GlobalFoundries (GF), the major US semiconductor foundry, announced on October 14, 2026 that it has been selected as the technology provider and operator for two semiconductor fabs to be built in Malta, New York. Under a US-Japan investment initiative, Japan would supply up to $21 billion in funding, while GF would provide manufacturing technology and run the fabs. The first fab is scheduled to begin operating in 2029, and existing support under the CHIPS Act will also be redirected toward investments in optical communications and advanced packaging.
The initiative combines Japanese funding with American manufacturing capability to strengthen the domestic US semiconductor supply base. Even so, securing large amounts of capital and building fabs is a separate challenge from establishing a system that can reliably supply chips. The plan's success will depend not only on GF's volume-production technology but also on whether it can secure enough customer demand to make use of the new capacity.
Japan to Provide Up to $21 Billion, With GF Handling Technology and Fab Operations
The plan calls for an initial investment of about $10 billion to build the first fab, which will process 300mm silicon wafers. A further investment of about $10 billion would then allow construction of a second fab to begin.
The announced ceiling is $21 billion, but GF has not disclosed how the difference from the roughly $10 billion needed for each fab will be allocated. Nor does the announcement say that funding for both fabs has already been fully secured and committed.
The division of roles is that Japan supplies the funding, while GF provides volume-production technology and operates the facilities.
GF CEO Tim Breen also explained that building large-scale chip fabs requires not only vast capital but also manufacturing technology and the ability to operate the plants.
A fab cannot ship large volumes of products simply by having buildings and equipment in place. Stable volume production requires introducing and tuning manufacturing processes, improving yields, and building quality-control systems. GF will be responsible for establishing this manufacturing know-how and running the fabs.
The broader $550 billion US-Japan strategic investment framework is also important for understanding Japan's side of the investment.
At a press conference on February 20, 2026, Japan's Ministry of Economy, Trade and Industry (METI) said it intends to use equity investments and loans from the Japan Bank for International Cooperation (JBIC), as well as guarantees on private-sector loans from Nippon Export and Investment Insurance (NEXI).
At that time, the ministry indicated that it would proceed with projects while confirming that principal, interest, and guarantee fees could be expected to be recovered and that Japan would also benefit.
However, this was an explanation of the overall US-Japan investment framework and does not specify the financial terms that will apply to the GF project.
GF's announcement does not detail how much of the up-to-$21 billion will take the form of equity versus loans, how ownership of the new fabs will be divided, or on what terms investment returns will be shared.
Two New Fabs With 1.8 Million Wafers a Year in Capacity; First Fab to Start in 2029
According to the plan GF announced on October 14, each new fab is expected to have a full-operation capacity of about 900,000 wafers per year, measured in 300mm wafers.
The second fab is planned to have comparable capacity, so if both reach full operation, they could process about 1.8 million wafers a year by simple addition.
Based on GF's official announcement, the investment scale and operating plans can be summarized as follows.
| Construction phase | Investment scale | Planned start of operations | Annual capacity at full operation |
|---|---|---|---|
| Fab 1 | Initial investment of about $10 billion | 2029 | About 900,000 300mm wafers |
| Fab 2 | Additional investment of about $10 billion envisioned | Not announced | Comparable to Fab 1 |
| Total for both fabs | Up to $21 billion for the overall initiative | Full-operation timing for both fabs not announced | About 1.8 million by simple sum |
The figure of 1.8 million wafers a year, however, represents capacity if both fabs are built as planned and reach full operation. It does not guarantee actual shipment volumes. Wafer throughput also differs from the number of chips produced from those wafers, which varies with die size, yield, and other factors.
Furthermore, 2029 refers only to the planned start of operations at the first fab. It has not been stated that the fab will reach full capacity of 900,000 wafers a year in 2029, and the start date for the second fab has not been announced.
The eventual 1.8 million-wafer expansion and the point at which that capacity actually becomes available should therefore be considered separately. The specific manufacturing processes the new fabs will use, and how capacity will be allocated among chip types, have also not been disclosed.
On the construction side, GF can take advantage of an existing site.
GF plans to use its Malta site, where the necessary permits have already been obtained and preparations for groundbreaking are complete. It says leveraging existing infrastructure will help speed construction. Construction of the first fab is scheduled to begin within 2026.
On jobs, GF expects about 27,000 positions through construction and related activities, plus more than 1,000 people directly employed in semiconductor manufacturing.
However, the employment impact from construction and related business differs from the workforce that will work at the fabs on an ongoing basis. The roughly 27,000 jobs do not all mean permanent employees at the new fabs.
$955 Million in CHIPS Act Support Reallocated to Optical Communications and Advanced Packaging
Alongside the new fab plan, GF will also redirect existing support from the US government.
GF agreed with the US Department of Commerce to modify the support it receives under the CHIPS and Science Act and to reallocate $955 million.
The reallocated funds will go toward technology development and capacity expansion in silicon photonics, silicon germanium (SiGe), and advanced packaging at its sites in Malta, New York, and Burlington, Vermont.
Silicon photonics uses semiconductor technology to handle optical signals, supporting high-speed, high-capacity data transmission in data centers and communications infrastructure. Advanced packaging, meanwhile, integrates multiple chips and memory at high density to boost overall system performance.
As demand for AI computing grows, it matters not only how fast chips can compute but also how efficiently large volumes of data can move between chips and between devices.
The reallocation can be seen as an effort to use funding from Japan to advance construction of the new fabs while using existing US government support to strengthen the technologies needed for data communications and chip packaging in the AI era.
