Taiwan Semiconductor - Newsy Today

Alphabet has reportedly ordered 3 million Tensor Processing Units (TPUs) from Intel for delivery through 2028 to diversify its AI chip supply chain. This move shifts Google away from its exclusive reliance on Taiwan Semiconductor Manufacturing (TSMC) as the massive demand for artificial intelligence hardware threatens to outpace current manufacturing capacity.

Why is Alphabet shifting its chip manufacturing to Intel?

The decision to partner with Intel stems from the need to avoid production bottlenecks. While Google has collaborated with Intel on specialized processors for years, those chips were historically manufactured by TSMC. According to reporting from The Motley Fool, Google’s recent confidence in Intel follows months of testing Intel’s specific chip packaging technology to ensure it meets Google’s rigid engineering standards.

By adding Intel as a primary manufacturer, Alphabet gains a vital second source for its application-specific integrated circuits (ASICs). These chips are purpose-built for the specific matrix and vector-based mathematics required to run large-scale AI models. Relying on a single provider like TSMC creates a single point of failure in an era where AI compute demand is growing exponentially.

Did you know? An ASIC is a chip designed for one specific task. Unlike a general-purpose CPU in a laptop, Google’s TPUs are engineered solely to accelerate AI workloads, making them much more efficient for machine learning.

How do Google’s new TPU architectures improve AI performance?

Google is moving away from “all-purpose” AI chips in favor of specialized hardware. At the recent Cloud Next conference, Google introduced two distinct architectures: the TPU 8t and the TPU 8i.

Amin Vahdat, Google’s senior VP and chief technologist for AI and infrastructure, stated that the community benefits from chips specialized for individual tasks. The distinction is clear:

TPU 8t: Dedicated specifically to training workloads, where models learn from massive datasets.
TPU 8i: Designed for inference, which is the process of a trained model providing answers or predictions to users.

Vahdat noted that this specialization allows the company to run its most demanding AI workloads two to four times faster. Additionally, the company reports these specialized chips operate at a 30% lower cost than previous-generation TPUs.

Comparing TPU Architectures

Feature	TPU 8t	TPU 8i
Primary Use	Model Training	Model Inference
Optimization Goal	Computational Throughput	Serving Latency/Efficiency

What happens next for Google’s AI business model?

Alphabet is transitioning its TPU strategy from internal use to a commercial product. While Google historically used these chips only for its own services, executives recently announced they will sell TPUs to a select group of external customers. This shift is intended to significantly expand the company’s total addressable market.

Intel shares soar on reported in-house chip deal with Alphabet

This commercial expansion is already reflected in the company’s financial metrics. Google’s backlog has nearly doubled year over year, reaching a reported $460 billion. This surge suggests that enterprise demand for specialized AI hardware is outpacing the availability of general-purpose chips like those produced by Nvidia.

Pro Tip: When analyzing semiconductor trends, watch the “foundry” relationship. Companies like Google are increasingly acting as “fabless” designers, meaning they design the brains of the AI but rely on giants like Intel or TSMC to actually build the physical hardware.

Frequently Asked Questions

Why is Google using Intel instead of just TSMC?

Google is using Intel to diversify its supply chain. By not relying solely on TSMC, Google can avoid production delays and capacity shortages caused by the global AI boom.

What is the difference between training and inference?

Training is the process of teaching an AI model using massive amounts of data. Inference is the process of the AI actually working—such as when you ask a chatbot a question and it generates a response.

How much has Google’s AI backlog grown?

According to reported data, Google’s backlog has nearly doubled year over year to approximately $460 billion.

What do you think about Google’s move to Intel? Will this help them catch up in the AI race? Let us know in the comments below or subscribe to our newsletter for more deep dives into the semiconductor industry.

The Shift Toward Custom Silicon: Is the AI Hardware Gold Rush Evolving?

