CHIP INDUSTRY NEWS Chip industry update: A review of Q2 2025

Related Vendors

ROHM Semiconductor GmbH

Schunk Sonosystems GmbH

Heraeus Electronics GmbH & Co. KG

The global semiconductor industry in Q2 2025 saw record-breaking growth driven by AI demand, but this masked widening fractures in traditional markets and supply chains. As geopolitical tensions and infrastructure constraints escalate, the sector faces a future defined as much by strategic resilience as by technological innovation.

The global semiconductor industry closed out the second quarter of 2025 in a state of paradox. On one hand, it saw record-breaking revenues and continued investment in cutting-edge technologies. On the other, it grappled with bottlenecks, mounting geopolitical pressures, and an emerging shortage that threatens to upend automotive supply chains once again.

At the heart of the story is an increasingly bifurcated market. Explosive growth in AI-centric applications is masking flat or declining performance in traditional sectors like smartphones, PCs, and industrial electronics. While headlines tout a thriving industry, a deeper look reveals a fragile, imbalanced system vulnerable to external shocks and structural blind spots.

AI: The engine driving unprecedented growth

The most visible force in Q2 was the generative AI boom. Demand for GPUs, NPUs, and custom AI chips surged past all expectations, with projected annual sales of AI semiconductors now exceeding $125 billion, accounting for more than 20 % of global chip revenue.

This appetite for compute is not confined to the data center anymore. AI-powered PCs and smartphones, featuring on-device neural processing, are revitalizing mature markets. AI PCs are forecast to make up around 50 % of all units sold in 2025, while AI-infused smartphones are expected to surpass 1.24 billion units globally. However, these gains are modest compared to the surging valuation of AI infrastructure players and suppliers like TSMC and NVIDIA.

CHIP INDUSTRY NEWS

Chip industry update: A review of Q1 2025

TSMC, in particular, has responded with unmatched aggression, announcing a $42 billion capex plan and doubling its CoWoS (Chip-on-Wafer-on-Substrate) packaging capacity. Advanced packaging, not just silicon, has become the key battleground for performance and differentiation in AI hardware.

But success has bred pressure. NVIDIA, the undisputed king of AI chips, has stumbled with manufacturing delays on its next-gen Blackwell GPUs, triggering a temporary $300 billion loss in market cap. Meanwhile, hyperscale cloud providers are increasingly exploring alternatives, creating rare competitive openings for challengers like AMD.

Traditional markets show hidden cracks

While AI supercharged one part of the industry, the rest of the chip ecosystem showed only modest gains or, in some cases, signs of stress. Smartphones and PCs saw minimal recovery, and overall wafer shipments declined in 2024 before rebounding narrowly in 2025. The value of the market is now being carried disproportionately by a few high-margin AI components rather than broad-based volume growth.

This asymmetry is fueling a potentially dangerous underinvestment in older process nodes (40nm and above) that still serve vital roles in automotive, industrial, and consumer electronics. The result? A second semiconductor shortage is quietly taking shape. This time, it’s not due to pandemics or panic buying, but because the AI gold rush is pulling foundry capacity away from sectors that can’t match AI’s margins.

Auto manufacturers are bracing for a supply crunch in late 2025 and into 2026, as foundries prioritize newer nodes over mature ones used in MCUs, analog chips, and power components critical to vehicles. Compounding the issue, new U.S. tariffs are expected to increase car prices, recalling the disruptions of 2021.

PART 2: SEMICONDUCTOR MANUFACTURING

The Semiconductor “Wafers”

Geopolitical upheaval creates new challenges

The return of the Trump administration has introduced a wave of tariff threats and regulatory reversals. A July 9 deadline has been set for new bilateral trade deals, with tariffs on chip materials and equipment looming. These levies could add billions to the cost of projects like TSMC’s Arizona fab, effectively neutralizing the incentive offered by the CHIPS Act.

Simultaneously, the administration rescinded the AI Diffusion Rule, a wide-reaching Biden-era export control measure. The reversal aims to restore market access for U.S. chipmakers like NVIDIA and AMD but introduces fresh legal uncertainty around compliance and enforcement.