However, it has not been disclosed how much of the silicon photonics, SiGe, and advanced packaging products will be manufactured at the two planned new fabs.
$2 Billion Manufacturing Agreement With TSMC: AI Chip Components to Be Made in the US
On October 8, ahead of this announcement, GF also announced a $2 billion manufacturing agreement with TSMC.
According to the announcement of the GF-TSMC agreement, GF will manufacture and supply, in the US, the silicon interposers used in TSMC's advanced packaging technology, CoWoS. Volume production is scheduled to begin in the first half of 2028, with an initial contract term of five years.
Silicon interposers are a key component used to connect compute chips such as GPUs with high-bandwidth memory (HBM) at high density.
Being able to make this component in the US would allow part of the supply chain needed for advanced packaging of AI accelerators to be built domestically. For GF, it would also broaden its role beyond conventional chip manufacturing to supporting the production and packaging of AI semiconductors.
That said, the interposer volume-production plan and the plan to build the two new fabs should be treated separately for now.
Volume production of interposers for TSMC is scheduled to start in the first half of 2028, while the first new fab is planned to begin operating in 2029. GF has not said that the 2028 interposer production will take place at the new fabs.
Even if initiatives proceed at the same Malta site, individual manufacturing plans cannot simply be treated as the operating schedule of a single fab.
In June 2025, GF also announced a $16 billion investment plan to expand semiconductor manufacturing and advanced packaging capacity in New York and Vermont.
It has not been disclosed how much funding and equipment overlap between the current initiative of up to $21 billion and the 2025 plan for $16 billion. It would therefore not be appropriate to simply add the two together and conclude that GF will newly invest $37 billion in the US.
The announcement also revealed that the US Department of Commerce plans to acquire about 4% of GF's shares under a separate agreement.
However, that is a stake in the company GF, not an indication of ownership share in the two new fabs. Details such as the consideration paid for the shares and voting rights have not been announced, and it has not been explained that the reallocated $955 million in support will go directly toward the share purchase.
Renesas Plans to Use the New Fabs, With a Japanese Company as Customer
In this US-Japan investment initiative, Japanese companies are involved not only as providers of funding but also as users of chips made in the US.
Renesas Electronics CEO Hidetoshi Shibata explained that the company positions the new capacity GF is building in New York as an extension of its own manufacturing base and has incorporated it into its business plan.
By using US manufacturing capacity, the company aims to ensure a stable long-term supply of the chips needed in areas such as electrification, communications, and AI.
This cooperation builds on manufacturing collaboration the two companies have already developed.
In February 2026, Renesas and GF announced an expansion of their multibillion-dollar strategic partnership.
Under this collaboration, Renesas will be able to use GF's FDX (FD-SOI) and BCD technologies, as well as CMOS processes with non-volatile memory capabilities.
Targeted products include SoCs, power-management chips, and microcontrollers (MCUs). The companies envisage manufacturing chips used in automotive radar, battery management systems, and industrial equipment communications.
The two companies had indicated plans to begin manufacturing in the US first and then expand to GF's manufacturing sites in other regions.
They are also considering transferring some of GF's manufacturing processes to Renesas's fabs in Japan.
In other words, beyond using US manufacturing capacity, GF's manufacturing technology could also be used to strengthen Japan's production base.
However, the transfer of processes to Japan has not been completed, and it has not been announced that Renesas has signed a contract to purchase a certain amount of capacity at the two planned new fabs.
How much customer demand has actually been secured for the new fabs will need to be checked going forward.
Tokyo Electron, a longtime equipment and technology partner of GF, also expressed its intention to contribute to expanding US semiconductor production capacity.
If the plan advances, Japanese companies could gain opportunities both to supply chipmaking equipment and to procure semiconductors produced in the US.
Still, how much business benefit they actually gain will depend on how concrete equipment orders and chip supply contracts become.
Can the Up-to-$21 Billion Plan Be Realized? Customer Demand Is the Key
Sustained customer demand will be crucial as GF moves ahead with the new fabs.
In October 2025, GF Chairman Tom Caulfield explained that the pace of investment in expanding US chip manufacturing capacity depends on customers showing willingness to move production to the US and committing to specific orders.
In remarks on CNBC that GF shared, he also expressed the view that building new capacity without customer demand would be pointless.
The US-Japan investment initiative gives GF a path to securing the funds needed to build the new fabs. But turning that into fabs with capacity of about 900,000 wafers a year requires introducing specific manufacturing processes and securing customers to use them.
It also takes time to move from installing chipmaking equipment through starting volume production and improving yields to reaching full operation. Even if the first fab begins operating in 2029, it may not be able to deliver its planned capacity immediately.
Going forward, important factors to watch will be the start of construction scheduled within 2026 and progress toward the 2029 start of operations, as well as the ramp-up of volume production, specific customer contracts, and whether the additional investment for the second fab is made.
If GF can secure demand from customers such as Renesas and put a volume-production system in place as planned, the new fabs could become an important manufacturing base for Japanese companies to reliably procure chips for the US market.
If investments in silicon photonics and advanced packaging using CHIPS Act support lead to actual product supply, GF could also strengthen its role in supporting communications and advanced packaging in the AI era, beyond conventional chip manufacturing.
What is at stake in this plan is not the investment scale of up to $21 billion itself. It is whether enormous funding can be converted into actual semiconductor production capacity and whether that capacity can be tied to customer demand. Whether the new fabs built through US-Japan cooperation function as a sustainable supply base will become clear through future volume-production results and how customers use them.