This week, reports from multiple automotive manufacturers confirmed what many of us in the embedded and IoT space have been feeling for months: there simply aren’t enough chips to go around. Ford, Toyota, and Volkswagen have all announced production cuts due to semiconductor shortages, and the ripple effects are spreading far beyond cars. If you’re building anything that depends on microcontrollers, sensors, or custom silicon, you’re likely already feeling the squeeze — or you will soon.
How We Got Here#
The current shortage has been building since mid-2020, driven by a perfect storm of factors. When COVID-19 hit, automotive manufacturers slashed their chip orders anticipating a prolonged slump. Meanwhile, demand for consumer electronics — laptops, gaming consoles, networking equipment — surged as the world shifted to remote work and entertainment. Chip fabricators, particularly TSMC and Samsung, reallocated capacity to serve the booming consumer and data center markets.
Now that automotive demand has bounced back faster than expected, those manufacturers find themselves at the back of the queue. But this isn’t just a car problem. The shortage spans multiple process nodes and chip types. Reports from semiconductor industry analysts indicate lead times for some components have stretched to 26 weeks or more — double the normal timeframe.
The geographic concentration of fabrication capacity makes this worse. TSMC alone produces roughly 54% of the world’s contract-manufactured chips, and the vast majority of advanced node production (7nm and below) happens in Taiwan and South Korea. A single facility disruption — or a geopolitical incident — could turn a shortage into a crisis.
What This Means for IoT and Embedded Development#
For those of us building IoT systems, the implications are immediate and practical. Popular microcontrollers from STMicroelectronics, NXP, and even the ubiquitous ESP32 from Espressif are seeing extended lead times. I’ve been working on a sensor project that relies on an STM32L4 series chip, and my usual distributor is showing backorders stretching into Q3.
This forces some uncomfortable decisions. Do you redesign around a different MCU that happens to be available? Do you stockpile components, tying up capital and warehouse space? Do you delay product launches? None of these are great options, especially for smaller teams and startups that can’t throw purchasing power at the problem.
The Raspberry Pi Foundation has acknowledged supply constraints as well. While the Pi itself uses a Broadcom SoC that’s somewhat insulated from the broader shortage, the ecosystem of HATs, sensors, and peripheral components that make Pi projects useful is definitely affected.
The Deeper Structural Problem#
What bothers me about this situation is that it exposes a vulnerability we’ve been ignoring for years. The global semiconductor supply chain is extraordinarily concentrated and optimized for efficiency, not resilience. Just-in-time manufacturing works beautifully until it doesn’t — and when it breaks, there’s no buffer.
Building a new fabrication facility takes 2-3 years and costs $10-20 billion. You can’t spin up chip production the way you can spin up cloud instances. This physical reality means the current shortage will take time to resolve, regardless of how much money gets thrown at it. Intel’s new CEO, Pat Gelsinger (who just took the helm on January 13th), has signaled that revitalizing Intel’s foundry capabilities is a top priority. But even Intel’s manufacturing ambitions won’t produce chips tomorrow.
There’s also a design complexity angle worth noting. Modern chips are increasingly specialized — AI accelerators, 5G modems, automotive safety controllers — which means you can’t easily substitute one for another. The days when a shortage of one chip could be solved by swapping in a pin-compatible alternative are largely behind us.
Practical Steps for Engineering Teams#
If you’re leading a hardware or IoT project right now, here’s what I’d recommend based on what I’m seeing:
Diversify your BOM. If your design is locked to a single-source component, start evaluating alternatives now. Even if you don’t need them today, having a validated second source could save your project in six months.
Extend your planning horizon. If you normally order components 8-12 weeks out, push that to 20-26 weeks. Yes, this ties up more working capital, but it’s better than halting production.
Talk to your distributors. The major distributors like Mouser, Digi-Key, and Farnell have allocation teams that can help prioritize orders if you have a relationship and can provide demand forecasts.
Consider design flexibility. If you’re starting a new project, architecting for multiple MCU targets isn’t trivial, but frameworks like Zephyr RTOS and PlatformIO can make it more feasible to port between chipsets.
My Take#
I’ve been building hardware-adjacent systems for long enough to have lived through previous semiconductor supply hiccups, but this one feels different in scale and duration. The combination of pandemic demand shifts, concentrated fabrication capacity, and increasing chip specialization creates a structural challenge that won’t be resolved quickly.
The silver lining, if there is one, is that this shortage is finally generating political will to diversify chip manufacturing. The US, EU, and Japan are all discussing incentives to build domestic fabrication capacity. Whether those plans materialize into actual fabs remains to be seen, but at least the conversation has moved from “interesting idea” to “national security priority.”
For now, order early, plan for delays, and keep your designs flexible. The chip shortage is the tech industry’s supply chain wake-up call — and we’d be wise to listen.
