Corning Nvidia Optical Connectivity Partnership Expands US AI Infrastructure Supply Chain

Corning

Corning Nvidia optical connectivity plans will expand US manufacturing capacity for the optical systems needed to support artificial intelligence data centres. Corning will build three new manufacturing facilities in North Carolina and Texas as it targets a tenfold increase in optical connectivity output.

Corning Nvidia optical connectivity investment also includes a more than 50% expansion in Corning’s fibre production. The move reflects rising demand for high-speed data movement across AI infrastructure, where advanced optical links are becoming as important as chips themselves.

Corning Nvidia optical connectivity partnership strengthens the domestic supply chain around Nvidia’s AI computing ecosystem. Nvidia chips require high-performance optical fibre connectivity to move data quickly and at scale across large data centre networks.

The agreement also has a strategic materials angle. The fibre-optics industry is the largest US end-user of germanium, making AI data centre buildout increasingly relevant to minor metals demand.

AI Data Centres Drive Optical Connectivity Demand

AI workloads require massive data movement between chips, servers and storage systems. As computing clusters grow, copper-based connections face performance, distance and energy-efficiency limits in some high-speed applications.

Optical connectivity helps solve that problem. It allows data to move faster and across longer distances, supporting the scale required by advanced AI data centres.

Corning’s planned facilities in North Carolina and Texas will increase domestic capacity for these optical systems. That is important because AI infrastructure is becoming a national industrial priority, not only a technology market.

For Nvidia, the partnership supports the physical network behind its chips. AI accelerators create value only when data can move efficiently through the system.

For Corning, the deal gives stronger exposure to one of the fastest-growing infrastructure markets. Optical fibre, cable assemblies and connectivity products are becoming critical components in the AI supply chain.

Germanium Demand Links AI Growth to Critical Materials

The partnership also connects AI infrastructure to germanium demand. Germanium is used in optical fibre production, making fibre expansion relevant to critical minerals and specialty materials markets.

This matters because germanium supply is already strategically sensitive. It is used in fibre optics, infrared systems, semiconductors, defence electronics and solar applications.

If AI data centre construction accelerates, optical fibre demand could strengthen further. That would increase attention on germanium availability, recycling, refining and origin security.

The transaction also includes a financial component. Nvidia has the right to purchase up to 15mn shares of Corning stock at a fixed price of $180/share, as well as a pre-funded warrant to purchase up to 3mn shares for a total price of $500mn.

That structure shows how strategic customers are moving closer to upstream and midstream suppliers. Nvidia is not only buying components. It is helping secure the manufacturing base needed for future AI infrastructure.

For the US, the partnership supports domestic manufacturing around semiconductors, photonics and critical materials. It also reinforces the wider shift toward regionalised supply chains for high-value technology infrastructure.

The Metalnomist Commentary

The Corning-Nvidia partnership shows that AI supply chains are moving beyond chips into optical fibre, photonics and specialty materials. Germanium demand could become a hidden beneficiary as data centres require faster and more resilient optical connectivity.

Nvidia invests $5bn in Intel to co-develop next-gen x86 and AI infrastructure

Nvidia & Intel

Nvidia invests $5bn in Intel to accelerate advanced chip development and AI infrastructure. The firms will co-design future x86 platforms. They will integrate Nvidia chiplets into Intel products. Nvidia invests $5bn in Intel through common stock at $23.28 a share, subject to closing. As a result, both aim to strengthen US semiconductor capacity.

Strategic co-development and chiplet roadmap

The partnership targets modular designs using advanced chiplets. Nvidia’s accelerators will link with Intel cores and I/O. Consequently, OEMs could build denser AI servers with lower latency. The roadmap focuses on packaging, interconnects, and memory bandwidth. Nvidia invests $5bn in Intel to scale these platforms quickly.

Implications for US semiconductor supply chains

The deal complements recent CHIPS funding for Intel. It reinforces domestic manufacturing and design sovereignty. Meanwhile, the collaboration diversifies supply away from single-node risks. It also supports AI data center growth and edge computing. Therefore, US buyers gain a deeper, on-shore ecosystem.

