Greedy AI industry leaves other supply chains struggling to find chips

AI Processor

The AI sector’s voracious appetite for memory chips is taking a toll on supply chains for other industries, where companies scramble for new suppliers, facing scant supply, long wait times and soaring costs.

The observation that cars have become ‘computers on wheels’ has ominous repercussions for the industry that makes them.

Traditionally the auto industry has been at the head of the line for parts, including memory chips, but today manufacturers are struggling to find them in sufficient numbers, and on time, which has spawned tales of factories with finished vehicles waiting for a late part that involves a memory chip.

Lead times on engine sensors, control modules ,and fuel injectors have stretched to 8-12 weeks in many cases, one logistics executive dealing with the auto industry remarked.

Chip supplies have been squeezed by the unquenchable need of the AI-related industry to equip its servers and data centres. By one estimate, this sector hoovers-up about 70% of global chip supply, leaving firms in other verticals competing for the rest.

“The increase in demand is real,” noted Bryann Lim, specialist, global sales and marketing, at Dimerco Express Group, pointing to the widening trade imbalance of the US with Taiwan, South Korea, Vietnam, Malaysia, and Thailand.

He added that some of Dimerco’s clients in the hi-tech sector were expecting growth in the latter half of this year into 2027.

A spokesman for AIT Worldwide Logistics reported that ‘hyperscalers’ were commanding a lot of airfreight capacity, especially out of Taiwan, pushing up rates and the transport costs for automotive semiconductors.

“The market is transitioning to a semiconductor supply-constrained environment,” he said. “The bottlenecks are no longer data centre construction timelines or server assembly, it’s advanced AI chips, high bandwidth memory, advanced packaging capacity (packaging multiple chips together with memory for AI applications), networking components and power infrastructure.”

And the chip producers are exacerbating the problem by allocating a larger portion of production to high-bandwidth memory (HBM) chips, reducing the number of other, less lucrative types. Mr Lim noted that the recent profit surges reported by SK Hynix and Samsung had been largely attributed to HBM demand.

However, auto makers are not the only ones left scrambling for supplies. Bob Imbriani, EVP international at forwarder Team Worldwide., noted: “Multiple industries have been affected.”

Besides obvious sectors like producers of cell phones, laptops and tablets, companies in the machining industry have also faced tight supply, as a lot of their machines are computer-controlled, he said, adding that Apple had raised prices of its products.

One of Team’s customers has said the problem was not only availability of memory chips, but also the fact that available inventory often goes to the highest bidder, and AI firms appear willing to outbid traditional users, gobbling up stock.

Prices for memory chips in the auto industry have reportedly climbed as much as 70%. As US carmakers are approaching the point where they can no longer absorb increased costs from tariffs, US consumers stand to see a painful rise in prices.

Manufacturers have been looking for new suppliers in locations like Costa Rica, Mr Imbriani reported. However, those new suppliers are in the process of gearing up their production, which takes time, he added.

Stuart Mitchell, VP of finished vehicles at Circle Logistics, described the repercussions for automotive supply chains as “transformative”. Historically the underlying assumption has been that supply constraints were temporary, but this has changed with semiconductor capacity permanently re-allocated to AI infrastructure. This new reality calls for a structural re-alignment of supply chains, he said.

He argued that auto makers could no longer treat freight as something to arrange after product has been ordered. The freight part should already be in motion before orders for parts are confirmed. This requires better visibility of supply chains, allowing for problems to be identified before their impact, he noted.

Another factor he regards as critical is to have dedicated capacity to move orders. Rather than the relationship with the supplier, it is often freight execution speed that can determine the outcome of a scramble for supply, he noted.

From AIT’s vantage point, capacity is the key factor at the moment.

“Supply chain strategy has shifted from cost optimisation to capacity acquisition,” the spokesman commented.

 

Добавить комментарий

Больше на Mikhail Family Investment

Оформите подписку, чтобы продолжить чтение и получить доступ к полному архиву.

Читать дальше