Raspberry Pi surfs memory crunch to record first half

Raspberry Pi posted record first-half revenue of $256.9 million, up 90% year-on-year, while pre-tax profit more than tripled to $19.6 million; gross profit per board rose to $12.20 from $8.00. A strategic memory stockpile enabled product availability and helped lift gross profit 79% to $59.4 million, though these exceptional unit economics are moderating as low-cost inventory is exhausted and memory costs rose to $13.30/GB from $3.60/GB. The company holds sufficient memory inventory and confirmed orders to meet its 2026 production targets, with strong demand in higher-memory boards, smart-home applications, and aerospace and defense despite Pi Zero supply constraints.

Analysis

RPI's earnings quality is likely peaking rather than compounding at the first-half rate: low-cost memory embedded in cost of goods sold created a temporary gross-margin tailwind, while replacement inventory is being purchased at materially higher prices. The key 1-3 month debate is whether price increases and mix toward higher-memory boards can fully offset the lagged COGS reset; if not, consensus estimates anchored to the first-half run rate will require downward gross-margin revisions despite continued unit growth.

The more durable positive is allocation-driven share capture. Smaller single-board-computer vendors without inventory financing capacity or direct supply relationships are likely losing OEM design wins now, and those wins can persist for 6-18 months because industrial, aerospace and smart-home customers face qualification and software-porting costs. That supports a higher share of recurring commercial demand, but it also converts RPI from a low-inventory hobbyist hardware model into one with substantially greater working-capital and obsolescence exposure.

The non-obvious bottleneck is packaging, not just DRAM: constrained Taiwanese backend capacity can prevent RPI from monetizing secured memory inventory. This is constructive for LPDDR4 suppliers and outsourced semiconductor assembly/test firms with relevant Taiwan exposure, but it weakens the simple thesis that stocked RAM guarantees revenue. A reversal would be signaled by declining backlog, board gross profit falling below roughly $10, or inventory days rising without matching revenue growth; each would imply either demand elasticity or a supply-chain conversion problem.

Key Decisions for Investors

• Maintain a tactical long RPI only through the next results/production update, sized modestly: the catalyst is evidence that second-half volume growth absorbs higher-cost inventory without a material gross-margin reset. Take risk down if management guides board gross profit below $10 or signals inventory write-downs; absent valuation and consensus-margin data, do not underwrite a longer-duration position.
• Use any post-results strength to consider a 1-3 month RPI short or put spread if management confirms that replacement-memory costs are flowing into COGS faster than realized price increases. The asymmetry improves if revenue remains strong while gross margin guidance falls, since the market may initially reward shipments rather than recognize the working-capital burden.
• Watch Micron (MU) and SK Hynix exposure as indirect beneficiaries of persistent specialty/legacy mobile-memory tightness; prefer MU as the liquid proxy, but require evidence that LPDDR4 pricing—not only leading-edge HBM demand—is supporting realized pricing before initiating.
• Monitor outsourced assembly and test capacity in Taiwan as a supply-chain alert rather than a trade: confirmation that packaging constraints are easing would remove a near-term cap on RPI shipments, while further allocation delays would undermine the second-half delivery thesis even with adequate DRAM inventory.