AustinIO > Texas Nexus > Texas Triangle Cluster > Brazos Valley Semiconductor Concentration > Terafab Production Facility (Grimes County)

Terafab Production Facility Grimes County

SpaceX-Operated Semiconductor Production Fab at Gibbons Creek Reservoir

The Terafab Production Facility is the production-scale element of the broader Terafab program — the volume manufacturing complement to the Tesla-led Research Fab at Giga Austin North Campus. SpaceX (Space Exploration Technologies Corp.) filed a Grimes County, Texas tax abatement application on May 6, 2026 designating the production facility's location at "SpaceX Reinvestment Zone No. 1 – 2026-001" at the Gibbons Creek Reservoir and surrounding areas off Highway 30, approximately 20 miles east of Bryan-College Station, 70 miles northwest of Houston, and 90 miles northeast of Austin. The filing characterizes the project as a "multi-phase, next-generation, vertically integrated semiconductor manufacturing and advanced computing fabrication facility" representing "a transformative investment in domestic semiconductor manufacturing capacity." Initial-phase capital commitment of $55 billion plus up to $119 billion total if all phases are built positions Terafab Grimes County as one of the largest single US semiconductor capital commitments on record. The Grimes County Commissioners Court will hold a formal public hearing June 3, 2026 at 9:00 AM at the Grimes County Justice & Business Center (270 FM 149 W, Anderson, Texas) to consider approving the property tax abatement agreement.

What distinguishes Terafab Grimes County at the Texas Nexus level is the operational separation of production manufacturing from research and development: SpaceX handles high-volume chip manufacturing at the Grimes County production fab while Tesla operates a smaller R&D pilot line at Giga Texas Austin (per Musk's Tesla Q1 2026 earnings call). The split is structural rather than incidental — the Research Fab on Giga Austin's North Campus iterates on process recipes and chip-and-fab co-design experiments at experimental scale (a few thousand wafer starts per month); the production fab at Grimes County executes high-volume manufacturing of validated designs at the 100,000-wafer-starts-per-month prototype scale and the 1-million-wafer-starts-per-month full-scale ambition Musk described at the original Terafab announcement event. Combined with the announced Intel 14A process technology (per Musk's Tesla Q1 2026 earnings call) plus the broader Terafab program's vertically integrated chip design, lithography, fabrication, memory production, advanced packaging, and testing scope, Terafab Grimes County would represent the most consequential US semiconductor investment outside the established Arizona-Texas-New York corridor when fully built.

Site and Substrate

The Gibbons Creek Reservoir site occupies approximately 6,000 acres of brownfield substrate in rural Grimes County, inherited from the former Texas Municipal Power Agency (TMPA) coal-fired power plant that operated 1982-2018. TMPA closed the plant 2018; Charah Solutions acquired the approximately 6,000-acre property in 2021 for demolition, remediation, and parcel-by-parcel resale. Local resident reports indicate Musk-affiliated entities completed property acquisition approximately one week before the May 2026 county filing, with surveyors documenting work across approximately 1,900 acres in the days preceding the public notice.

The brownfield substrate provides structural advantages relative to greenfield semiconductor fab sites that no other major US semiconductor project has been able to leverage at comparable scale. Existing transmission infrastructure inherited from the prior TMPA coal generation provides multi-hundred-megawatt power supply capacity ahead of new construction substrate that would otherwise require ERCOT interconnection queue clearance plus new transmission line construction. Water access via the Gibbons Creek Reservoir provides ultra-pure water source capacity that semiconductor fab operations require at scale. The remediated site condition reduces site preparation timeline and environmental compliance overhead relative to greenfield substrates. Reduced rural land cost relative to metropolitan brownfield substrate provides additional economic advantages.

The site's geographic position approximately 20 miles east of Bryan-College Station along Highway 30 connects the production fab to the broader Brazos Valley Semiconductor Concentration substrate including the Texas A&M Semiconductor Institute (TSI) at RELLIS Campus, the Cyclotron Institute (radiation testing capability supporting SpaceX orbital compute), AggieFab Nanofabrication, and the Center for Microdevices and Systems. Substrate Inc.'s prospective $10-13 billion Project Factory One commercial semiconductor manufacturing facility on the 288-acre RELLIS reinvestment zone provides parallel commercial concentration supporting continued semiconductor supplier ring development across the region. The combined Brazos Valley substrate plus Highway 30 corridor connectivity supports continued workforce flow plus supplier ring development across the concentration.

