Multibagger Selection: Claude Sonnet 5.5's picks
Claude Sonnet 5.5 picked AMBQ, ASPN, SILC, ETON, CDNA with a twelve-month thesis, measured against SPY at each checkpoint.
| Experiment | Multibagger Selection |
|---|---|
| Model | Claude Sonnet 5.5 |
| Prompt version | multibagger-selection-v1 |
| Asked | |
| Horizon | 12-month |
| Benchmark | SPY |
| Status | Being measured |
The picks
| Rank | Symbol | Return | Confidence | vs benchmark | Called right |
|---|---|---|---|---|---|
| 1 | AMBQ | — | 60% | — | — |
| 2 | ASPN | — | 48% | — | — |
| 3 | SILC | — | 55% | — | — |
| 4 | ETON | — | 64% | — | — |
| 5 | CDNA | — | 58% | — | — |
Checkpoints
| Checkpoint | Portfolio return | Return vs SPY | Best peak multiple | Picks reaching 2x |
|---|---|---|---|---|
| 1 week | +1.48% | +1.77% | 1.06x | 0 |
The model's market view
Small caps have led in 2026 (S&P small-cap growth was up about 26% YTD in mid-August, and a Russell 2000 earnings rebound is expected), but the Russell 2000 fell about 5.7% over the last month, and long-term Treasury yields are near 24-year highs with the Fed hiking and oil above 100. That favors companies with already visible revenue acceleration over concept stories. A 3x return in 12 months is economically plausible only for the smallest names here (SILC, ASPN). For the larger caps (AMBQ, ETON, CDNA) it requires roughly 2x or more revenue growth plus sustained multiples, so those are scenarios, not expectations. Several picks are volatile with betas of about 2.4-3.1, and the figures I could not verify are marked as such.
Why each pick
AMBQ
Ambiq Micro (NYSE, IPO July 2025) sells ultra-low-power chips for edge AI, wearables and IoT. Demand has stepped up sharply this year: Q2 2026 net sales grew about 90% year over year to $33.9M with a sequential gain for a fifth straight quarter, and management guided FY26 net sales to about $135M versus $72.5M in FY2025 (consensus before the call was about $122M). Gross margin was 47.2% in Q2. The stock is about 46% below its 52-week high of 91.61 after a steady decline in September and October, with seven analysts rated Buy on average (targets from 73 to 125, including new initiations at Jefferies and Roth). It is not cheap on sales (about 8.8x FY26 guided sales on current market cap, balance-sheet cash not verified).
ASPN
Aspen Aerogels makes aerogel thermal barriers for EV battery packs (PyroThin) and insulation for energy-industrial projects. Revenue collapsed (TTM $202M, about -49% y/y) after an East Providence plant incident and an EV slowdown, but the recovery is now visible: Q2 2026 revenue $49.8M (+32% sequentially) and Q2 thermal-barrier revenue $29.5M (+81% sequentially). The company guided Q3 2026 revenue to $65M-$80M against a $56.9M consensus and adjusted EBITDA to $7M-$15M, i.e. a swing to positive EBITDA. A General Motors supplier award and a Jaguar Land Rover award for vehicles entering production in 2027 support the thermal-barrier ramp. At about 2.2x TTM sales on a depressed base, the market cap is low relative to the planned recovery.
SILC
Silicom (NASDAQ, Israel-based) makes server network interface cards, smart cards and edge networking devices. Revenue is accelerating: Q1 2026 +33%, Q2 2026 +59% to $23.8M, and FY26 guidance was raised twice to $93M-$95M (about +50% vs $61.9M in 2025), with Q3 guided to $25M-$26M. Operating expenses grew only 16% against 59% revenue growth, narrowing the GAAP operating loss to $2.0M, and management projects a return to quarterly profitability by year-end. New AI-inference production orders (about $3M in 2026 scaling to about $10M in 2027 from one customer) and several $5M+/year design wins add visibility. It is the smallest company on this list and very thinly held (5.75M shares).
ETON
Eton Pharmaceuticals commercializes rare-disease medicines (Increlex, Alkindi Sprinkle, Khindivi, Galzin, Hemangeol and others). It has the best-evidenced inflection on this list: FY2025 revenue $79.95M (+105%), Q2 2026 revenue $37.6M (+99% y/y) against a $27.1M consensus, and FY26 revenue guidance raised to more than $145M with an adjusted EBITDA margin of at least 35% (up from 30%). The company is GAAP profitable (TTM net income $12.7M), supports debt reduction from cash flow, and added Impavido (launched Sept 28), the Libervant license (Oct 14) and ASN-001 (acquired Aug 5). Four analysts rate it Strong Buy with a $71.50 average target, and several raised targets after Q2. The stock is already up about 174% over the year, so the question is how long growth lasts.
CDNA
CareDx is a transplant diagnostics company (AlloSure, AlloMap, HeartCare) that is expanding into specialty oncology through the Naveris acquisition (completed July 1, 2026) after selling its lab products business (completed June 30). Q2 2026 revenue rose 52% to $131.9M against a $114.1M consensus, testing services grew 61%, and FY26 guidance was raised to $490M-$500M revenue with $66M-$78M adjusted EBITDA. The Medicare local coverage determination for transplant molecular testing was finalized in July 2026, removing a reimbursement overhang. Cash was $373.6M at June 30. It is the largest and most fully valued company here, so it carries the weakest multibagger case.
Other models on the same run
| Model | Picks |
|---|---|
| Muse Spark 1.1 | LUNR, PL, SOUN, BKSY, EXTR |
| Perplexity | RXRX, SOUN, ACHR, BBAI, CRSP |
| Gemini 3 Pro | ASTS, RKLB, SMR, LUNR, ACHR |
| Grok 4.5 | INOD, PGY, TMDX, SEZL, CALX |
| DeepSeek | INOD, OUST, AEVA, LAES, AREC |
| GPT-6 | ARDX, AMPX, FLNC, TMDX, SERV |
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.