Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problemsSome links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
TSMC’s “endless” AI-demand remark was not a promise of unlimited growth. Chairman and CEO C.C. Wei used the phrase during the company’s January 15, 2026 earnings call to describe a potentially multiyear need for computing power. He also cautioned that he could not guarantee several consecutive years of strong semiconductor-industry growth.
The bullish case has since strengthened. TSMC’s July 2026 results showed record quarterly profit and revenue, higher capital-spending plans, and a sharply improved full-year outlook. But the evidence supports a more precise conclusion: AI-chip demand is currently broad, strong, and supply-constrained—not permanently immune to a downturn.
What TSMC actually meant by “endless”
Wei’s comment came during TSMC’s Q4 2025 earnings call, held on January 15, 2026. “Endless” was an informal description of the long-term AI trend, not a formal financial forecast or guarantee that TSMC’s revenue will rise forever.
Free tools Windows power users keep installed
One-click scans. No signup required.
Wei’s point was that increasingly capable AI models, inference workloads, reasoning systems, and agentic applications will continue to require substantial computing power. However, he also said he did not know whether the semiconductor industry could sustain strong growth for three, four, or five consecutive years.
#1 Best Overall
- System Compatibility Note: This 2-slot card measures 271 x 112 x 39 mm and requires a single 12V-2x6-pin power connector. Please verify chassis and PSU compatibility before purchase.
- Dedicated Support: Please contact us directly through Amazon for any product questions or assistance you may require.
- Professional Intel Arc Pro B70 GPU: Built on the Intel Xe2-HPG architecture, it features 32 Xe cores and 256 XMX engines, designed to accelerate AI, rendering, and complex visualization workloads.
- Massive 32GB GDDR6 VRAM: Equipped with 32GB of high-speed GDDR6 memory on a 256-bit bus, running at 19 Gbps, which allows for handling large AI models and complex datasets locally.
- High-Performance Engine Clock: Delivers an engine clock of 2540 MHz, providing the compute power needed for demanding professional applications and AI inference.
That distinction matters. TSMC sees orders, capacity requests, and customer forecasts moving through its manufacturing business. It does not have perfect visibility into whether every AI application will be profitable, whether hyperscalers will maintain their spending plans, or whether end users will generate enough revenue to justify the infrastructure investment.
TSMC’s Q4 2025 earnings-call transcript provides the primary context for the remark.
Why Q4 2025 was important
The January call showed that AI was becoming a larger and more durable part of TSMC’s business, rather than merely producing a short-term inventory rebound.
- AI accelerators represented a high-teens percentage of TSMC’s 2025 revenue.
- TSMC raised its expected AI-accelerator revenue growth to a mid-to-high-50s compound annual growth rate for 2024 through 2029.
- The company projected overall long-term revenue growth approaching a 25% CAGR for the five-year period beginning in 2024.
- At the time, TSMC expected 2026 revenue to increase by close to 30% in U.S.-dollar terms.
TSMC’s AI-accelerator category includes AI GPUs, custom AI application-specific integrated circuits, and high-bandwidth-memory controllers used in data-center training and inference. It is therefore broader than sales of branded GPUs alone.
Rank #2
- NVIDIA Volta GV100 Architecture — 4,608 CUDA Cores, 640 1st-Gen Tensor Cores delivering 14 TFLOPS FP32 and 112 TFLOPS deep learning performance for AI training, inference, HPC, and scientific computing workloads
- 32GB HBM2 ECC Memory — 900 GB/s Bandwidth — High-bandwidth memory on a 4096-bit bus with ECC error correction provides the memory capacity and throughput required for the largest AI models, simulations, and datasets
- PCIe 3.0 x16 Interface — 250W TDP — Standard PCIe Gen3 connectivity with passive cooling designed for enterprise rack server deployment in HPE ProLiant, Dell PowerEdge, and Supermicro platforms with adequate chassis airflow
- NVLink — Scale to 96GB Unified Memory — Connect two V100 GPUs via NVLink at 300 GB/s bi-directional bandwidth to scale GPU memory from 32GB to 96GB for larger AI training and HPC workloads
- Multi-Precision Computing — Supports FP64 (7 TFLOPS), FP32 (14 TFLOPS), FP16 (112 TFLOPS) and INT8 precision modes for flexible deployment across training, inference, and scientific simulation workloads
Management also said all four of its major growth platforms—smartphones, high-performance computing, the Internet of Things, and automotive—would contribute to growth. AI accelerators were expected to be the largest incremental contributor, but AI was not the only source of expansion.
Why TSMC is an important AI-demand signal
TSMC is a leading contract chip manufacturer. It fabricates advanced processors designed by companies including Nvidia, AMD, Apple, and other chip designers, as well as custom chips developed for large technology customers. It also provides advanced process technology and packaging, both of which are crucial to high-end AI accelerators.
This gives TSMC a wider view of the semiconductor cycle than a single chip company. Demand can arrive through Nvidia GPUs, custom hyperscaler ASICs, smartphone processors, and other advanced designs. TSMC is not a perfect proxy for any one customer, and it does not manufacture every product sold by the companies mentioned above. But its results provide meaningful evidence about orders for leading-edge manufacturing and packaging capacity.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →The limitation is equally important: strong foundry demand does not prove that every AI application is commercially successful. It shows that chip designers and their customers are willing to commit capital and reserve manufacturing capacity.