The investment includes a planned $5bn equity purchase by Nvidia. Intel will jointly develop products that embed Nvidia chiplets. In parallel, Intel advances foundry services for high-volume AI parts. The US government’s 10pc stake in Intel adds policy backing. Together, these steps signal durable public-private alignment.

The Metalnomist Commentary

This tie-up tightens the US grip on high-end compute and packaging. If execution matches ambition, chiplet-based x86 platforms could reset AI TCO. Watch interoperability standards and packaging yields as leading indicators.

Lumentum Indium Phosphide Capacity Expansion Targets AI Optical Demand

Lumentum

Lumentum indium phosphide capacity is set to expand in the US as the photonics manufacturer increases production for artificial intelligence and cloud infrastructure. The company is preparing a new North Carolina facility focused on six-inch InP wafer production.

Lumentum indium phosphide capacity is strategically important because InP is a critical material for high-speed laser chips used in optical transceivers. These devices move large volumes of data between servers, switches and accelerators inside AI data centres.

Lumentum indium phosphide capacity expansion also reflects a broader shortage of compound semiconductor manufacturing capacity. AI infrastructure is now creating a new demand channel for InP beyond traditional telecom and sensing applications.

The company has ordered multiple InP production systems from German equipment supplier Aixtron. The North Carolina site is expected to open by mid-2028.

Nvidia Deal Anchors US InP Manufacturing Growth

Part of Lumentum’s future production is already committed to Nvidia under a multibillion-dollar supply agreement for advanced InP laser components. Nvidia is also investing $2bn in Lumentum’s manufacturing expansion.

That commercial backing is important because compound semiconductor fabs require large capital investment and long qualification cycles. Firm customer commitments reduce the risk of adding capacity ahead of demand.

AI data centres require increasingly fast optical links as computing clusters grow larger. Copper interconnects remain important at shorter distances, but optical systems become more critical as bandwidth and transmission distance increase.

InP laser chips sit at the centre of that shift. They enable high-speed optical transceivers used to connect servers and networking equipment inside cloud and AI infrastructure.

These AI-related applications now account for around 10% of global InP demand. That share could rise as data-centre operators move toward higher-speed optical architectures.

The expansion therefore gives Lumentum exposure to one of the fastest-growing segments of the semiconductor materials market.

Structural Supply Constraints Lift InP Investment

Global InP manufacturing capacity remains limited, prompting producers in several regions to expand.

UK semiconductor producer IQE has said it could convert some gallium arsenide capacity to InP if customer demand continues to strengthen. That illustrates how manufacturers are reallocating compound semiconductor assets toward higher-growth photonics markets.

Chinese producer AXT is also expanding aggressively. The company plans to double InP capacity in 2026 and again in 2027 to serve growing data-centre demand.

This competition shows that InP is becoming a strategic material rather than a niche semiconductor substrate. Supply availability, wafer quality and production scale will increasingly affect optical component growth.

For AI infrastructure, the bottleneck is shifting beyond processors. High-performance computing also depends on optics, power electronics, cooling and materials capable of moving data efficiently.

The next phase of AI investment will therefore benefit companies positioned in these supporting supply chains. Lumentum’s North Carolina expansion places InP wafer production directly inside the US manufacturing ecosystem.

The Metalnomist Commentary

AI growth is turning indium phosphide into a strategic semiconductor material, not just a telecom input. The companies that secure wafer capacity and qualified laser production early could gain significant leverage as optical connectivity becomes a core data-centre bottleneck.

Amazon Corning Fiber Optics Deal Expands US AI Infrastructure Supply

Corning

Amazon Corning fiber optics investment will expand US production of optical fiber, cable and connectivity systems used in data centers. Amazon plans multibillion-dollar investment under a multi-year agreement with Corning.

Amazon Corning fiber optics partnership reflects the rapid infrastructure buildout behind artificial intelligence and cloud computing. Faster data transfer between servers, accelerators and data centers is increasing demand for high-performance optical networks.

Amazon Corning fiber optics expansion also has a direct critical minerals angle. Fiber optics are the largest end-use market for germanium in the US, linking AI infrastructure growth with demand for a strategically important minor metal.