Why Gibbons Creek: The Constraint-Satisfaction Logic

The Gibbons Creek site selection is not optimization across one variable — it is the simultaneous solution to a stacked set of binding constraints where no single variable dominates but each variable rules out most alternatives. Working through the binding constraints applied to the specific Terafab production fab requirements clarifies why Gibbons Creek emerged as essentially the unique US site that satisfies all constraints at the announced scale within the announced timeline. The broader constraint-satisfaction framework that applies across Texas AI-Industrial buildout is covered at Why Texas: The Structural Logic of AI-Industrial Concentration; this section applies that framework specifically to Gibbons Creek.

Power timeline. A leading-edge production fab at the announced 100K-1M wafer-starts-per-month scale draws 0.5-10 GW continuous depending on phase. ERCOT interconnection queue clearance for greenfield sites currently runs 5-7+ years from application to energization. New transmission line construction adds 3-5 additional years. Combined greenfield power timeline can lose 8-12 years to infrastructure development before construction begins. The Gibbons Creek brownfield substrate inherits multi-hundred-MW transmission infrastructure from the prior Texas Municipal Power Agency coal plant (peak ~470 MW operational 1982-2018). The transmission rights-of-way exist. The substation footprint exists. The interconnection point exists. Charah Solutions' 2021 demolition removed the generation but preserved the transmission. Time-to-power at Gibbons Creek is years faster than any greenfield Texas alternative — and substantially faster than any greenfield US alternative outside the established semiconductor manufacturing corridor.

Water rights. A 100K-WSPM fab consumes 5-10 million gallons per day of ultra-pure water plus 2-4x that volume in source water inputs. Full multi-phase buildout extrapolates to 50-100 MGD UPW. Source water at this scale within Texas concentrates near major river systems, reservoirs, or Gulf-adjacent groundwater systems — and most candidate sources compete with municipal supply, agricultural rights, and existing industrial users. The Gibbons Creek Reservoir is approximately 2,400 acres, originally constructed by TMPA specifically for cooling water at the coal plant. Water rights at Gibbons Creek are largely intact from the prior TMPA framework — the rights flow with the property to the new owner. The reservoir's existing intake infrastructure, discharge permitting, and water rights framework substantially compresses the water-supply timeline that greenfield sites face. The pre-existing water rights and infrastructure substrate is structurally distinctive even relative to other Texas brownfield sites.

Contiguous acreage at price. Multi-phase Terafab buildout at announced scale requires 3,000-6,000 acres for full vertical integration (logic plus memory plus advanced packaging plus utilities plus supplier co-location plus buffer plus future expansion). Contiguous brownfield acreage at this scale within reasonable proximity to a major metropolitan region is genuinely scarce in the United States — most US sites of comparable acreage are Western federal lands, Gulf Coast petrochemical complexes (incompatible neighboring operations), or rural agricultural land with no industrial substrate. The Gibbons Creek property is approximately 6,000 acres total (Charah Solutions' 2021 acquisition footprint), with reports indicating ~1,900 acres surveyed in the days before the May 2026 filing. Giga Austin's 2,500-acre footprint plus the 5.2M sqft North Campus expansion is itself substantially full once the dedicated Optimus factory, Cybercab production scaling, Cortex AI compute expansion, and Research Fab are accommodated; production capacity has to go somewhere with available greenfield acreage at brownfield-substrate prices.

Workforce within commuting reach of two major metros. A multi-phase fab at full scale employs 8,000-15,000+ direct workers plus supplier ring multiplier. The Brazos Valley substrate provides a structurally distinctive workforce pipeline through TAMU System: top-10 engineering school, largest research expenditure base in Texas, AggieFab Nanofabrication existing semiconductor research substrate, the Texas A&M Semiconductor Institute coming online 2028, Cyclotron Institute expertise, Center for Microdevices and Systems. TAMU College Station enrollment exceeds UT Austin. Plus the site sits within commuting reach of both Austin metro (~90 miles SW) and Greater Houston (~70 miles SE), providing multi-metro workforce draw redundancy that single-metro substrates lack. The Brazos Valley positioning is structurally why the site selection includes the entire Brazos Valley Semiconductor Concentration framing — it isn't just the Grimes County site, it's the multi-county engineering and operational workforce substrate.

