AI news ·
Microchip says AI buyers take over supplier factories and push everyone out
Microchip Technology beat its forecast with $1.622 billion in revenue but shares plunged 22.55% after AI buyers seized its suppliers' factory capacity, blocking order fulfillment.
Microchip Technology shipped $1.622 billion in the September quarter, beating its own forecast. The stock still dropped 22.55% over two days after CEO Steve Sanghi told investors that AI buyers had seized factory capacity at its packaging and test suppliers, blocking the company from filling all its orders.
Sanghi described the situation bluntly: AI companies "came in and took over the whole factory. They pushed everybody out." The shortage sits in outside assembly and test houses, substrate supply, and trailing-edge manufacturing lines. Sanghi said moving that work to other suppliers will take about six months.
Shares closed at $58.35 on November 6, 2026, down from $75.34 two sessions earlier. The selloff began with a sector-wide chip decline on November 5, before Microchip's report, and accelerated after the details emerged.
One-time gains mask the margin picture
CFO Eric Bjornholt disclosed roughly $53 million in non-recurring revenue during the quarter, all at 100% gross margin. The items included a $21 million distributor inventory price adjustment, $22 million in extra patent licensing, and $10 million from a customer contract. Those gains helped push non-GAAP gross margin to 66.5%.
GAAP earnings per share came to $0.68, including a 13-cent tax benefit. Non-GAAP EPS reached $0.97. The December guidance calls for non-GAAP gross margin between 64% and 65%, reflecting the absence of those one-time boosts.
Lead times stretch, backlog builds
Standard products that once shipped in four to eight weeks now take six to 32 weeks, Sanghi said. Orders placed inside those extended lead times often go unfilled, pushing revenue into later quarters. The book-to-bill ratio stayed above one, and Microchip entered the December quarter with more backlog than it had entering September.
Some high-margin orders that slipped out of the September quarter are now scheduled for March. Sanghi noted that March-quarter backlog is already stronger than December's was at the same point in the cycle.
Data center revenue surges despite constraints
Microchip's data center business topped $250 million in the quarter, up 55.7% from a year earlier. The segment is on pace to reach $1 billion in calendar 2026. Sanghi said a slowdown in AI demand would actually help in the near term, freeing capacity the company needs for other customers.
Supplier costs are also climbing. Major foundries and OSATs plan price increases around January. Microchip intends to absorb those costs initially but may pass them through to customers if the trend continues.
Analyst targets fall but stay above the stock
Truist cut its price target to $62 from $90. Raymond James moved to $80 from $125, Citi to $75 from $95, and JPMorgan to $100 from $120. Every revised target remains above Friday's closing price.
The board declared a quarterly dividend of 45.5 cents per share, payable December 8 to shareholders of record on November 23.
Why this matters for operations and supply chain leaders
Microchip's situation is a signal for any operations team that relies on outsourced semiconductor packaging and test capacity. When AI buyers with deeper order books can monopolize OSAT lines, smaller-volume customers get pushed into longer lead times regardless of their relationship with the chip designer. The six-month timeline Sanghi cited for qualifying alternative suppliers is a real constraint - production planners mapping Q2 and Q3 2027 capacity should factor that lag into their own lead-time assumptions. AI Production Forecasting Courses can help teams model these kinds of supply shocks before they hit the P&L.
For supplier relationship managers, the dynamic creates a hard conversation: foundries and OSATs are signaling price increases, and absorbing them erodes margin. The alternative is renegotiating contracts or qualifying new partners, both of which take time. AI Vendor Management Courses offer frameworks for evaluating supplier risk when capacity gets concentrated among a few deep-pocketed buyers.