The Hidden Cost of AI: When Data Centers Devour Your Supply Chain

Analysis of artificial intelligence’s place in the world is as ubiquitous (and occasionally, insufferable) as those chatbots cluttering up many businesses’ websites. Not unironically, then, am I adding to the din.

More than 5,400 data centers operate in the US alone today, including more than 50% of the world’s hyperscale AI facilities. Last year alone, global spending on data centers reached an estimated $580 billion and is expected to continue to grow at a 28% clip per year through 2030, with AI investments alone reaching $200 billion by 2030. Assuming much of that investment is financed, interest charges from that capex could exceed $1.5 trillion.

Much of the impact – realized or projected – has focused on energy consumption and location, and for good reason. While local folks might want access to AI so they can make quick work of otherwise time-intensive projects, they tend not to want to see their energy bills spike to underwrite the data center across town.

The Next Great Technology Challenge: Designing without Rare Earths

As demand for AI and advanced electronics grows, the industry faces mounting pressure to reduce its reliance on scarce rare earth materials.

Much has been written and said in all areas of the world and in all walks of life about the challenges facing the world of technology. Whether it is developing and harnessing AI (artificial intelligence), utilizing electric vehicles, reducing pollution to leave a smaller carbon footprint, or training the next generation of employees to fill the multitude of jobs required to manufacture the advanced technologies that all the above will require, the number and magnitude of all these challenges is staggering. It is nothing, however, compared to the granddaddy of them all: creating the next generations of technology without depleting rare earth minerals.

Rare earths are front-page news today, but regrettably the focus is more toward geopolitical wrangling among countries that have them and those that do not. But all the saber-rattling among competitors fails to address the root challenge: how can new technologies be developed that perform the functions we want while using far less – or no – rare earths?

First, what are rare earth minerals? They include minerals that most people have heard of, such as cobalt, lithium, manganese, graphite, and silicon derived from quartz. There are far more than those, however, most with names that are hardly household: neodymium, praseodymium, dysprosium, terbium and samarium, to name a few, are among the ones used to form industrial-quality magnets, sometimes for high-heat applications. These can be found in the battery of just about any hybrid or electric vehicle.

Do You Have a 2026 PCB Buying Strategy?

A PCB buying strategy shifts focus from unit price to total cost, design discipline and supplier governance.

This is not the year for buyers to accept PCB price increases blindly. This is the year to buy PCBs like a professional who understands the job isn’t just about placing orders; it’s about controlling total cost of ownership while protecting delivery, quality and continuity.

Here’s how I believe board buyers should approach 2026, including what to change in your quoting process, how to structure your supplier base and how to use logistics and design to reduce costs. These are concrete buyer moves you can implement immediately.

Stop comparing PCB quotes and start comparing total cost of ownership. Most buyers still shop boards like it’s 2019; i.e., comparing unit prices, choosing a vendor and putting off dealing with shipping and tariffs until later. In 2026, that approach will have your company bleeding money.

Dashboards and Controlled Chaos

In an EMS environment built on controlled chaos, dashboards help program managers catch problems early and keep accounts on track.

The electronics manufacturing services (EMS) industry is controlled chaos by design. The basic EMS value proposition is that outsourcing relieves original equipment manufacturers (OEMs) of manufacturing challenges. The EMS provider covers fixed costs during manufacturing and carries the associated inventory costs on its balance sheet. Market slowing down? Revise your EMS provider’s forecast. Market improving? Tell your EMS provider to pull some production in. Engineering challenges? Send your EMS provider a set of engineering change orders (ECOs) as the team determines what needs to change. The OEM side of the equation has the chaos part of the business model down. Even when things are running smoothly, the number of unplanned emails and calls in keeping projects on track is a significant part of the program management team’s daily activities. When more of the day gets consumed than planned, activities designed to control the chaos may not get done.

This underscores the importance of having a control system that makes sense for managing that chaos. Walk a production floor, and you’ll see control limits established for critical processes such as solder paste deposition and reflow. If a process exceeds those limits, either the monitoring software in that equipment or the inspection equipment immediately following that process tells the operator a problem has developed so that the operator can identify the root cause and bring it back within control limits before significant amounts of defective products are built. Program managers need to create control systems that provide that intelligence as well. A dashboard is an effective way to highlight developing issues when they are small.

The basics of a good dashboard include:

• It educates. You can’t expect customers to change behavior if you can’t show them why that behavior is costing them unnecessary money. Dashboards need to be data-driven, and those data need to be able to support the program manager’s business case.

Rebuilding America’s Electronics Backbone, One Member at a Time

The PCBAA is on a mission to ensure domestic security via a revitalized American manufacturing base.

Across the United States, a quiet but urgent realization has been taking shape – one that echoes through defense briefings, aerospace reviews, medical device evaluations and critical infrastructure planning sessions. The nation that once led the world in electronics manufacturing now faces a stark truth: the ability to design and manufacture advanced electronics on domestic soil has diminished to dangerous levels.

This isn’t speculation. It’s not a theory. It’s a matter of national security.

Again and again, leaders across government and industry have expressed the same concern:
A country that cannot build its own electronics cannot secure its own future.

That fundamental truth fueled the creation and mission of the Printed Circuit Board Association of America (PCBAA) in 2021. The organization exists because the stakes are no longer academic – they are strategic, economic and deeply tied to national security and resilience.

Inspecting Printed Circuit Boards and Assemblies

How do inspection discipline and sampling plans decide whether a shipment ships or gets torn apart?