Rank #3
- Professional AI & Creator Workstation: AMD Radeon AI PRO R9700 GPU with 32GB GDDR6 is engineered for AI development, professional content creation, and compute-intensive workloads.
- Massive 32GB Memory Capacity: 32GB of GDDR6 memory on a 256-bit bus provides ample bandwidth for large AI models, 8K video editing, and complex 3D rendering.
- Advanced RDNA 4 with AI Accelerators: 64 Compute Units with 3rd Gen Ray Tracing and dedicated 2nd Gen AI Accelerators for groundbreaking AI performance and visual computing.
- Professional Blower Cooling: Efficient single blower design exhausts heat directly out of the chassis, ideal for multi-GPU workstation and server configurations.
- Enterprise-Grade Thermal Solution: Vapor chamber heatsink with industrial Honeywell PTM7950 thermal interface material ensures reliable cooling under sustained professional loads.
The immediate problem is not a lack of demand—it is capacity
TSMC said AI-related capacity remained very tight. Customers were engaging with the company two to three years in advance because advanced-node production and packaging require long planning cycles.
TSMC was trying to increase output through productivity improvements, process optimization, and reallocating capacity between manufacturing nodes. Those measures can improve near-term supply, but they cannot replace new fabs indefinitely.
The January call included 2026 capital-expenditure guidance of $52 billion to $56 billion. TSMC said that spending would add little supply during 2026 and only limited capacity in 2027. New fabs typically take two to three years to build, so more meaningful capacity relief was expected around 2028 and 2029 if the AI trend continued as management expected.
This creates the central tension in the story: customers want more AI chips now, while the factories and packaging capacity needed to make them arrive years later.
Rank #4
- System Compatibility Note: 2-slot card, 271x112x39mm, single 8-pin power, 200W TDP. Verify chassis clearance and PSU capacity before purchase.
- Dedicated Support: Please contact us directly through Amazon for any product questions or assistance you may require.
- 24GB GDDR6 on 192-Bit Bus: Massive 24GB memory with 456 GB/s bandwidth – ideal for LLMs, AI inference, 3D rendering, and generative design.
- Intel Xe2-HPG Architecture: Built on Intel's next-gen architecture with 20 Xe cores and 160 XMX engines for AI acceleration (197 INT8 TOPS).
- PCIe 5.0 Support: PCI Express 5.0 x16 interface for maximum bandwidth with the latest workstation platforms.
What happened after the Q4 call?
July 16, 2026: TSMC’s Q2 2026 results provided a stronger near-term test of the January outlook. The company reported record net profit of NT$706.6 billion, up 77% year over year, and revenue of NT$1.27 trillion, up 36%.
TSMC also raised its 2026 capital-expenditure range to $60 billion–$64 billion and lifted its full-year revenue-growth forecast from above 30% to slightly above 40% year over year. It announced another $100 billion of planned U.S. investment, bringing its stated Arizona investment total to $265 billion. The additional buildout was expected to include four more fabs, including facilities targeting 2-nanometer and more advanced technologies.
The Associated Press reported the Q2 results and additional Arizona investment, while the Q2 earnings-call transcript provides further management context.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchAs of August 18, 2026, these developments made Wei’s January assessment look directionally prescient. They did not convert “endless” into a guarantee. The July figures are realized results and updated guidance; the 2029 horizon remains a management forecast.
Best Value
- PLEASE NOTE: Exporting an NVIDIA RTX Pro 6000 GPU outside the US requires strict adherence to the U.S. Export Administration Regulations (EAR) and issuance of an export license from the Bureau of Industry and Security (BIS). Compliance and Know Your Customer (KYC) screening may be required as a condition of order acceptance. [NVIDIA Blackwell Streaming Multiprocessor] The new SM features increased processing throughput, and new neural shaders that integrate neural networks inside of programmable shaders | DLSS 4: Multi Frame Generation ensures ultra-smooth frame pacing for lifelike simulations.
- [Double-Flow-Through Design] The RTX PRO 6000 Blackwell features a double-flow-through cooling design, optimizing efficiency and airflow to sustain peak performance under 600W power loads. | [5th Gen Tensor Cores] Deliver up to 3X the performance of the previous generation and support for FP4 precision for faster AI model processing times with reduced memory usage, enabling local fine-tuning of LLMs and generative AI | [4th Gen Ray Tracing Cores] Double the ray-triangle intersection rate of the previous generation to create photoreal, physically accurate scenes and immersive 3D designs with RTX Mega Geometry, which enables up to 100X more ray-traced triangles.
- [PCIe Gen 5] Support for PCIe Gen 5 provides double the bandwidth of PCIe Gen 4, improving data-transfer speeds from CPU memory and unlocking faster performance for data-intensive tasks like AI, data science, and 3D modeling. | [GDDR7 Memory] With 96 GB of GPU memory and 1.8 TB ps bandwidth, it can tackle massive 3D and AI projects, fine-tune AI models locally, explore large-scale VR environments, and drive larger multi-app workflows.
