In 2001, there was so much excess component inventory in the supply chain that more than $13 billion worth of semiconductors alone was written down or off by component makers, distributors and customers. People still wince at the memory.
In 2021, there’s such a severe shortage of semiconductors that the automotive industry is set to lose more than $100 billion in sales because it can’t get chips. “Desperation” is a term that’s been kicked around.
For two decades the electronics supply chain has been minimizing the level of inventory in its pipeline via just-in-time (JIT), build-to-order (BTO) and lean. The goal is delivering the “right amount” of components to manufacturing lines just as they are needed. These parameters are determined by end-customers, largely OEMs and EMS providers.
Forecasting in the electronics industry is notoriously bad.
“The first time I get an accurate customer forecast will be the only time I get an accurate customer forecast,” one chip maker recently said. When demand for electronics skyrocketed at the beginning of the year, no one was prepared. The current chip shortage is expected to last well into 2022.
One of the constants in the electronics industry is components take between 16 and 18 weeks to manufacture. So, lead times -- the span between placing a component order and receiving it – are an indicator of how well supply is meeting demand.
Currently, lead times for a basket of electronic components -- both active and passive -- average 22 weeks, according to market intelligence firm LevaData, and for some products are stabilizing. “After experiencing a dramatic 2x to 3x increase in average semiconductor lead times over the last eight to 12 months, it appears that the situation may be shifting again. The delays are shortening or, in some cases, beginning to flatten at the least.”
In its most recent report, ECIA also noted significant easing on product lead times across all categories from July to August. Discrete semiconductors and capacitors remain under the most pressure. Some memory products are quoting lead times up to a year.
There are a number of ways to interpret this trend:
- Supply/demand are beginning to moderate and trend back toward normal
- Customers, realizing they can’t get devices when they need them, are placing orders farther out
- Manufacturers have adjusted their forecasts downward due to constrained supply
- Component makers are producing more parts
“For the first time in months there’s a several-month streak of lead times either flattening or declining – especially in those categories that initially led the lead time increases,” LevaData reported. “While supply constraints such as seasonality or new Covid-19 labor shortages are undoubtedly still present, the impact of these factors varies. They’re worse at various nodes in the supply chain and affecting specific components within a group more than others. Most component manufacturers' facilities are near or at 100 percent capacity.”
Another possibility, the firm noted, is less hoarding. Many companies’ cash flow was greatly affected by having to tie up more cash than usual to establish sufficient component supplies for enabling continuity of product supply to their customers. This change in lead time momentum might be pointing toward brighter days for companies, where “extreme” excess inventory is no longer needed.
There are guardrails against this practice since hoarding only makes a shortage situation worse. Component makers and distributors monitor customers’ historic buying patterns and flag unusually high orders. They also work with customers on finding alternative devices for their designs.
Stabilizing lead times could mean the semiconductor market is beginning to shift back toward more balance between supply and demand, LevaData noted. However, unforeseen volatility factors, like Covid-19 shutdowns, will continue to impact that trend over the next three to six months.
Covid-19 has exposed some fundamental weaknesses in the supply chain, according to the McKinsey Global Institute and other consultancies.
“The shift to just-in-time and lean production systems has helped companies improve efficiency and reduce their need for working capital,” reported McKinsey. “But now they may need to strike a different balance between just-in-time and ‘just in case.’ Having sufficient backup inventory of key parts and safety stock is a critical buffer that can minimize the financial impact of disrupted supplies. It can also position companies to meet sudden spikes in demand.”
Don’t count on an increase in ready-to-ship inventory. Component makers don’t want to store it; publicly traded distributors are hammered if stock levels increase; and customers have become accustomed to JIT, BTO and lean.
It’s fair to count customer forecasts as a volatility factor in the supply chain. The degree to which forecasts were wrong in 2001 and 2021 is remarkable. The supply chain reinvented itself after 2001. The chip industry plans to build roughly 29 fabs over the next few years.
Overall demand for components isn’t going to decrease, but if near-term demand for 5G, EVs, AI or any other high-growth industry misses the mark we could be looking at excess by 2023.