Rank + list placement
Show exactly where the name sits in the monthly product.
A repeatable package that turns each Upside Ranks ticker from a scored list entry into a publishable research card: rank, catalyst, key numbers, bull path, bear case, source links, and what would change our mind.
Show exactly where the name sits in the monthly product.
Force the idea into a clean investment claim before adding detail.
Explain the current catalyst, not just the long-term story.
Anchor the thesis in hard data instead of vibes.
Make the upside path explicit and falsifiable.
Show what could break the idea before readers discover it for us.
Create a clean future update trigger for the next monthly refresh.
Make every material claim auditable.
Credo is the highest-conviction 2x name because it attacks a real AI cluster bottleneck: moving data inside scale-out systems without blowing up power budgets.
2x rank #1; probability 14.6%; quant score 82.6.
AEC unit ramp at hyperscale AI data-center customers
A 2x needs three things: FY2027 revenue visibility moving toward the $2.5B-$3.0B zone, sustained gross margins above 60%, and evidence that multiple hyperscaler programs are ramping. If those show up together, Credo can be valued as a core AI connectivity platform rather than a single-cycle AEC winner.
The biggest risk is concentration. Two customers drove 87% of Q3 FY2026 revenue, so a pause, design loss, or inventory digestion event could break the thesis fast. The valuation also leaves little room for execution mistakes.
Upgrade: SEC filing attributes the revenue increase primarily to higher AEC shipments.
Cut/demote: Demote if customer concentration does not broaden beyond two hyperscalers by the next two reporting cycles.
NVIDIA remains the cleanest mega-cap expression of the AI compute bottleneck, but this is no longer a discovery story — it is an expectations and revision story.
35% rank #6; probability 37.1%; quant score 78.4.
Record Q1 FY2027 revenue and Data Center revenue, with Q2 revenue guide stepping up again to $91B.
+35% is plausible if Blackwell/Rubin demand keeps exceeding supply, Q2 guidance proves conservative, inference revenue broadens beyond training clusters, and the market keeps raising out-year earnings estimates. This is a high-quality, high-expectation setup: the stock needs continued upside revisions, not just another strong quarter.
The risk is not business quality; it is the altitude of expectations. A pause in hyperscaler capex, export-control pressure, faster custom ASIC substitution, or even modest margin compression could make strong execution insufficient for another +35% move.
Upgrade: Blackwell, Vera Rubin, NVLink Fusion, optics partnerships, and Dynamo inference software deepen the full-stack AI factory moat.
Cut/demote: Sequential Data Center growth slows sharply without a clear supply-transition explanation.
Aehr is the purest small-cap test-equipment moonshot in the June list.
5x rank #2; probability 8.7%; quant score 89.1.
April 16, 2026 record $41M follow-on production order from lead hyperscale customer for package-level burn-in of custom AI processor ASICs used in data-center training and inference workloads.
A 5x requires Aehr to move from niche supplier to critical test-equipment beneficiary across AI-adjacent silicon carbide, photonics, and advanced semiconductor reliability markets. The base bull case is smaller than 5x; the 5x case needs a new order cycle and broader customer adoption.
Aehr is small, order-driven, and customer-concentrated. Revenue can be lumpy, orders can slip, and weak FY2027 conversion would quickly undermine the thesis. This is risk-capital only, not a quality compounder.
Upgrade: Second-half FY26 bookings exceeded $92M as of April 16, already above the high side of the previously increased $60M-$80M expectation.
Cut/demote: Lead hyperscale AI customer pushes out Sonoma deliveries or next-gen ASIC production.
A ranked name is not ready for the final monthly package until this thesis card, the scoring audit, and the source-confidence gate agree. If the card cannot explain “why now” and “what would change our mind,” the rank needs more research before publication.