RMBS is a pure-play beneficiary of the CPU memory renaissance in AI, with multiple product cycles driving sustained 20%+ revenue growth.
Execution and timing risks, coupled with a premium valuation, create downside if growth fails to accelerate.
The company is seen as a clean way to play the 'CPU memory renaissance' in AI infrastructure, driven by agentic AI and inference requiring more persistent, memory-intensive, CPU-heavy workloads.
Management explicitly stated that standard server demand is growing, the server market is now modeled at low-double-digit growth, and the CPU:GPU mix in inference/agentic AI is moving in the company's favor.
The product revenue shortfall in the guide was attributed to supply/cadence issues, not a deterioration in demand.
Multiple layers of content expansion are expected, including DDR5 RCDs transitioning from Gen2 to Gen3, Gen5 shipments starting, companion chips moving to mid-double-digit percentages of product revenue, and MRDIMM representing a ~$600M SAM.
SOCAMM2 provides an early foothold in LPDDR-based server modules, with LPDDR6 potentially becoming more content-rich, and silicon IP is benefiting from AI custom silicon, hyperscaler ASICs, and other IP offerings.
Bulls project product revenue growth of 20%+ in CY26 and 25-30%+ in CY27/CY28, with a base bullish case of CY27 revenue near $975M and EPS of $3.50-$3.60, potentially exceeding $4.00, leading to a $160+ stock at a high 30s multiple.
New products were a low-double-digit percentage of product revenue in Q1, expected to be similar in Q2, and management anticipates them to exit 2026 at a mid-double-digit percentage, supporting the 'one-stop shop' narrative.
Management reiterated a ~$600M SAM and 2027 ramp tied to next-gen Intel Diamond Rapids and AMD Venice platform launches, modeling a conservative attach rate due to platform timing and pricing uncertainties.
A lot of this does seem backward looking and before OAI pivoted to enterprise and the Codex ramp, but with Oil +5% encouraging a risk off morning, semis having had a huge run MTD, froth reading near highs, along with some weaker earnings from RMBS/CLS, all...
The stock had already priced in more than an 'in-line' performance, trading at a premium multiple, and investors were looking for visible EPS upside and a sharper product revenue recovery, which was not delivered.
Q1 was acceptable, but Q2 product revenue guidance was below Street and likely below buy-side expectations, EPS guided lower, and operating margin guided below consensus due to higher operating costs.
The bear view is that the story is not broken, but the stock is expensive and requires clean beat-and-raise quarters, not in-line quarters with supply caveats.
Bears focus on execution and timing risk, as the highest-value opportunities like MRDIMM and Gen5 depend on Intel and AMD platform ramps, and management continues to cite backend supply tightness.
SOCAMM is strategically interesting but not a 2026 revenue driver, and LPDDR could be both an opportunity and a risk if architectures evolve to reduce DDR/MRDIMM attach.
Competition from Montage and Renesas in DDR chipset offerings, along with broader risks from Chinese competitors, customer concentration, memory cyclicality, and potential pricing/module-cost sensitivity, are concerns.
Bears believe the stock will re-rate lower if CY27 EPS remains around $3.00-$3.50 instead of quickly moving towards $4+, suggesting a trade closer to 25-30x rather than high 30s, resulting in a $100 stock.
Research through 2026-04-28 · 2026-04-28 to 2026-04-28 · 1 source reports.
| Range | Prices through |
|---|---|
| 1W | |
| 1M | |
| 3M | |
| 6M | |
| 1Y | |
| 2Y | |
| 3Y |
Research through 2026-11-03. Artifact validity does not establish factual accuracy or research freshness.