What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Jensen Huang’s explanation for preferring Taiwan Semiconductor Manufacturing Co. (TSMC) over Samsung for leading-edge NVIDIA production comes down to five connected capabilities: system-level factory coordination, sustained technology focus, customer service, advanced process execution and trust. Huang did not present a numbered list using the exact phrase “five reasons”; the five-part breakdown separates the two cultural strengths he described.
His comments describe an advantage in running a global foundry system, not proof that Samsung lacks important technology or that NVIDIA has stopped working with it. Samsung remains an NVIDIA memory partner, and the companies are collaborating on a large AI-factory project.
What Huang says TSMC does better
Huang called TSMC’s manufacturing system “completely miraculous.” His point was that advanced chip production depends on coordinating an enormous, changing network of customers and factories, not simply owning a smaller transistor geometry.
1. A system that orchestrates hundreds of customers
TSMC can respond to the “dynamic demands of hundreds of companies,” according to Huang. That includes changing wafer starts, stopping or restarting production, handling emergencies and adjusting schedules as customers’ forecasts move.
This is an operating system for manufacturing capacity. A foundry must decide which designs run on which tools, when a lot should be paused, how scarce equipment is allocated and how a late change affects every downstream process. Huang’s praise implies that TSMC has institutionalized those decisions across its network rather than treating each customer as an isolated project.
2. Technology is a continuous operating priority
Huang described TSMC as advancing process technology while serving customers at the same time. The advantage is cultural as much as technical: engineering teams keep pushing the next process generation without losing the discipline required to deliver today’s products.
For NVIDIA, that matters because a GPU program can span several process generations and involve long qualification cycles. A supplier that can move from current production to a new node while maintaining customer coordination reduces the risk that a design transition becomes a manufacturing bottleneck.
Rank #2
3. Customer service is treated as a core capability
Huang said TSMC balances customer service with bleeding-edge technology, a combination he said many companies fail to achieve together. Customer service in this context means more than account management. It includes technical support, schedule changes, yield learning, escalation during defects and the ability to adapt factory priorities when a customer’s product mix changes.
That balance is particularly valuable for complex accelerators. NVIDIA needs a foundry to work through design-for-manufacturing issues, packaging constraints and production ramps while protecting confidential information and delivery commitments.
4. Leading-edge process execution is unusually difficult
Huang separated TSMC’s technology capability from its broader culture and emphasized how hard advanced manufacturing is. “Building advanced chip manufacturing is extremely hard,” he said, because it requires “the engineering, the science, and the artistry of doing what TSMC does for a living.”
The wording highlights why a process node cannot be judged by a marketing label alone. Real production requires thousands of tightly controlled steps, specialized equipment, process recipes, defect monitoring, packaging coordination and engineering feedback. A company may announce a node yet still face a long path to dependable, high-volume output for a particular chip.
Free tools Windows power users keep installed
One-click scans. No signup required.
5. Trust lowers the risk of putting a major product on another company’s factories
Huang called trust an intangible capability and said, “I trust them to put my company on top of them. That’s a very big deal.” The statement covers execution history, confidentiality, predictable escalation and confidence that the foundry will protect a customer’s schedule when capacity is constrained.
Trust is cumulative. It is built through repeated production ramps and difficult recoveries, then becomes part of a customer’s decision about where to place its most valuable designs. It is also why a foundry relationship can persist even when another manufacturer advertises a comparable process specification.
Rank #4
TSMC and Samsung: what the evidence supports
Huang’s remarks support a TSMC advantage in foundry-system execution from NVIDIA’s perspective. They do not establish a current, apples-to-apples yield ranking for the same process node, and no such numerical comparison should be inferred from his comments.
| Comparison area | What is established | What is not established here |
|---|---|---|
| Manufacturing orchestration | Huang specifically praised TSMC’s ability to coordinate changing demand across hundreds of customers. | A published numerical score comparing TSMC and Samsung’s factory systems. |
| Process technology | Huang described TSMC’s advanced manufacturing as requiring engineering, science and artistry. | A current same-node yield comparison or universal performance lead. |
| Customer service | Huang said TSMC combines customer service with bleeding-edge technology. | A standardized customer-support metric for either company. |
| Trust and execution | Huang said he trusts TSMC with NVIDIA’s products. | Proof that Samsung cannot meet NVIDIA’s requirements. |
| HBM and related relationships | NVIDIA said in a 2024 Q&A that it works with SK hynix, Micron and Samsung as its three HBM partners, with qualification and integration continuing. | Evidence that NVIDIA has ended its Samsung relationship. |
| Geographic resilience | NVIDIA and TSMC announced the first production Blackwell wafer at TSMC’s Arizona Fab 21 using TSMC’s 4N process. | A complete, directly comparable resilience assessment of Samsung’s footprint. |
Samsung still matters to NVIDIA
Samsung remains an active NVIDIA partner in high-bandwidth memory (HBM). NVIDIA’s 2024 Q&A identified Samsung alongside SK hynix and Micron and said qualification and integration work was continuing. That relationship concerns memory, while Huang’s five-part explanation focuses mainly on the foundry system used to manufacture logic chips.
NVIDIA later announced an AI-factory collaboration with Samsung involving more than 50,000 NVIDIA GPUs. The project is designed to use CUDA-accelerated lithography and simulation, digital twins, predictive maintenance and factory automation. This is evidence of cooperation, not an announcement that one supplier has replaced the other.
Best Value
Why TSMC’s scale changes the decision
The Associated Press reported that TSMC produces more than 90% of the world’s most cutting-edge chips and supplies companies including NVIDIA and Apple. That scale gives TSMC a dense base of process learning and customer demand, although scale alone does not guarantee every product ramp will succeed.
TSMC has stated a 2026 capital-expenditure range of $60 billion to $64 billion. TSMC chairman and chief executive C.C. Wei said AI demand remained above the company’s current supply capability, indicating that even the leading supplier is operating under capacity pressure.
Capacity pressure makes Huang’s first and fifth reasons more consequential. When demand exceeds available tools, the supplier’s scheduling system and the customer’s confidence in allocation decisions can matter as much as the nominal node.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Blackwell production and the move beyond Taiwan
The first production Blackwell wafer made at TSMC’s Arizona Fab 21 used the company’s 4N process. Later phases are planned for TSMC’s N3, N2, A16 and A14 technologies. The Arizona production demonstrates that TSMC’s operating model is being extended beyond its Taiwanese base, while the phased roadmap shows that geographic expansion does not eliminate the complexity of qualifying each new process and site.
What “beating Samsung” means in practice
For an NVIDIA chip, “better” is not a single specification. It means the foundry can deliver the required process, packaging and volume while responding to design changes and protecting the schedule. Huang’s argument is that TSMC is unusually strong at combining those jobs.
- Choose TSMC when coordinated capacity, rapid schedule changes and leading-edge process execution are the central risks.
- Do not interpret the comments as a claim that Samsung has no advanced manufacturing or memory role.
- Do not use Huang’s remarks as a substitute for product-specific yield, cost or delivery data; those figures depend on the exact node, design, package and production period.
The practical conclusion is narrower and more useful than a blanket declaration that one company is superior: Huang trusts TSMC’s integrated foundry system to manage NVIDIA’s most demanding production requirements, while NVIDIA continues to use Samsung in important memory and factory collaborations.
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.
