical median, representing the most dangerous sub-sector in the customer chain. ODMs are 21 days above average, also in a high position. Conversely, automotive suppliers have returned to near historical medians.

In the same industrial chain, some are desperately stocking up while others are purchasing based on demand. The absence of a unified demand signal indicates that the terminal market has not yet reached a consensus.

Three segments can thrive without waiting for inventory replenishment

Morgan Stanley's stock selection framework is clear: seek segments where supply-demand relationships have tightened, without waiting for demand to fully recover.

The first segment is analog chips. ADI and NXP are reducing inventory, and pricing power is returning. In the analog/MCU sub-sector, DOI decreased by 2 days, one of the few segments in the semiconductor industry where inventory is decreasing. Analog chip products have long lifecycles, diverse customers, and strong price rigidity. When supply tightens, the profit elasticity is greater than that of other digital chips.

The second segment is computing networks. NVDA, AVGO, and CRDO see AI demand digesting inventory. Although DOI in the computing/mobile sub-sector is still 12 days above the historical median, structural growth in AI computing is consuming this inventory. Morgan Stanley believes this demand is not cyclical; even if overall industry demand weakens, the shipment volume of AI-related chips will not decline proportionally.

The third segment is memory chips. MU and SNDK are the only sub-sectors where year-on-year inventory has not expanded, month-on-month inventory remains stable, and absolute amounts have decreased by only 1% year-on-year. Memory is the most volatile category in semiconductors. While other segments are still digesting inventory, memory inventory is no longer increasing; this is a signal of a cyclical bottom.

Equipment manufacturers are also highlighted. MKSI, KLAC, LRCX, ONTO saw DOI decrease by 6 days, outperforming peers in reducing inventory. Equipment is a leading indicator of capacity expansion, and Morgan Stanley highlighting this segment indicates that improvements in supply have already transmitted upstream.

No small companies in Morgan Stanley's recommendation list

One detail in this list is more important than all the data: there are no small to medium-sized semiconductor companies.

Morgan Stanley says a replenishment has not yet come, yet they promote stocks that would benefit from it. The only explanation is that these companies do not need to wait for inventory replenishment. Analog chips and memory have already cleared supply, and structural demand supports AI computing. The speed of inventory digestion, financing capacity, and customer negotiation positions of small to medium-sized companies cannot support a turnaround in this cycle.

Previously, once inventory hit bottom, smaller companies had the greatest elasticity. However, now the cost of capital, customer concentration, and technological barriers have changed. Morgan Stanley's silence itself is a form of judgment.

Trend perspective

The most poignant question in this report lies within the recommendation list. Morgan Stanley bets that the supply side can complete its own clearance without needing demand assistance. This logic has already worked for analog chips and memory, but when applied to the AI computing chain, inventory data does not support it.

Computing/mobile DOI is 12 days above the historical median, and telecom is 23 days above. The narrative of AI digesting inventory is still a logical deduction that has not been validated by inventory data. If AI capex marginally slows in the second half of the year, inventory risks along this chain will be repriced. Morgan Stanley strongly recommends NVDA and AVGO, but the data supporting this recommendation is significantly weaker than for the other two segments.

Another issue the report avoids is the position of small to medium-sized companies in this cycle. Morgan Stanley's lack of promotion indicates a lack of confidence, suggesting that they believe the structure of this cycle has changed. If this judgment is correct, semiconductor investment may enter an era of "large-cap racing", necessitating a rewrite of the cyclical elasticity logic for smaller stocks.

Disclaimer

This article is a summary and interpretation of third-party broker research reports by Trend Navigation Research. The ratings, target prices, earnings forecasts, and related judgments quoted in the article are the views of the respective broker analysts and only represent their institutions' positions, not those of Trend Navigation Research, nor do they constitute any investment advice.

Markets are risky; decisions require independence. This article should not be used as the basis for buying or selling any securities.