Regulatory framework speed and continuity. Grimes County is a rural county with low population density (~30,000 residents), low-friction Commissioners Court framework, Texas property tax abatement framework via Chapter 312 plus Chapter 313 (now JETI under HB5), and pre-existing brownfield environmental remediation completed by Charah Solutions reducing TCEQ permitting friction. Texas state-level regulatory framework continuity (no state income tax, Texas Semiconductor Innovation Fund, Texas Enterprise Fund, Texas CHIPS Act framework, Senate Bill 6 implementation, ERCOT FERC-excluded grid) supports the multi-decade operator commitment that Terafab requires beyond any single political administration. California's regulatory framework would extend semiconductor fab construction timeline by 2-5+ years through CEQA, water rights litigation, environmental review, and broader regulatory friction; the Grimes County framework supports rapid permitting at the pace Terafab requires.

Geographic centrality to Tesla / SpaceX / xAI Texas operational portfolio. Gibbons Creek sits geographically central to the Texas-based Musk operating-entity portfolio: Giga Austin (~90 miles SW), Tesla Robstown lithium refinery (~250 miles S), SpaceX McGregor engine production (~100 miles NW), SpaceX Starbase (~400 miles S), SpaceX Bastrop semiconductor R&D (~85 miles SW), broader Texas SpaceX and Tesla footprint. Workforce flow plus supplier ring development plus inter-facility logistics plus integrated decision-making across Tesla / SpaceX / xAI operational coordination favor proximity over distance. Texas-state-level regulatory framework continuity across all Musk operating entities (vs. multi-state regulatory complexity) provides additional integration advantages.

Why not the obvious alternatives. Williamson County near Samsung Taylor: land prices have escalated substantially since Samsung commitment 2021, available contiguous acreage at the scale Terafab requires (3,000-6,000 acres) is genuinely scarce, ERCOT interconnection queue at Williamson County concentration is approaching saturation per 2025-2026 filings, workforce competition with Samsung Taylor and broader Williamson County operators raises labor costs, and higher cost basis vs. Grimes County brownfield substrate. Expanded Giga Austin North Campus footprint: the 2,500-acre Giga Austin substrate is substantially full at the 5.2M sqft North Campus expansion currently permitted, and production-scale fab requires physical separation from vehicle manufacturing operations (vibration, contamination, traffic patterns). Permian Basin behind-the-meter site: workforce availability constraints are binding (Permian engineering substrate is petroleum-focused, not semiconductor), distance from Austin / Houston / DFW workforce hubs ~400+ miles, water supply constraints (groundwater depletion, produced-water management complexity), and behind-the-meter natural gas works for data centers but doesn't supply the grid-stability characteristics semiconductor fabs require. Gulf Coast petrochemical complex: petrochemical operational substrate creates contamination, traffic, and regulatory compatibility friction; hurricane risk and broader weather event exposure; workforce availability mostly petroleum-focused; existing Gulf Coast industrial substrate has limited contiguous acreage available. Sherman TX area near TI: DFW metro proximity provides workforce but Sherman is ~75 miles north of Dallas, increasing commute friction; Texoma Lake water supply less abundant than Brazos River system or Gibbons Creek; Grayson County land prices have escalated since TI announcement. Arizona near TSMC / Intel: Phoenix metro water constraints (Colorado River allocation reductions through 2030) increasingly binding, APS interconnection queue saturation, workforce competition with TSMC Phase 1+2+3 plus Intel Arizona scaling, and distance from Tesla / SpaceX Texas operational hub.

The structural conclusion. Gibbons Creek emerges as essentially the unique US site that simultaneously satisfies all binding constraints at the scale Terafab requires within the timeline operators target. Greenfield sites fail the power timeline constraint. Williamson County fails the contiguous acreage constraint. Permian Basin fails the workforce constraint. Arizona fails the water and proximity constraints. New York and Ohio fail the proximity constraint plus higher cost basis. The brownfield substrate at Gibbons Creek is the structurally distinctive feature that makes the site selection possible at all — the prior TMPA coal plant created the power infrastructure substrate, the water rights framework, the transmission rights-of-way, the substation footprint, and the contiguous acreage profile that Terafab requires. Charah Solutions' 2021 demolition and remediation removed the operational generation while preserving the infrastructure substrate. The 2018-2026 interim eight-year window allowed regulatory simplification (decommissioning completed, environmental remediation completed, parcel-by-parcel disposition completed). The site became, for the specific use case of a leading-edge semiconductor production fab at announced scale, uniquely valuable in a way it had not been to other potential industrial users during the 2018-2026 interim. Musk's framing of the site as "one of several locations under consideration" hedges against Commissioners Court approval risk, but the constraint analysis suggests that no other US site simultaneously satisfies the binding constraints at the scale Terafab requires within the timeline the broader program targets. Substitution to alternative sites would extend timeline by 5-10 years through binding power and water constraints alone.

