Samsung will begin construction of its second chip fabrication plant in Taylor, Texas, next month, the company confirmed, with the expansion timed to serve accelerating demand for AI processors.

The date matters for the foundry market. Samsung's contract chipmaking arm (the business that manufactures chips designed by other companies) has been racing to add US capacity, and a second Taylor plant doubles down on a site the company has already committed to.

Taylor, a city of roughly 16,000 people northeast of Austin, hosts Samsung's existing US semiconductor investment. The second plant adds capacity at a moment when AI accelerators — the high-end processors that train and run large models — are absorbing a growing share of global advanced-node output.

What does the timing tell us?

Three things stand out:

• Speed. A construction start "next month" means site work moves from planning to execution on a near-term clock, not a vague multi-year horizon.
• Demand signal. Samsung explicitly tied the decision to AI chip demand, joining a pattern of foundries and memory makers expanding capacity around AI workloads.
• US footprint. The Taylor expansion deepens Samsung's position in American semiconductor manufacturing, where rivals including TSMC are also building.

AI chips need advanced packaging and leading-edge process nodes (the manufacturing recipes that determine how small and dense a chip's transistors are). Any new fab dedicated to that demand adds supply in a segment where customers have complained about tight capacity for over two years.

Why Taylor, and why now?

Samsung picked Taylor in 2021 for its first US fabrication site in years, drawn by state incentives and proximity to its Austin operations. A second plant on the same campus gives the company shared infrastructure, an existing workforce pipeline, and room to scale without repeating a site-selection process.

The AI demand backdrop explains the urgency. Datacenter operators keep ordering accelerator chips in volume, and foundries that can make those chips — or the surrounding components — are booking out capacity well in advance. Samsung has pushed to win AI-related foundry customers as it competes with TSMC, the market leader with roughly two-thirds of global foundry revenue.

What does this change for buyers and rivals?

For chip designers, a second Taylor plant means one more source of US-based capacity coming online in the coming years. Fabs take time — construction is only the first phase, followed by tool installation (fitting the lithography and etch machines that do the actual chipmaking) and qualification runs before volume production. So the near-term supply picture doesn't shift tomorrow.

For competitors, the message is that Samsung intends to fight for AI-era foundry business on US soil. TSMC is expanding in Arizona; Intel is building in Ohio and Arizona; Samsung is doubling down in Texas. The three-way race for American advanced manufacturing now has another committed entrant with a shovel-ready date.

For Taylor itself, a second plant means extended construction employment now and a larger permanent workforce later, on top of what the first fab already brought to the region.

The road ahead

Next month's construction start is a firm, dated commitment in an industry where announcements often outrun timelines. Watch for follow-up details on the plant's process technology target, capacity figures, and production schedule — the numbers that will determine whether this fab serves leading-edge AI silicon or supporting nodes.

Samsung hasn't publicly detailed those specifications for the second plant yet. Until it does, the confirmed facts are simple: construction starts next month, the site is Taylor, and the driver is AI chip demand.

That's enough to move the story. The rest of the details will decide how much.