The U.S. CHIPS Act, once seen as a game-changer, is now under review. Over 100 semiconductor projects announced under the act are in limbo, with major recipients like Intel experiencing significant delays. Intel’s $20 billion Ohio facility, originally slated for 2026, is now pushed out to 2030–2031.

In contrast, the EU’s Chips Act has mobilized over €80 billion in investment, approving seven new fabs. Yet, an audit by the European Court of Auditors questions whether the EU can realistically meet its goal of capturing 20 % of global market share by 2030.

Asia remains aggressive: China continues its self-sufficiency drive with a $47.5 billion “Big Fund 3.” South Korea’s $450 billion strategy and Japan’s ¥10 trillion commitment are reshaping the regional landscape. Even India is stepping up, easing SEZ rules to fast-track domestic fabs under its India Semiconductor Mission.

SEMICONDUCTOR INDUSTRY

Foundry market outlook 2025: AI and advanced technologies drive growth

What the big players are currently doing

TSMC remains the clear leader, responding to AI-driven demand with unmatched speed and scale. Its capital expenditure plan for 2025 exceeds $40 billion, with advanced node capacity and packaging facilities scaling rapidly across Taiwan, Japan, Germany, and the U.S. This expansion has fortified its position as the primary supplier of the world's most critical AI hardware, pushing its valuation beyond $1.1 trillion.

Intel, in contrast, is fighting an uphill battle. While it continues to make progress on its 18A process node and remains the first customer for ASML’s High-NA EUV systems, its operational troubles are deep. The firm has announced sweeping layoffs, delayed flagship construction projects, and now faces investor skepticism over whether it can realistically challenge TSMC in foundry services.

Samsung is in a period of strategic reorientation. After lagging its peers in AI-specific packaging and memory optimization, it is moving decisively to reposition its product lines. The company is targeting increased HBM production and exploring a transition to glass interposers by 2028—an engineering bet aimed at winning back design wins from NVIDIA and AMD.

On the fabless side, NVIDIA remains financially dominant but vulnerable. Manufacturing delays in its Blackwell GPUs and a $5.5 billion write-down of Chinese inventory due to export rules have put pressure on its execution. AMD has seized the opportunity, rolling out a broad portfolio across PCs, AI processors, and GPUs, establishing itself as the preferred alternative for many customers wary of supply chain risk or price concentration.

Beyond silicon: The physical constraints of AI expansion

The AI revolution is also facing physical constraints that are increasingly defining the market. Data centers designed to house AI accelerators are now running into power grid limitations. In many regions, new data center construction is limited not by space, but by access to sufficient power infrastructure. Lead times for new transmission lines can exceed four years.

Cooling is another frontier. Traditional air-cooled systems are no longer sufficient to dissipate the heat generated by AI compute racks. Liquid cooling, once a specialty feature, is becoming the default for new builds. In some cases, immersion cooling is being adopted as the only viable solution. Energy is now a gating factor on semiconductor performance.

This constraint is already reshaping strategy. Companies like NVIDIA are investing in nuclear energy partnerships, with small modular reactors emerging as a promising solution for a reliable, scalable, and carbon-neutral energy supply tailored to high-density AI workloads.

ARTIFICIAL INTELLIGENCE

AI semiconductors: The talk of the hardware industry

The stakes of an industry in transition

The second quarter of 2025 represents an inflection point for the semiconductor industry. The AI boom has created a gravity well around which the rest of the industry is now forced to orbit. It is driving revenue, innovation, and investment, but also causing underinvestment in legacy sectors and amplifying geopolitical risk.

Traditional market indicators no longer tell the whole story. Growth is no longer broad-based, and supply chain resilience is now as critical as technological capability. Government policy has become a primary market mover, while physical infrastructure is now as decisive as transistor design.

As the industry enters the second half of the year, success will not be determined solely by product roadmaps or quarterly earnings. It will hinge on a company’s ability to navigate a volatile mix of AI ambition, geopolitical turbulence, and structural constraint. The future belongs to those who can move fast, but also build deep, stable foundations in a world that is anything but.

(ID:50459960)