The Two-Facility Architecture

The Terafab program operates as a two-facility architecture with the Research Fab at Giga Austin North Campus and the Production Fab at Gibbons Creek in Grimes County. The two facilities are different buildings, at different locations, with different scales, and different roles in the broader Terafab silicon strategy. The architectural separation is structurally significant rather than incidental — leading-edge semiconductor process iteration cannot share line capacity with high-volume revenue production, and high-volume manufacturing cannot accommodate the experimental recipe variation that R&D requires.

Dimension	Research Fab (Giga Austin)	Production Fab (Grimes County)
Location	Giga Austin North Campus, Austin, Travis County	Gibbons Creek Reservoir, Grimes County
Operator	Tesla-led with Intel as foundry partner	SpaceX-operated (Space Exploration Technologies Corp. as filing entity)
Output Scale	A few thousand wafer starts per month (experimental)	100K–1M wafer starts per month at full scale (high-volume manufacturing)
Primary Role	Rapid iteration of process recipes, transistor architectures, packaging variants for AI5 successors, AI6, AI7	Volume manufacturing of validated designs supporting Tesla autonomy, SpaceX orbital compute, AI data centers
Capital Commitment	$3B Tesla initial cash commitment; $25-30B realistic equipped cost	$55B initial-phase commitment; up to $119B total full buildout
Status	Groundbreaking April 22, 2026; active earthwork underway	Tax abatement filing May 2026; Commissioners Court hearing June 3, 2026

The operational division Musk confirmed during Tesla's Q1 2026 earnings call clarifies the structural pattern: SpaceX handles high-volume chip manufacturing at the production fab, Tesla operates the smaller R&D pilot line at the Research Fab. The split entity structure (SpaceX as Grimes County filing entity, Tesla as Research Fab capital partner) plus the broader Terafab joint venture (Tesla, SpaceX, xAI as JV partners; Intel as foundry partner) plus the Intel 14A process technology selection collectively positions Terafab Grimes County as the production manufacturing complement that the broader Terafab program's announced 100-200 GW Earth compute plus 1 TW space compute target depends on. For comprehensive treatment of the Research Fab side of the program see Terafab Research Fab at Giga Texas.

The May 2026 Grimes County filing's $55 billion initial-phase commitment substantially exceeds Morgan Stanley's $35-45 billion pre-filing estimate and approaches the lower bound of Bernstein's full-buildout math when scaled across the multi-phase construction schedule. The $119 billion full-buildout figure validates the analytical position that the production fab requires capital substantially beyond what Musk's original March 21, 2026 announcement framed at $20-25 billion. The semiconductor industry's collective annual capex globally runs $200-250 billion; the Terafab Grimes County initial $55 billion commitment alone represents approximately 20-25% of one year of global semiconductor capex, sustained across the multi-phase construction schedule.

The split filing entity structure (SpaceX rather than Tesla) clarifies the capital allocation across the broader Terafab program. Tesla's prior $3 billion initial commitment to the Research Fab plus the SpaceX-led $55 billion Production Fab commitment plus the broader Intel foundry partnership plus the prospective xAI integration collectively positions the program's capital scale beyond what any single Musk operating entity could fund independently. SpaceX's June 2026 IPO targeting approximately $1.75 trillion valuation provides the public-market capital substrate that the Production Fab buildout depends on; the Production Fab capital commitment plus the broader IPO timing plus the Grimes County permitting timeline are structurally linked.

Process Technology and Customer Base

Musk has indicated during Tesla's Q1 2026 earnings call that the Production Fab will utilize Intel's 14A process technology — Intel's most advanced node currently in development, scheduled for high-volume manufacturing introduction at Intel's Arizona and Oregon facilities. Intel 14A is roughly 1.4nm-class and represents the next-generation node beyond Intel 18A (which is the process the Research Fab targets for AI6 and successor silicon). The 14A roadmap includes High-NA EUV lithography plus continued PowerVia backside power delivery plus continued advanced packaging integration.

Musk has separately referenced 2-nanometer chip production targets for Terafab, suggesting the Production Fab may host both Intel 14A capacity (for Tesla autonomy and SpaceX avionics applications) and 2nm-class capacity (potentially via TSMC-equivalent node licensing or Intel co-development) supporting the broader compute target. Musk's quote characterizing the facility as "the largest and most advanced chip fabrication facility in the world" suggests multi-node capacity at full buildout. The broader Terafab program's announced 100-200 GW Earth compute plus 1 TW space compute target requires advanced-node manufacturing capacity at scales that no single foundry currently operates.