For years, Nvidia has been the undisputed face of the artificial intelligence revolution. With its share price climbing significantly over the past three years, the company’s graphics processing units (GPUs) became the gold standard for data centers worldwide. However, as the AI sector matures, a new trend is emerging: the rise of custom silicon.

While Nvidia remains a powerhouse, major tech players are increasingly turning toward application-specific integrated circuits (ASICs) to gain a competitive edge. This shift suggests that the future of AI hardware may not belong to a single entity, but rather to a diverse ecosystem of chip designers and manufacturers.

Why Custom Processors Are Gaining Traction

General-purpose GPUs have fueled the initial boom in AI, but they are not always the most efficient solution for every workload. Leading tech companies are discovering that custom semiconductors can be tuned to work more effectively with their specific AI models.

By designing chips tailored to their own unique architectures, companies can optimize performance and reduce operational costs. Industry data suggests that custom processors could significantly lower computation expenses compared to using standard GPU models. As the race to develop more powerful AI intensifies, this efficiency could be the key to long-term success.

Did you know?
Custom ASIC processors are projected to see faster growth this year compared to the general-purpose GPU market, signaling a fundamental shift in how hyperscalers approach infrastructure.

Key Players Shaping the AI Hardware Landscape

Several companies are positioning themselves to benefit from this demand for specialized hardware. Marvell and Broadcom have become essential partners for major hyperscalers looking to implement custom silicon solutions.

Broadcom: The company has seen its ASIC sales double in recent periods, driven by strong demand from major cloud providers. Broadcom continues to expand its work on custom designs for large-scale AI data centers.
Marvell: Known for its custom ASIC solutions, Marvell has become a key design partner for major tech firms, including Microsoft. Notably, the company collaborated on the design of the Maia 200 chip, aimed at improving the economics of AI token generation.
Taiwan Semiconductor (TSMC): As the premier manufacturer for these chip designers, TSMC holds a dominant position in the global processor market. With a significant market share in advanced AI processors, TSMC stands to benefit regardless of which chip designer leads the market.

these custom chips are generally intended to work alongside Nvidia’s GPUs, rather than replace them entirely. This collaborative approach ensures that the ecosystem remains robust, with Nvidia still playing a critical role in the broader infrastructure.

The “Megatrend” of AI Manufacturing

For investors and industry observers, TSMC serves as a bellwether for the health of the AI hardware sector. TSMC leadership has characterized AI as a “megatrend,” noting that the surge in demand for high-end processing power is driving substantial growth across the board.

Broadcom $AVGO Analysis: AI Custom Silicon, VMware Integration, and Q1 2026 Financial Strategy

With companies like Microsoft, Amazon, and Alphabet all investing in proprietary chip designs, the manufacturing capacity provided by TSMC has become a vital bottleneck and a massive opportunity. As long as the world’s leading AI firms continue to innovate, the demand for advanced manufacturing will likely remain a persistent force in the tech economy.

Frequently Asked Questions (FAQ)

Q: Why are tech companies moving away from general-purpose GPUs?
A: They aren’t necessarily moving away, but they are augmenting their infrastructure with custom silicon. Custom chips can be tuned for specific AI models, offering better efficiency and lower long-term costs.

Q: Is custom silicon replacing Nvidia’s technology?
A: No. In most cases, custom ASICs are designed to work in conjunction with existing GPU hardware to handle specific tasks more efficiently.

Q: Why is TSMC considered a key player in this trend?
A: TSMC is the primary manufacturer for many of the world’s leading chip designers. Because they produce the hardware for various competitors, they are positioned to benefit from the growth of the AI industry as a whole.

Pro Tip:
When evaluating the AI hardware space, look beyond the headline-grabbing chip designers and consider the entire supply chain, including the companies responsible for the manufacturing and interconnect technologies that make these systems possible.

What are your thoughts on the transition toward custom AI silicon? Do you believe this will eventually challenge the dominance of general-purpose GPUs? Let us know your take in the comments section below, or subscribe to our newsletter for more deep dives into the future of tech.