Before going into PCB design, my employer was in the telecom business. I started out putting PCBs into antistatic bags, then into individual boxes with appropriate labels. A group of eight distinct boards was placed in a larger box to form a die group. The big box labeling reflected the part-dash number and revision for each board. This was called “final prep” and was the last step prior to shipping.

A few selected to represent the whole shipment. Someone would audit the shipping boxes, looking for incorrect information on the labels and any other obvious defects. If they found a problem with the ones sampled, the entire lot would be re-examined by final prep. A missing antistatic awareness label could jeopardize the whole shipment. It could happen, especially toward the end of the quarter when we were sprinting to make our quotas.

Mastering Signal Integrity: From Fundamentals to AI-Powered Analysis

Signal integrity underpins reliable electronic design, and emerging AI tools are reshaping how engineers optimize it across all system speeds.

Ever wonder what makes our modern electronic devices so incredibly reliable, whether they’re processing gigabits of data or simply ensuring a sensor reads accurately? While many factors contribute, one unsung hero often working behind the scenes is signal integrity (SI). Think of it as the invisible guardian of your electrical signals, ensuring they travel cleanly and efficiently from point A to point B. Ignoring it can lead to headaches, delays and even product failures, regardless of your operating speed. But embracing it? That’s where the magic happens.

In today’s electronic systems, from the fastest data centers to the most sensitive medical devices, a deep understanding of how signals propagate and how to mitigate issues that degrade signal quality has become critical. It’s not just about “high-speed” anymore; it’s about any signal path where clean, reliable transmission is paramount. Let’s dive into why understanding and prioritizing signal integrity is crucial for every designer across all design spectrums.

Getting SI Right from the Start

Imagine building a skyscraper without a proper blueprint or a solid foundation. It wouldn’t stand a chance, right? The same applies to your electronic designs. The very first steps you take in your PCB layout can make or break your signal integrity.

Value Chain Reshaping is the Latest Maturing Stage of the Global Electronics Industry

Under Foundry 2.0, the semiconductor value chain is moving back into strategic focus.

When Joni Mitchell recorded Big Yellow Taxi, singing “You don’t know what you’ve got till it’s gone,” she was reiterating the proverbial warning that we often appreciate things properly only after losing them. It’s an observation that transcends context and can be applied even in today’s electronics industry. For decades, Western companies have outsourced significant parts of their value chain to achieve cost-down and to focus on core competencies in pursuit of efficiency. Today’s geopolitical tensions are drawing attention to the loss of sovereignty that results from exporting control of critical processes like packaging and testing as part of the semiconductor value chain.

While repatriation initiatives like the US Chips Act have focused on silicon, it’s becoming clear to all involved – politicians included – who depend on experts from the industry to provide technical and practical insights, that achieving any meaningful future shift in the balance of economic power is a much broader task than setting up wafer fabs alone. Although establishing those fabs is a huge undertaking that demands massive investment, the Act has not ascribed comparable value to activities like packaging and testing.

Managing Moisture and Heat in Flex and Rigid-Flex Assembly

Successful flex and rigid-flex assembly depends on controlling moisture.

Questions about baking and assembly come up at least monthly, if not more often. Several factors can impact successful assembly with flex and rigid-flex.

Environment. As a rule, delamination is usually due to retained moisture. The rapid rise in temperature during reflow causes moisture to change from liquid to expanding vapor or steam. This expansion can result in delamination.

Storage and factory conditions can certainly impact assembly success. Baking times and temperatures are definitely humidity-dependent. Most guidelines are not defined based on a specific condition. I would treat them as a baseline, and if humidity is high or if parts were exposed to a long liquid cleaning, I would increase bake times.

robert
boguski

Force Majeure

A holiday-season scramble exposes how “always on” offshore manufacturing ideals collide with real-world AXI limits.

A Yuletide surprise, courtesy of The Anthill. The same Anthill that wouldn’t answer emails, phone calls, texts or carrier pigeons the other 332 days of 2025. Things change, and snubs become embraces overnight when year-end revenue is threatened. Lucky us. Their AXI machine was down indefinitely; ours was definitely up. Meanwhile, an impossibly unrealistic, unreasonable, irrational name-brand tech superstar du jour was expecting results, no excuses, to fulfill a product launch. 

This being the season of sharing, the Anthill proposed sharing their misery with us. They asked if our small, energetically earnest staff (13) would drop everything else and forego sleep for, well, the next six weeks. At least until Lent, but in pagan time: forget Christmas, and observe abnormal work hours. You’ll sleep and find leisure when you die, goes the motivational thinking, and, apparently, the Anthill practice. Through unremitting toil shall you be saved is progress made. Scrooge lives. Joy to the world. For the right price.

Once upon a time, when dinosaurs roamed the Earth and youthful male tech leaders were still admired (2011), a leading fruit-emblemed toy company CEO told an inquisitive American president that 24/7/365 mass production of certain specialized electronic devices was impossible in the United States. This visionary supplied indigestion to our president, in the form of dinner table examples of the other side’s manufacturing superiority. Deeper supply chain. Better engineering support. Faster response time. Greater flexibility in making changes. Ability to scale from prototype to pilot build to production faster than anywhere else. Regimented social structure. Unquestioning willingness to work 3000 people 36 hours straight to implement an ECO. Worker docility in the guise of hyper-attentive customer service, a techbros’ wet dream. All the attributes of an anthill. This, proclaimed the famous CEO, was manufacturing as it should be. He insisted to our incredulous president that such characteristics could not be replicated in the United States. The implied truth: we’re lazy, as he defines the term, and they’re not.