The Production Fab's customer base maps directly to the broader Musk operating-entity portfolio plus the AI-Industrial convergence thesis the program advances. Tesla self-driving systems (Cybercab autonomous-native EV plus broader vehicle inference) plus Tesla Optimus humanoid robotics inference plus xAI training compute plus SpaceX orbital compute (Starlink and successors plus radiation-tolerant LEO compute) plus Tesla and SpaceX AI data centers collectively represent multi-hundred-billion-unit annual chip demand at full operating scale. Musk has framed the program's necessity in direct terms: "We either build the Terafab, or we don't have the chips, and we need the chips, so we build the Terafab" — a framing that the broader semiconductor industry's allocation patterns at TSMC, Samsung, and Intel substantiates.

Permitting and Community Context

The Grimes County Commissioners Court will hold a formal public hearing June 3, 2026 at 9:00 AM at the Grimes County Justice & Business Center (270 FM 149 W, Anderson, Texas) to consider approving the property tax abatement agreement for SpaceX Reinvestment Zone No. 1 – 2026-001. As of the May 2026 filing, neither SpaceX, Tesla, Elon Musk, nor Grimes County leadership had publicly commented on the project beyond the legally-required 30-day notice. Grimes County Judge Joe Fauth indicated availability for resident input but unavailability for media interview during the public notice period.

Local community response to the project has been mixed and reflects broader concerns about rapid industrial development in rural Grimes County. A resident community group formed months prior to the SpaceX filing in response to data center development proposals in the area has raised civic engagement concerns about the speed and limited disclosure of the SpaceX project. Earlier community meetings indicated some confusion about whether the project involved xAI (data center) or SpaceX (semiconductor manufacturing); the May 2026 filing formally identified Space Exploration Technologies Corp. as the business prospect. The brownfield character of the Gibbons Creek substrate (former coal-fired plant rather than greenfield rural land) reduces some land-use concerns relative to greenfield industrial development but does not eliminate broader concerns about water, infrastructure, traffic, and rural character impacts.

Beyond the June 3, 2026 Commissioners Court hearing, additional permitting and approval processes likely include Texas Commission on Environmental Quality (TCEQ) air quality permits for semiconductor fab operations, water rights and wastewater permitting for ultra-pure water and effluent management, ERCOT interconnection queue applications for power supply beyond the inherited TMPA capacity, Federal Communications Commission and Federal Aviation Administration approvals for any wireless infrastructure or airspace impacts, and broader regulatory compliance across federal CHIPS Act framework, federal export control compliance, and broader semiconductor industry regulatory substrate.

Outlook

The May 2026 Grimes County filing materially shifts the Terafab program's risk profile. The production fab moves from "TBD" to a confirmed brownfield substrate with existing power infrastructure, water access, and transmission capacity inherited from prior coal generation. The site selection plus the $55-119 billion capital framework substantially de-risks the production fab side of the program; the Research Fab on the Giga Austin North Campus remains the chip-and-fab co-design substrate that the production fab cannot replace.

Near-term operational questions resolving in 2026-2027 include whether the June 3, 2026 Commissioners Court approves the tax abatement agreement on workable terms, whether SpaceX June 2026 IPO completion provides the capital substrate that the Production Fab buildout depends on, whether continued Grimes County permitting plus broader Texas regulatory framework supports continued operational ramp, whether the Intel 14A process technology high-volume manufacturing introduction at Intel's Arizona and Oregon facilities translates to Production Fab capability transfer on a workable timeline, and whether continued Brazos Valley supplier ring development plus Highway 30 corridor infrastructure supports continued operational scaling.

Longer-term capital and operational questions resolving over the back half of the decade include whether the $55 billion initial-phase commitment converts to operational capacity on the implied timeline, whether the $119 billion full-buildout figure is achievable across the multi-phase construction schedule, whether the broader Terafab announced 100-200 GW Earth compute plus 1 TW space compute target is achievable on any timeline at any cost, and whether Bernstein's multi-trillion-dollar full-buildout estimate represents an unrealistic ambition or a generational program that the Grimes County substrate can host.

This page captures the May 2026 announcement-stage substance. As Commissioners Court approval, SpaceX IPO completion, Intel partnership operational scaling, EUV procurement, and continued construction substrate scaling become public, this page plus the broader Terafab program coverage will update accordingly. A Terafab Grimes County supplier-and-facility CSV dataset (paralleling the Phase 1.5 supplier datasets for Tesla Giga Texas, Samsung Taylor, Meta Temple, SpaceX Starbase, and Fermi America) becomes appropriate once construction substrate plus operational supplier ring development reaches sufficient operational disclosure.

Related Coverage