Corning expects the partnership to create about 1,000 advanced manufacturing jobs in North Carolina. The companies also plan to expand technician training to support the larger manufacturing base.

AI Data Centers Drive Optical Fiber Capacity Growth

Amazon will use Corning's optical fiber, cable and connectivity products across its data center network. The agreement gives Corning a major anchor customer as it expands domestic manufacturing.

AI computing requires enormous volumes of data to move quickly between processors, memory and networking systems. As computing clusters grow larger, optical connectivity becomes increasingly important for maintaining high bandwidth and low latency.

This means AI investment is spreading beyond semiconductor chips into photonics, fiber networks, power equipment, cooling systems and grid infrastructure.

Corning expects to increase fiber production by more than 50%. It also plans three new manufacturing facilities in North Carolina and Texas through a separate partnership with Nvidia.

The combination of Amazon and Nvidia demand gives Corning greater visibility for future capacity utilization. It also supports additional investment in US optical communications manufacturing.

Germanium Demand Gains New AI Infrastructure Exposure

The expansion is important for germanium because the metal is widely used in fiber optic manufacturing. Germanium compounds help control the refractive properties of optical glass, supporting efficient transmission of high-speed data.

Rapid fiber deployment therefore creates another demand channel for the minor metal. Individual fiber products use relatively small quantities, but large-scale data center construction can materially increase aggregate consumption.

Supply security is becoming more important because germanium production and processing remain concentrated. Greater AI-related demand could increase competition for refined and high-purity material.

For technology companies, optical connectivity is becoming as critical as access to processors and electricity. For materials suppliers, that means AI demand increasingly reaches deep into specialty metals and advanced materials supply chains.

Corning's expansion shows how data center investment can translate into new demand for physical infrastructure and strategic raw materials far beyond the semiconductor sector.

The Metalnomist Commentary

AI infrastructure is becoming a broad materials story, not just a chip story. As hyperscale data centers expand, germanium and optical fiber capacity could become increasingly strategic bottlenecks.

Amazon Corning Fiber Optics Deal Expands US Data Center Supply Chain

Corning

Amazon Corning fiber optics investment will expand US manufacturing capacity for optical fiber, cable and connectivity systems used in hyperscale data centers. Amazon plans multibillion-dollar investment under a multi-year supply agreement with Corning.

Amazon Corning fiber optics partnership reflects the growing infrastructure requirements behind artificial intelligence and cloud computing. Higher computing density requires faster data transmission between servers, accelerators and data centers, making optical connectivity an increasingly strategic component.

Amazon Corning fiber optics expansion also has implications for specialty metals. Fiber optics are the largest end-use market for germanium in the US, linking rapid data center construction with demand for a highly concentrated critical mineral.

The partnership will also create around 1,000 advanced manufacturing jobs in North Carolina and expand training for fiber optic technicians.

AI Data Centers Drive Optical Fiber Capacity Expansion

Amazon will use Corning's optical fiber, cable and connectivity products across its data center network. The agreement gives Corning a large anchor customer as it expands domestic production.

Data center growth is changing the materials required for digital infrastructure. AI systems need enormous computing capacity, but processors are useful only if data can move between them quickly enough.

Optical fiber addresses that bottleneck by supporting high-speed, high-bandwidth transmission over distances where conventional electrical connections become less efficient.

This means AI investment is increasingly spreading beyond semiconductor chips into photonics, optical networking, power equipment, cooling systems and grid infrastructure.

Corning expects to increase fiber production by more than 50%. The company is also planning three new manufacturing facilities in North Carolina and Texas through its partnership with Nvidia.

The combination of Amazon and Nvidia demand gives Corning strong visibility for future capacity utilisation. It also strengthens the case for further US investment in optical connectivity manufacturing.

Germanium Becomes More Exposed to Data Center Growth

The expansion carries a direct critical minerals angle because germanium is widely used in optical fiber production. The US fiber optics industry is already the country's largest end-user of the metal.

Germanium is used to modify the refractive properties of glass in fiber optic systems, making it an important material for high-performance communications infrastructure.

That creates a new demand link between AI investment and minor metals. Rapid data center expansion can increase germanium consumption even though only relatively small quantities are used in individual fiber products.

