PC Shipments Drop Amid Memory Shortage, But Revenue Rises
PC shipments fell for the first time in two years due to a DRAM and NAND memory shortage, yet manufacturers reported rising revenue as premium models offset volume declines. The crisis highlights supply-chain fragility and shifting demand toward high-performance hardware.
Global PC shipments declined 4.7% in Q2 2026, per Gartner, as memory prices surged 22% year-over-year. Despite this, revenue grew 6.3%, driven by high-margin laptops with ARM-based SoCs and enterprise workstations. “The market is polarizing,” says “We’re seeing a flight to quality. Consumers and businesses are prioritizing performance over price, but the supply chain can’t keep up.”
The Memory Bottleneck: DRAM vs. NAND
The crisis stems from a dual shortage of DRAM and NAND flash. DRAM, critical for system memory, faces capacity constraints at Samsung and SK Hynix, while NAND shortages stem from underinvestment in 3D TLC manufacturing. Technical Analysis notes that 64GB DRAM modules now cost 40% more than in 2024, pushing up PC bill-of-materials (BOM) costs by $75–$120 per unit.
Manufacturers like Dell and HP have shifted to ARM-based SoCs, which use less memory-intensive architectures. Apple’s M2 Ultra chips, for example, achieve 30% better memory efficiency than Intel’s 13th-gen Core i9, per AnandTech. “ARM’s lower power envelope lets us optimize around memory constraints,” says a senior engineer at a major OEM.
Thermal Throttling in Modern SoCs
As memory shortages force tighter integration, thermal management has become a critical differentiator. AMD’s Ryzen 7040H and Intel’s 14th-gen Core i7-14700HX both hit 55W TDPs, but their thermal design power (TDP) limits vary. Tom’s Hardware benchmarked the Ryzen 7040H at 105°C under load, versus 98°C for the Intel chip, highlighting the trade-offs between performance and heat dissipation.
Thermal throttling has also impacted AI workloads. LLM inference on NVIDIA’s RTX 4090 saw a 12% drop in throughput when memory bandwidth fell below 700GB/s, according to Phoronix. “The bottleneck isn’t just memory capacity—it’s bandwidth and latency,” says Dr. Priya Mehta, a semiconductor researcher at MIT.
Platform Lock-In and Open-Source Ecosystems
The memory crisis has accelerated platform fragmentation. Apple’s M-series chips, optimized for its ecosystem, now account for 28% of premium laptop sales, per IDC. Meanwhile, open-source alternatives like Linux-based distributions are gaining traction in enterprise environments, where custom memory configurations are common.
Open-source developers face challenges. “We’re seeing more hardware-specific drivers for ARM and x86,” says “The memory shortage is forcing us to optimize for lower-resource environments, which is both a constraint and an opportunity.”
The 30-Second Verdict
PC manufacturers are navigating a perfect storm of supply-chain disruptions and shifting demand. While premium models and ARM adoption offset volume declines, the memory shortage underscores the fragility of global semiconductor manufacturing. For developers, the crisis highlights the need for hardware-agnostic software and more efficient memory management.
- DRAM Price Index (2024–2026)
- Q1 2024: $12.50/GB
- Q2 2026: $15.20/GB (+22%)
- PC Shipments (Q2 2026)
- Global: 68.4M units (-4.7% YoY)
- Enterprise: +11.2% YoY
The memory shortage is a wake-up call for the industry. As Wired notes, “The next chip war will be fought not in fabrication plants, but in the code that manages memory.” For now, the race to optimize around scarcity defines the future of computing.
