AI chip demand stays strong, but component shortages emerge as next challenge: Nomura
New Delhi, July 2
The artificial intelligence boom is entering a new phase where the biggest challenge is no longer demand for AI chips but shortages of critical components needed to build AI infrastructure, according to a Nomura Report, which said supply constraints across the hardware ecosystem are likely to intensify through 2027 even as AI investment remains strong.
The report said the AI investment cycle is far from over, with demand continuing to be driven by aggressive spending on AI data centres by hyperscalers such as Microsoft, Google, Amazon and Meta. However, it warned that shortages are shifting beyond advanced chips to a wider range of components across the supply chain.
"We believe there will be an unprecedented component supply-mismatch period in 2H26F, and we expect it will get worse in 2027F," Nomura said in the report. It added that, apart from advanced packaging, memory and CPUs, "PCB/CCL, IC substrate, higher-end capacitors, power management IC (PMIC), and optical components are also already in shortage currently."
According to the report, these shortages could spill over into other industries as well.
"This could further affect the supply for non-AI subsectors such as consumer and auto," it said, adding that "supply chain price hikes could continue or increase with worsening shortages."
Despite the growing supply bottlenecks, Nomura said it does not believe the AI infrastructure investment cycle has peaked.
"There could be a few valid factors that may explain the recent share price pullback, but likely we have not yet reached the peak of this cycle," the report said. It attributed this to continued upside in hyperscalers' capital expenditure plans and a steady pipeline of new data centre projects globally.
The brokerage said its latest tracking of global data centre projects indicates that demand for AI hardware will remain strong over the next few years.
It added that incremental AI data centre capacity deployment is expected to rise to 32 gigawatts in 2027 from 28 gigawatts estimated earlier.
The report said the current constraints are no longer centred only on chip manufacturing capacity but shortages are shifting deeper into the semiconductor supply chain.
Reflecting its bullish outlook, Nomura raised its global server market forecasts, expecting AI infrastructure spending to remain robust over the next two years.
"We are raising our 2026-27F server market forecasts on stronger AI and general/CPU server sales," the report said. It now expects global server revenue to grow 74 per cent year-on-year in 2026 and a further 65 per cent in 2027, supported by continued AI deployment.
The report noted that while supply shortages and higher component prices could create short-term volatility, they also reinforce its view that demand for AI infrastructure remains strong and that the current investment cycle still has room to run.
— ANI
Reader Comments
Interesting analysis from Nomura. The shift from chip shortage to component shortage is quite telling. I work in supply chain and we're already seeing lead times extend for capacitors and PMICs. This is going to hit everyone—from data centers to your car's infotainment system. Buckle up.
While I appreciate the optimism about AI demand, I'm a bit concerned about the environmental cost of all these data centers. 32 gigawatts of new capacity by 2027 is massive. Where will all that power come from? We need sustainable AI, not just more chips. Renewable energy investment should match AI investment.
This is a classic boom-bust cycle pattern. Everyone rushes to build AI infrastructure, then component shortages hit, prices spike, and eventually some projects get delayed. But the demand is real. India should position itself as a manufacturing hub for these components now. Taiwan and China can't be the only suppliers forever. Good wake-up call for Indian policymakers. 👍
Finally someone is talking about the real bottleneck! Everyone just talks about AI chips but forgets the thousands of other components needed. My uncle runs a small electronics factory in Pune and he's already struggling to get basic capacitors. The supply chain is more fragile than people realize. 😅
As someone working in a semiconductor fab, I can confirm this is spot on. The advanced packaging bottleneck alone is massive. But I question whether the 74% growth forecast is realistic. Component shortages will likely temper that. Still, good for Indian IT stocks if they can navigate these supply issues.