Supply security therefore matters. Germanium production and processing remain geographically concentrated, making diversified sourcing increasingly important as optical infrastructure expands.

For Amazon and other hyperscale data center operators, secure connectivity supply will become as important as access to processors and power. For Corning, the opportunity lies in combining domestic manufacturing scale with long-term demand from major technology customers.

The wider industrial message is clear. AI infrastructure is creating new demand not only for copper and semiconductors, but also for specialty materials embedded deeper inside optical communications systems.

The Metalnomist Commentary

Amazon's investment shows that the AI infrastructure race is moving into optical connectivity and its upstream materials. Germanium could become one of the quieter beneficiaries as data centers require increasingly dense, high-speed fiber networks.

Indium Phosphide Exports Become China’s New Chokepoint in AI Data Centre Supply Chain

AI data centre

Indium phosphide exports have become a strategic pressure point in the global AI data centre supply chain as China’s licensing controls delay shipments of a material essential for high-speed optical chips. The restrictions are exposing a new vulnerability in AI infrastructure: the physical materials behind silicon photonics and optical interconnects.

The issue has moved quickly from a specialist semiconductor concern to a high-level trade and industrial policy problem. Coherent, a key optical components supplier backed by Nvidia, warned in early May that indium phosphide shortages were already affecting the market. Its chief executive then joined a US business delegation to China as companies sought relief from export licence delays.

Indium phosphide exports matter because AI data centres are moving beyond copper-based interconnects. As AI workloads grow, hyperscalers need faster, lower-latency and more energy-efficient data transmission between processors, accelerators, switches and optical modules. Indium phosphide is one of the core materials enabling that shift.

The material is used in high-speed optical chips, lasers, detectors and photonic components. These devices support the optical links that move huge volumes of data across AI clusters. Without reliable indium phosphide substrates and wafers, the expansion of advanced AI data centre networks could slow.

China’s control over indium phosphide exports shows that critical materials policy is becoming more granular. Beijing no longer needs to restrict only rare earths or finished technology products. It can also influence upstream compounds, substrates and wafers that determine whether advanced semiconductor supply chains can scale.

Export Controls Expose a Hidden Bottleneck in Silicon Photonics

Silicon photonics has become a critical technology for AI infrastructure because it allows data to move through light rather than electrical signals. This reduces energy use per bit and supports the bandwidth required by large AI systems.

But silicon photonics is not only a silicon story. The most advanced optical systems often require compound semiconductor materials such as indium phosphide, gallium arsenide, gallium nitride and germanium-based compounds. Indium phosphide is especially important for lasers and high-speed optical devices.

This creates a difficult supply chain problem. AI companies, hyperscalers and chipmakers are racing to scale optical modules, but one of the key substrate materials remains highly concentrated. China is the world’s largest indium producer, accounting for about 70% of global output in 2024.

That concentration became more serious after China introduced export restrictions on indium phosphide in February 2025. Since then, licence delays have created backlogs for companies that manufacture or source InP substrates from China.

AXT, one of the world’s largest indium phosphide substrate producers and a major supplier to Coherent, said export permits were its most significant challenge. The company manufactures most of its InP substrates in China and only received its first permits last June. It still faces a large order backlog.

The effect has spread beyond individual suppliers. Coherent, Lumentum, VPEC and LandMark Optoelectronics all sit inside the optical components ecosystem that depends on reliable substrate supply. When permit delays hit upstream InP material, the impact moves through wafers, chips, optical modules and AI data centre equipment.

Prices show the severity of the shortage. Since China introduced export restrictions, the average price of a 6-inch indium phosphide wafer has surged by 250% to about $5,000. That price increase reflects both physical scarcity and the strategic premium attached to non-disrupted supply.

The supply squeeze also comes at a time of aggressive photonics investment. Nvidia announced $2bn investments each in Coherent and Lumentum in March. Marvell Technology also moved into photonics through its acquisition of Celestial AI, reflecting stronger demand for optical technology in AI computing.

These investments show where the industry is heading. AI infrastructure needs optical interconnects to manage power, latency and bandwidth. But China’s indium phosphide controls mean that materials availability could become a gating factor for deployment.

