Intel shares dropped more than 20% in extended trading on Thursday after a weak Intel revenue forecast for the fourth quarter trailed Wall Street targets. The Santa Clara chipmaker blamed an intensifying slump across the personal computer market and supply chain snags in its data centre operations for the reduced projections.
Key points
- Intel projected fourth-quarter revenue of $15 billion to $16 billion, missing Wall Street expectations of $17.1 billion.
- Adjusted earnings per share are forecast at 25 cents, compared with analyst estimates of 43 cents.
- Personal computer chip revenue fell 5% in the third quarter as memory component prices surged.
- Data centre sales grew 78% to $7.3 billion, but raw material shortages and an Israeli fab closure capped shipments.
Why the outlook missed expectations
Intel said on Thursday that revenue in the final quarter of 2026 will sit between $15 billion and $16 billion. Analysts polled by LSEG had expected average revenue of $17.1 billion, with several earlier forecasts topping $18 billion, Bloomberg reported. The company forecast adjusted net profit of 25 cents a share, well below the 43 cents predicted by financial analysts.
The projected decline follows a period of heavy enthusiasm for the company. Intel shares had gained roughly 150% earlier in the year before retreating 23% during October. That pullback came amid wider tech volatility, visible when Alphabet shares slumped after heavy AI spending raised fresh questions among investors about the near-term returns on computing infrastructure.
The weakness in the projection contrasts with Intel’s actual performance in the third quarter, which exceeded top-line expectations. Third-quarter revenue rose 21% from a year earlier to $16.51 billion, ahead of the $16.5 billion consensus. Adjusted earnings per share reached 52 cents, beating the 40 cents estimated by analysts.
Memory shortages hit personal computers
The core problem for Intel sits in its client computing and physical AI division, which manufactures processors for laptops and desktop computers. Division sales fell 5% from a year earlier to $8.1 billion during the third quarter, falling short of internal targets.
A sharp rise in memory component prices has forced computer makers to trim production plans. According to Counterpoint Research data, global computer shipments fell 4% in the second quarter of 2026 as component expenses climbed. Research firm TrendForce reported in September that device assemblers had reached the limit of what they could pay for dynamic random-access memory and flash storage.
Lip-Bu Tan, who took over as Intel’s chief executive to lead a corporate restructuring, told Bloomberg that memory pricing pressures would not ease quickly. “Memory pricing went up a lot. I think my prediction is that in 2027 it is going to be a very severe issue to all of us,” Tan said.
Intel now expects the broader personal computer market to contract by 12% across 2026, followed by a further 10% contraction in 2027. David Zinsner, Intel’s chief financial officer, said in a company release that the slowdown would persist through at least the first six months of next year.
Data centre growth faces supply bottlenecks
While consumer computing slowed, demand for enterprise infrastructure expanded rapidly. Intel reported that revenue in its data centre and artificial intelligence unit climbed 78% year-on-year to $7.3 billion during the third quarter.
That momentum was held back by operational constraints. Tan confirmed that the data centre division was unable to meet total customer orders due to shortages of specialised raw materials and an unscheduled factory shutdown in Israel. Without those disruptions, server processor sales would have been significantly higher.
Demand for Intel’s central processing units has grown as cloud providers construct automated software agents. While training massive foundation models requires graphics processors built by Nvidia and Advanced Micro Devices, autonomous agent architectures depend on central processing units to run operating systems, coordinate external tools and manage database requests.
Intel secured an expanded agreement with Oracle this week. Oracle announced at its annual conference that it will deploy Intel’s latest Xeon processors alongside accelerated clusters from Nvidia and AMD to expand its cloud network.
Foundry turnaround and public backing
The production delays arrive as Intel attempts to establish an independent manufacturing business to rival Taiwan Semiconductor Manufacturing Company. Intel is currently the only commercial fabricator in the United States producing advanced logic chips at scale.
Intel has started high-volume manufacturing of its 18A process node at its fabrication site in Arizona. While Intel remains the primary buyer of its own 18A production line, industry analysts expect outside chip design companies to consider its next node, known as 14A. Dan Hutcheson, an analyst at TechInsights, said major fabless semiconductor companies have been evaluating the factory line to introduce pricing competition against TSMC.
The turnaround effort has received direct financial backing from governments and private investors. The United States government acquired a 10% equity holding in Intel in 2025 under the Trump administration after earlier manufacturing delays on older process nodes. Nvidia and SoftBank also purchased multi-billion-dollar equity stakes last year.
Tan has slowed the pace of capital spending set by his predecessor, Pat Gelsinger, who started the foundry expansion. Intel raised $23 billion through a public equity offering in August to balance its balance sheet while construction continues. Last week, workers fitted the first structural steel truss at Intel’s upcoming factory in Ohio, which was originally scheduled to begin commercial operations in 2025. A second advanced manufacturing plant in Arizona is scheduled to start chip fabrication in late 2028.
