Tag: TSMC

Guest Post: How AI Has Transformed Semiconductor Scheduling at NVIDIA and TSMC

Professor Misty Blessley raises an interesting AI issue-chip plant scheduling.

Semiconductors have been called the brains of the modern world. Switching between conducting and blocking electricity, they are essentially “on/off” switches that control the flow of power, and this unique characteristic makes them the building blocks of modern technology. They are behind computing, communications, and energy innovation, and our information dependent society is always hungry for more. Leading chip makers, NVIDIA and TSMC, both use an AI to advance semiconductor design, scheduling, and manufacturing, report ElectronicsUSA (June 4, 2026) and NVIDIA News (May 31, 2026).

TSMC ( Taiwan Semiconductor Manufacturing Company), the world’s largest semiconductor foundry, makes the most advanced chips on the planet. NVIDIA is a global leader in accelerated computing and AI. For decades, TSMC has manufactured NVIDIA’s chips, and this partnership has come full circle. TSMC uses NVIDIA’s AI technologies inside its fabrication plants (“fabs”), which in the semiconductor industry refers to highly specialized facilities where silicon wafers are processed into microchips.

NVIDIA’s AI models are now embedded directly into TSMC’s manufacturing workflow, transforming production scheduling. A single wafer, a subcomponent of a semiconductor, may require hundreds of tools and thousands of tightly sequenced steps. TSMC can now evaluate millions of scheduling combinations in seconds. This optimizes job sequencing, and instantly rebalancing schedules when tools go down or urgent orders arrive. The result is smoother production flows, less idle time, and higher overall fab productivity. AI comes as close as ever to “running the plant.”

AI driven scheduling is also advancing through emerging capabilities such as predictive dispatching, where models forecast bottlenecks hours ahead and reroute wafers to prevent delays. Another is crossfab load balancing, which evaluates capacity across multiple TSMC sites and shifts work to maximize throughput. Both approaches are expected to reduce fab cycle time 5–10%.

NVIDIA’s AI scheduling helped TSMC cut critical chip production workloads 20-50%. This addresses society’s insatiable demand for semiconductors. It also fuels the next wave of breakthroughs in computing, communications, and energy innovation, the domains powered by these tiny on/off switches.

Classroom discussion questions:

  1. What are the strategic and operational risks if a company relies too heavily on AI to automatically reroute work without human oversight?
  2. Discuss how AI’s ability to evaluate millions of combinations in seconds changes a manager’s approach to bottleneck scheduling compared to manual or sequential sequencing rules (like SPT or FCFS).

OM in the News: The Chips That Run the World

Taiwan Semiconductor Manufacturing Co.’s chips are everywhere, though most consumers don’t know it. The company makes almost all of the world’s most sophisticated chips, and many of the simpler ones, too. They’re in billions of products with built-in electronics, including iPhones, PCs and cars—all without any sign they came from TSMC. TSMC has emerged as the world’s most important semiconductor company, with enormous influence over the global economy, writes The Wall Street Journal (June 19-20, 2021). With a market cap of around $550 billion, it is the world’s 11th most valuable company. But its dominance leaves the world in a vulnerable position. As more technologies require chips of mind-boggling complexity, more are coming from this one company.

tsmc2
A TSMC employee with a 8-inch wafer.

It will be difficult for other manufacturers to catch up in an industry that requires hefty capital investments. And TSMC can’t make enough chips to satisfy everyone—a fact that has become even clearer amid a global shortage, adding to the chaos of supply bottlenecks, higher prices for consumers and furloughed workers. The situation is similar in some ways to the world’s past reliance on Middle Eastern oil.

It now makes around 92% of the world’s most sophisticated chips, which have transistors that are less than 1/1,000 the width of a human hair. (Samsung makes the rest). Most of the 1.4 billion smartphone processors world-wide are made by TSMC. It also makes 60% of the less-sophisticated microcontrollers that car makers need. As carmaker after carmaker has idled production because of chip shortages, TSMC has little incentive to reallocate production. The less lucrative auto chips make up only around 4% of its revenues.

TSMC’s hard-driving culture and deep pockets make it hard to create a more diversified semiconductor supply chain. Once a chip producer falls behind, it’s hard to catch up. Companies can spend billions of dollars and years trying, only to see the technological horizon recede further. A single semiconductor factory can cost $20 billion. One key manufacturing tool for advanced chip-making that imprints intricate circuit patterns on silicon costs $100 million. TSMC’s own expansion plans call for spending $100 billion over the next 3 years. 

Classroom discussion questions:

  1. Why are chips critical to the US economy?
  2. What will need to be done for the US to be “semiconductor independent?”

OM in the News: Auto Makers and the Global Chip Shortage

Auto manufacturers have spent decades streamlining their supply chains, using their muscle over parts makers to reduce their own costs by carrying little inventory and relying on suppliers to deliver components “just in time.” But car makers are finding they can’t dictate terms in the same way to the chip industry with its far-broader customer base, particularly now that chip demand is booming globally among a swath of industries. Car makers blame the shortage on tier-one parts suppliers, which generally do most of the chip buying. The shortage will drive the global auto industry to produce nearly 700,000 fewer cars than planned for the first three months of 2021. Auto makers including VW, Ford and GM, have furloughed tens of thousands of workers.

Ford has slashed production of America’s top-selling vehicle, the F-150 pickup, because of chip shortages

The car-chip pain is partly a self-inflicted wound that traces to the pandemic’s early days, writes The Wall Street Journal (Feb. 13-14, 2021). When the global economy went into stasis, preparations for future car production halted. Auto-parts suppliers reduced orders for electronics, betting that large volumes wouldn’t be needed well into the future.

Chip makers chose not to stockpile parts and wait for car makers’ orders—a decision made easier by surging demand from sales of laptops, servers, smartphones, video game consoles, 5G networks, and other electronics. The harsh reality is that these other products have have a much, much higher rate of return than making chips for cars. Plus only one chipmaker, Taiwanese TSMC, produces about 70% of the units used in the world’s autos.

Still, the car industry largely operated as if electronics suppliers were at its mercy. Normally, when they are calling on their suppliers, everyone is excited about the large volumes. But it must be compared to the billions of smartphones and PCs that are being sold. Unlike with some other parts, said a chip maker exec, referring to the auto makers, “They don’t know how the sausage gets made at the bottom.”

Classroom discussion questions:
1. What is meant by the quote at the end of the blog?

2. Table 11.3 (page 450) in your Heizer/Render/Munson text lists 10 supply chain risks. Which, if any, apply in this case?