Companies are trying to respond. Coherent plans to double its InP wafer capacity at its Texas plant this year and more than double it again by the end of 2027. US photonics firms are also seeking supply from non-Chinese producers such as Sumitomo Electric Industries.

However, capacity additions are slow. New substrate plants can take two to three years to bring online. Qualification cycles are also long because optical chipmakers cannot easily switch substrate suppliers without testing performance, reliability and consistency.

This makes the shortage difficult to solve quickly. Even if new capacity is announced, it may not arrive fast enough to meet near-term AI data centre demand. Meanwhile, many non-China producers already consume part of their own output internally, reducing the amount available to the broader market.

China’s Materials Chokepoint Strategy Strengthens Domestic Producers

China’s indium phosphide export controls are creating both pressure and opportunity. They restrict global supply, but they also support domestic Chinese substrate producers that are expanding capacity.

Yunnan Germanium, Guangdong Xiandao and Zhuhai Dingtai Xinyuan are among China’s leading domestic InP substrate players. Their role is becoming more important as Beijing uses materials controls to strengthen strategic leverage across semiconductor and AI supply chains.

Yunnan Germanium has already moved to expand. The company announced a 189mn yuan investment in April to raise production capacity to 450,000 single InP wafers annually. Its shipments of InP wafers rose by 74% in 2025, showing fast domestic market growth.

Guangdong Xiandao is also expanding through its subsidiary Guangdong Xianrui. The project is expected to produce 40 t/yr of indium phosphide crystals, which are used as raw material for substrates.

These investments fit a broader pattern. China is not only defending control over upstream critical materials. It is also building downstream processing capacity in higher-value compound semiconductor materials.

However, Chinese producers may not immediately solve the global shortage. Some are still seeking export approvals, and any overseas shipments may be limited. Domestic demand remains a priority, especially as China builds its own AI, optical communications and semiconductor ecosystem.

AXT

Supplier qualification creates another barrier. Companies such as Coherent and Lumentum are unlikely to switch easily from established suppliers. Coherent relies heavily on AXT, while Lumentum sources mainly from Sumitomo and JX Advanced Metals. New suppliers must pass demanding qualification cycles before they can enter critical optical chip supply chains.

This gives China’s export controls a long-lasting effect. Even if alternative suppliers exist, the market cannot instantly redirect demand. The bottleneck is not only production volume. It is qualified, high-quality, customer-approved substrate supply.

The strategic lesson is clear. AI supply chains are not only exposed to advanced chips, GPUs and packaging capacity. They also depend on a deep materials stack that includes indium, phosphorous chemistry, InP crystals, substrates, wafers, lasers, detectors and optical modules.

This is why indium phosphide exports have become so important. AI data centre buildouts need more optical links as clusters grow larger. Copper interconnects face limits in speed, distance and energy consumption. Photonics offers a solution, but only if the materials chain can scale.

For the US and its allies, the response will likely require more than emergency licence negotiations. It will require investment in indium recovery, InP crystal growth, substrate manufacturing, wafer capacity and long-term offtake agreements. It may also require strategic stockpiles for high-purity indium and compound semiconductor substrates.

The issue also strengthens the case for recycling and secondary recovery. Indium is often produced as a by-product, making primary supply difficult to expand quickly. Recovering indium from industrial scrap, displays, semiconductors and related waste streams could become more important if export controls persist.

For AI data centre developers, the risk is timing. Demand for optical modules is accelerating now, while new ex-China capacity may not fully arrive until 2027 or later. That mismatch could raise costs, delay deployments and intensify competition for qualified photonics suppliers.

The market may therefore see a split. Companies with secured InP supply will be better positioned to support hyperscaler demand. Companies exposed to licence delays, qualification bottlenecks or spot-market wafers may face higher costs and delivery risk.

The Metalnomist Commentary

China’s control over indium phosphide exports shows that the AI race is becoming a materials race. The next bottleneck may not be only GPUs or power supply, but the compound semiconductor substrates needed to move data fast enough inside AI clusters.