- [DisplayPort 2.1] Achieve unparalleled visual clarity and performance, driving high resolution displays at up to 8K at 240 Hz and 16K at 60 Hz. Increased bandwidth enables seamless multi-monitor setups while HDR and higher color depth support ensures superior color accuracy for precision work, such as video editing, 3D design, and live broadcasting.
- [Universal MIG] Divide a single RTX PRO 6000 Blackwell into multiple isolated instances, each with dedicated resources, allowing for concurrent execution of multiple workloads, optimized GPU utilization, and secure isolation of different applications or users. [WARRANTY] 3 YR Manufacturer's Warranty. Bulk OEM Packaging. Retail Packaging is NOT included.
Does this disprove the AI-bubble argument?
No. TSMC’s results are strong evidence that current demand for AI infrastructure is real. They are not proof that every dollar being spent on AI will earn an attractive return.
Evidence supporting the demand thesis
- Customer visibility: Leading-edge customers were discussing capacity two to three years ahead.
- Supply tightness: TSMC described AI-related capacity as very tight rather than readily available.
- Demand breadth: The opportunity includes GPUs, custom ASICs, and other accelerator designs from multiple customers.
- Capital-spending follow-through: TSMC raised its capex plan and accelerated expansion rather than simply benefiting from a one-quarter restocking cycle.
- Subsequent performance: Q2 2026 revenue, profit, capex, and full-year guidance all exceeded the more cautious January outlook.
Why caution remains necessary
- Hyperscalers could slow or defer capital spending if AI-service revenue does not justify infrastructure costs.
- Customers can over-order, revise forecasts, or delay capacity commitments even after initially booking space.
- Power availability, data-center construction, networking, memory, advanced packaging, engineering labor, and permitting can constrain deployments independently of wafer demand.
- A more efficient model architecture could reduce compute required for some workloads, even if total AI use continues rising.
- Export controls, tariffs, geopolitical tensions, and logistics disruptions could affect shipments or investment plans.
- Samsung or Intel could improve leading-edge yields and win meaningful business from TSMC.
- Large fabs are expensive and slow to repurpose. If demand weakens before 2028 or 2029, new capacity could contribute to overcapacity rather than solve a shortage.
A strong foundry cycle can still be cyclical. “Endless” is not the same as risk-free.
What the capacity race means for TSMC and chip designers
For TSMC, tight capacity can support utilization, pricing power, and strong financial performance. But expansion also increases capital intensity. A higher capex figure demonstrates management’s confidence and customer commitments; it does not guarantee that each new factory will earn an attractive return.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
For chip designers, access to leading-edge wafers and advanced packaging may be as important as chip architecture. Companies that can reserve capacity early may have an advantage when demand exceeds supply. TSMC’s Arizona expansion could also help U.S.-based customers diversify manufacturing geographically, although those fabs will take time to build and ramp.
The move toward custom AI ASICs could broaden TSMC’s customer base beyond Nvidia. At the same time, dependence on a relatively small group of large technology customers remains a concentration risk.
What investors should watch next
- Booked capacity versus realized shipments: Reservations are useful, but revenue, utilization, and completed shipments provide stronger confirmation.
- Demand breadth: Watch whether growth comes from multiple chip designers and workloads or remains concentrated among a few hyperscalers.
- Pricing and profitability: Demand is more durable when it produces healthy margins and returns, not simply more units.
- Advanced packaging and memory: A shortage in packaging or high-bandwidth memory can limit AI-system output even when wafer capacity expands.
- Capital-spending discipline: Rising capex supports the growth thesis but increases downside if customers reduce orders.
- End-market economics: Continued investment ultimately depends on whether AI services generate enough value to support the cost of compute, power, and data centers.
Timeline: from “endless” demand to a larger buildout
| Date | Development |
|---|---|
| January 15, 2026 | TSMC reports Q4 2025 results. C.C. Wei describes the AI trend as “endless” while warning that sustained semiconductor growth cannot be guaranteed. |
| January 2026 outlook | TSMC forecasts close to 30% 2026 revenue growth and $52 billion–$56 billion of 2026 capex. |
| July 16, 2026 | TSMC reports record Q2 profit and revenue, raises capex to $60 billion–$64 billion, and lifts its 2026 growth forecast to slightly above 40%. |
| July 2026 U.S. expansion | TSMC announces another $100 billion of planned U.S. investment, bringing its stated Arizona total to $265 billion. |
| August 18, 2026 | The latest available evidence supports a powerful multiyear AI-capacity cycle, but not permanent or unlimited demand. |
Bottom line for investors
TSMC has stronger evidence than most companies that AI demand is currently real, broad, and difficult to satisfy. The company is seeing long-lead-time customer commitments, tight leading-edge capacity, record results, and enough confidence to increase its investment plans.
But the most accurate translation of “endless” is potentially long-lasting, not unlimited. The near-term question is whether TSMC and its customers can build enough advanced manufacturing and packaging capacity. The long-term question is whether AI businesses can turn that compute into sufficient economic returns to sustain spending through the next semiconductor cycle.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
