MediaTek warns of increasing operational costs as AI demand strains chip supply

On February 4, Taiwan’s MediaTek issued a warning, stating that the growing demand for artificial intelligence (AI) is putting pressure on global chip supply chains, driving up costs and forcing the company to change its rates. According to the chip designer, the AI surge is increasing company confidence, but as demand continues to rise, supply limitations may worsen beyond 2026.

During a quarterly earnings conference call, MediaTek CEO Rick Tsai expressed great confidence in the company’s future but cautioned about supply chain issues. He claimed that the global supply chain is struggling to meet growing demand in 2026 due to AI, which is acting as a catalyst for industry expansion and driving a spike in demand, raising costs throughout the supply chain.

He also said that the company will adjust its pricing to reflect the growing supply chain costs and allocate supply across products based on the overall profitability.

Additionally, Tsai restated remarks made on the company’s most recent earnings call in October, stating that MediaTek anticipates generating billions of dollars in revenue by 2027 from its AI accelerator ASIC chips.

MediaTek reports strong revenue amid AI market surge

Notably, according to Channel NewsAsia (CNA), Taiwanese IT companies such as MediaTek and TSMC have reported increased sales amid the AI boom. On Wednesday, MediaTek reported fourth-quarter revenue of T$150.2 billion ($4.76 billion), up 8.8% from the same period last year. However, net income decreased 3.6% to T$23.1 billion.

MediaTek’s stock also increased. So far in 2026, the stock has risen by 26%, surpassing the benchmark index’s 11.5% increase. Prior to the release of its earnings on Wednesday, the stock ended the day up 0.3%.

Amid these strong financial results, Tsai revealed that the total addressable market (TAM) for data center ASIC chips is currently anticipated to be between $50 and $70 billion, which is $20 billion greater than the prior estimate.

MediaTek tackles AI chip shortages through strategic partnerships

MediaTek’s warning reflects a broader consensus across the semiconductor industry, where leading suppliers are indicating that AI-driven demand is surpassing the industry’s capacity-expansion capabilities.

Executives from TSMC, SK Hynix, Micron, Nvidia, Intel, and Samsung issued coordinated statements during the Q3 2025 earnings season, cautioning that demand for advanced logic nodes, high-bandwidth memory (HBM), and sophisticated packaging is expanding far more quickly than new capacity can be built. 

According to businesses, the present challenges are structural rather than transient and will have an impact on availability, lead times, and prices well into 2027.

Advanced packaging, particularly TSMC’s Chip-on-Wafer-on-Substrate (CoWoS) technology, has emerged as the most severe bottleneck, according to Fusion Worldwide.

CoWoS capacity is fully booked through 2025 and into 2026, according to TSMC executives, limiting output for Nvidia and AMD’s high-end AI accelerators. Even 3-nanometer wafers cannot be turned into working AI circuits without sufficient CoWoS capacity, which further increases supply pressures across the ecosystem.

Fusion also revealed that, according to major vendors, high-bandwidth memory has emerged as a crucial limitation. HBM capacity, including HBM3 and HBM3E, is sold out through 2026. According to Fusion, manufacturing complexity and lengthy validation cycles impede quick capacity development, while hyperscalers continue to lock in multi-year allocations.

Contract pricing is already rising as a result of these factors. Fusion noted that vendors predict double-digit increases for HBM in 2026.

Beyond supply constraints, MediaTek is leveraging strategic partnerships to capitalize on AI growth and expand its product capabilities.

On January 6, MediaTek announced a collaboration with Nvidia to design the GB10 Grace Blackwell Superchip for Nvidia Project DIGITS, a personal AI supercomputer. The partnership offers high-performance AI processing for R&D and educational applications by fusing Nvidia’s accelerated computing capabilities with MediaTek’s proficiency in Arm-based system-on-chip (SoC) technology.

Building on this collaboration, MediaTek revealed that the Dimensity Auto Cockpit processors, which combine next-generation GPU-accelerated AI with RTX graphics, are part of MediaTek’s ongoing work with Nvidia in automotive AI, which is expanded upon by the GB10 program. 

In addition to automotive applications, MediaTek has integrated Nvidia’s TAO AI model training toolkit into its NeuroPilot SDK to improve edge AI capabilities across IoT platforms and enable sophisticated AI applications in connected devices.

Want your project in front of crypto’s top minds? Feature it in our next industry report, where data meets impact.