Trump-Xi Tariff Talks Yield No Deal, TikTok Sale Path Clears

Trump-Xi

Trump-Xi tariff talks produced no tariff agreement on Friday. However, both sides signaled movement on a potential TikTok sale. Trump thanked Xi for “the TikTok approval” and cited progress on trade and fentanyl.

China’s readout framed the Trump-Xi tariff talks around fairness for Chinese firms. Beijing urged an “open, fair, non-discriminatory” US business environment. It also welcomed ByteDance’s negotiations with US buyers.

Earlier this year, Trump linked tariff relief to a TikTok sale. The Trump-Xi tariff talks did not resolve broad duties that now burden bilateral trade. US energy exports to China remain largely curtailed.

Tariffs, agriculture, and new measures

US soybean sales to China have lagged early in the 2025-26 season. Through two weeks, commitments trailed last year’s 5.94mn t pace. Meanwhile, Washington is assessing a 30pc broad tariff on Chinese imports.

China currently applies a broad 10pc tariff on US goods. It also adds 10–15pc on energy and farm commodities. Without a deal by 10 November, both sides warn rates could rise by 24 percentage points.

Beijing opened an antitrust probe into Nvidia last week. The US plans port fees on Chinese ship operators and vessels from 14 October. Those charges include $50/net ton and $18/net ton, respectively.

Diplomacy calendar ahead

Trump plans to meet Xi at APEC in South Korea on 31 October–1 November. He also outlined travel to China early next year. Xi may visit the US “at an appropriate time,” pending further progress.

The Chinese readout noted Xi’s 3 September military parade discussion. Both parties kept channels open despite unresolved tariff issues. Talks continue while agriculture and technology disputes persist.

The Metalnomist Commentary

Tariff uncertainty continues to cloud pricing and procurement cycles. Near-term decisions on duties and platform divestment will shape fourth-quarter trade flows and 2026 planning across trans-Pacific supply chains.

Tantalum Prices Surge as AI Capacitor Demand Meets Tight African Supply

Ta (Tantalum)

Tantalum prices have surged across concentrates, metal, and scrap as capacitor demand rises and supply disruption tightens the upstream market. The rally reflects a rare collision between stronger electronics consumption, AI data centre investment, and instability in central African raw material flows.

The price rise began before the latest disruption. However, the February landslide at the Rubaya mine in rebel-held eastern Democratic Republic of Congo intensified the market shock. Tantalum concentrate prices jumped sharply, and the pressure quickly moved into tantalum metal and scrap markets in Europe and the United States.

The supply impact has been especially important because much of the material linked to eastern DRC moves through Rwanda before entering international trade. Any disruption around Rubaya therefore affects more than one mining district. It also exposes how dependent the tantalum supply chain remains on politically fragile and difficult-to-monitor sources.

AI Data Centres Lift Tantalum Capacitor Demand

AI data centres have become a major new demand driver for tantalum capacitors. These components help regulate electricity flow on circuit boards and remain critical in high-performance electronics. As AI servers expand, demand for tantalum and tantalum-polymer capacitors is rising alongside advanced chips, power systems, and server hardware.

This demand is not theoretical. Boards used for Nvidia H100 cards can contain multiple tantalum and tantalum-polymer capacitors, making AI infrastructure a direct consumption channel for tantalum products. As Alphabet, Microsoft, Meta, and Amazon raise capital spending for AI infrastructure, smelters and traders expect stronger orders for next-generation capacitor materials.

The capacitor industry is already responding to cost pressure and higher demand. Manufacturers raised prices across several product ranges in 2025, including tantalum-polymer capacitors and multilayer ceramic capacitors. Higher tantalum powder costs contributed to the increases, but AI-related consumption has become an equally important factor.

Tight Supply Pushes Tantalum Metal and Scrap Higher

Tantalum prices are rising because the market faces pressure at both ends of the value chain. Upstream concentrate availability has tightened, while downstream buyers in capacitors and alloys continue to compete for material. This has lifted prices for concentrates, refined metal, and scrap at the same time.

The alloy sector has added another layer of demand stability. Even as capacitor demand accelerates, industrial users of tantalum metal continue to require reliable supply for high-performance applications. As a result, scrap has become more valuable because it offers an alternative source of tantalum units when mined supply becomes uncertain.

