Intel estimates Fab 28 blast will erase $600M, delay Xeon output

By Sophie Lindqvist

Intel expects the Fab 28 explosion to remove about $600 million from revenue across two quarters. Intel expects Q4 2026 revenue to fall by approximately $300 million. It anticipates a similar revenue reduction in Q1 2027.

The mid-October blast at Intel's Fab 28 near Kiryat Gat, Israel, killed two people, seriously injured one and forced the entire facility into a two-week shutdown. Intel expects lower shipments of server processors built on 10nm-class processes in both quarters. That will constrain Xeon supply across a roughly six-month period.

What does the blast disrupt?

Fab 28 is one of Intel's key plants for 10nm-class and Intel 7 production. Intel 7 was formerly called 10nm Enhanced SuperFin.

Because Fab 28 combines both process categories, the outage reaches current server I/O and older products, while its precise SKU-level effects cannot be separated from company-wide revenue and margin estimates.

Xeon 6 illustrates the manufacturing split. Its compute dies use Intel 3. Its I/O dies use Intel 7. Intel continues to sell Xeon 5 Emerald Rapids and Xeon 4 Sapphire Rapids, both on Intel 7. It stopped accepting orders for Sapphire Rapids in late September.

CFO Dave Zinsner said on the earnings call: “In Israel, we have been shut down for about two weeks. We feel like we are going to be starting up production here by the end of this week.” Intel expected Fab 28 to restart this Friday. The company has not said when the plant will regain full capacity.

How large is the revenue and margin hit?

Intel's $600 million estimate covers both quarters combined. It does not represent a second $600 million loss.

Zinsner said: “Please note almost all of the $300 impact from Fab 28 is related to 10nm server products.”

Intel separately expects the incident to cut Q4 2026 non-GAAP gross margin by approximately 2%.

Intel is still assessing work-in-progress inventory. The company must determine which wafers it can recover and which require write-offs. Intel expects most write-offs in Q4 2026. Shipment reductions then extend into Q1 2027. The revenue effect spans two quarters, while most damaged-wafer write-offs occur in Q4, concentrating the reported margin pressure in the first period of the disruption.

Why can Intel not simply move production?

Intel has several other 10nm-capable fabs. They include Fab 12, Fab 22, Fab 32 and Fab 42 in Arizona. Fab 28 is also among Intel's 10nm-capable plants.

The report says strong server demand prevents Intel from fully replacing the missing output. It assesses that Intel has already exhausted its practical options for now.

The report argues that each product requires a qualified manufacturing flow, so moving output can require requalification, added time and cost that Intel may not justify while Fab 28 restarts.

Other fabs may lack product-specific manufacturing qualifications. The report also says Intel will not repurpose Xeon 6 I/O-die lines to make previous-generation CPUs. It adds that Fab 28's pending restart makes a costly transfer less compelling.

What remains undisclosed?

Intel has not named the affected Xeon processor families. It has not disclosed the number of wafers damaged. Nor has it specified when Fab 28 will reach full capacity after the restart. Those disclosures therefore do not map the estimate to precise processor shipments.

The uncertainty extends to client CPUs. Intel has not disclosed whether Intel 7 client shipments will be disrupted. The node supports 12th, 13th and 14th Generation Core processors, including Raptor Lake. Because Intel has not said whether Fab 28 makes any client parts, and other Intel 7 fabs might not absorb the missing production, consumer availability remains unquantified.

The claim that this is the most devastating semiconductor-fab disruption in recent history lacks a comparative outage record. The source's reference to the largest memory shortage likewise lacks a memory-output or pricing series. Neither superlative is established by the available figures.

Fab 28's Friday restart is the next measurable capacity test. Even a successful restart cannot erase the lost production time, and Intel still expects lower Xeon shipments in Q1 2027 while any client-processor effect remains unquantified.