The current rally also echoes earlier technology investment cycles. The Dotcom boom drove strong demand for tantalum capacitors and pushed tantalite prices to historic highs. Today, AI infrastructure may be creating a similar demand shock, but with a more complex supply chain and tighter scrutiny around conflict-linked minerals.

The Metalnomist Commentary

Tantalum is becoming a hidden beneficiary of the AI infrastructure boom. The market risk is that capacitor demand can scale faster than responsible mining, refining, and recycling channels can respond.

Texas Instruments Expands Chip Manufacturing with Two New Fabs in Texas

Texas Instruments

Texas Instruments Expands Domestic Semiconductor Capacity

Texas Instruments (TI) will build two new semiconductor fabrication facilities in Sherman, Texas, as part of its $60bn investment plan. The company aims to strengthen domestic semiconductor production amid rising demand and strategic US policy support. These facilities will join seven other fabs already operating or planned across Texas and Utah, collectively expected to produce hundreds of millions of chips daily at full capacity.

Federal Support Boosts Semiconductor Manufacturing Investment

The US government is playing a central role in enabling TI’s semiconductor expansion. In 2024, TI secured $1.6bn under the US Chips and Science Act to support construction of its first two Sherman fabs and its second Lehi, Utah, facility. The company also expects $6bn–8bn in additional tax credit funding from the US Treasury Department’s advanced manufacturing investment program. If Congress raises the investment credit percentage from 25pc to 30pc, TI could gain further support for its projects.

Broader Impact on Semiconductor Supply Chains

TI’s semiconductor fabs will play a crucial role in securing critical supply chains for US industries. The company produces analog and embedded processing semiconductors for major customers such as Apple, Ford, Medtronic, Nvidia, and SpaceX. By expanding capacity, TI will help reduce reliance on foreign semiconductor sources, strengthen resilience against supply shocks, and enhance competitiveness across technology and automotive markets.

The Metalnomist Commentary

Texas Instruments’ expansion reflects the reshaping of global semiconductor supply chains under US industrial policy. While the $60bn investment is significant, its real impact lies in reducing import dependency and fortifying key technology sectors. The fabs will strengthen US capabilities in advanced manufacturing, ensuring long-term resilience in a geopolitically sensitive industry.

AI Demand for InP Drives Major Semiconductor Capacity Expansion

Indium Phosphide

AI demand for InP is triggering a global investment cycle as semiconductor and photonics manufacturers race to expand indium phosphide wafer capacity. Supply remains tight as AI data centres require increasingly powerful optical interconnects between processors, memory and networking systems.

AI demand for InP has prompted Coherent to plan successive capacity expansions at its Sherman, Texas, facility. The company intends to double production by the end of 2026 and double capacity again by the end of 2027.

AI demand for InP is also accelerating investment in Japan, China and downstream US photonics manufacturing. Long-term supply agreements and customer prepayments show that manufacturers increasingly view access to InP substrates and components as a strategic constraint.

The expansion creates a direct link between AI infrastructure investment and indium demand, pushing a traditionally small specialty metal deeper into the semiconductor supply chain.

Six-Inch InP Wafers Become Strategic for AI Photonics

Coherent's Texas facility is focused on high-volume 150mm, or six-inch, InP semiconductor manufacturing. The company sees the transition from three-inch to six-inch wafers as a major driver of future production efficiency.

Larger wafers allow manufacturers to produce more devices from each substrate and potentially improve manufacturing economics. Coherent has also reported stronger yields from its six-inch lines compared with three-inch production.

The company shipped its first transceivers containing components manufactured on six-inch InP lines during the first quarter.

US industrial policy is supporting the expansion. Coherent has reached a preliminary agreement for up to $50mn in direct funding under the CHIPS and Science Act.

Nvidia has also invested $2bn in Coherent and secured supply through the end of the decade. This highlights how AI companies are moving upstream to secure optical component capacity rather than relying solely on conventional procurement.

Coherent also plans to establish six-inch InP production at another facility in Zurich, further expanding its geographic manufacturing base.

Japan's JX Advanced Metals is pursuing an even larger capacity programme. The company plans to invest up to ¥120bn over four years and increase InP substrate capacity by seven to ten times.

Expansion will include additional production at Isohara Works and new capacity in Hitachinaka. JX is also discussing price revisions with customers as it seeks to build a more sustainable long-term supply model.

Tight InP Supply Extends Into Indium and Optical Manufacturing

China is seeing similar efforts to secure material. AXT's Beijing Tongmei subsidiary has signed a 2027 supply agreement with Nanjing Casela Technologies covering 173mn yuan of InP substrate wafers.

Casela will prepay 50% of the contract value, demonstrating the level of competition for future wafer availability. The agreement also gives the customer priority for additional volumes if capacity allows.

Export restrictions on Chinese InP have added another source of supply uncertainty, increasing the incentive for customers outside China to secure alternative production.

The shortage is also driving investment further downstream. Nokia is expanding its advanced photonic semiconductor test and packaging operation in Pennsylvania by up to ten times.

The facility packages photonic chips into optical modules used in AI and telecommunications infrastructure. Nokia is investing around $30mn alongside federal and state support.

These investments show that the bottleneck extends across the entire photonics chain, from InP substrates and epitaxial wafers to lasers, modulators, transceivers and packaging.

For metals markets, the expansion strengthens the outlook for indium. InP production requires high-purity indium, meaning sustained AI-driven wafer growth could place additional pressure on upstream refining and feedstock availability.

The Metalnomist Commentary

AI is turning indium phosphide from a specialised telecom material into strategic computing infrastructure. The next bottleneck could move upstream from photonic devices to high-purity indium and qualified six-inch wafer capacity.

STMicro targets new sources of silicon chip demand

STMicro

STMicro is betting on new sources of silicon chip demand from AI data centres and advanced EV platforms. The company sees silicon and silicon carbide (SiC) devices as core growth engines, even as it trims some near-term capital spending. As a result, STMicro is repositioning its portfolio toward photonics, high-voltage power electronics and next-generation vehicle systems that can unlock new sources of silicon chip demand across several end markets.

Photonics and AI servers anchor new sources of silicon chip demand

STMicro is using silicon photonics to capture new sources of silicon chip demand from hyperscale AI data centres. In the third quarter, the firm reported rising orders for silicon photonics integrated circuit prototypes from its 300mm fabs. This confirms internal expectations that photonics ICs will become a meaningful revenue driver as AI server build-outs accelerate worldwide.

Meanwhile, the company is tightening its integration into the AI infrastructure ecosystem through the Starlight Consortium. The consortium connects substrate suppliers, device makers and system integrators to develop high-speed optical solutions for data centres, telecoms and automotive. In parallel, STMicro is collaborating with Nvidia on an 800V DC AI data centre architecture that combines SiC, gallium nitride (GaN) and silicon technologies. The firm has also demonstrated a GaN prototype with over 98pc conversion efficiency, underscoring how power density and efficiency now drive silicon chip demand as much as raw compute.

SiC, EV power electronics and a more selective capex strategy

STMicro is broadening SiC usage beyond traction inverters and onboard chargers to new power roles in active suspension inverters. This expands the addressable market in EVs just as advanced driver-assistance systems increase demand for sensors and control chips. However, weaker-than-expected EV programmes in Europe and China have delayed the full impact, forcing the company’s main automotive customer to cut orders.

Therefore STMicro is trimming near-term SiC capex as it transitions production from 150mm to 200mm wafers. The company will slow some investments in SiC conversion while maintaining its 300mm expansion plans in Agrate, Italy, and Crolles, France. These fabs continue to see solid order visibility, particularly for data-centre power modules and optical components. STMicro expects EV growth and SiC restocking to resume in 2026, once inventory drawdowns run their course and regional electrification policies translate into firmer demand.

The Metalnomist Commentary

STMicro’s push into photonics and SiC power devices shows how new sources of silicon chip demand are shifting toward AI infrastructure and complex EV systems. For the wider materials chain, this means sustained pull for high-purity silicon, SiC substrates and GaN epitaxy, even if near-term EV softness delays some projects. Suppliers that align with 300mm and 200mm wafer roadmaps and can support AI-class power densities will be best positioned as these new demand waves crystallise from 2026 onward.