Multibagger Selection: Claude Sonnet 5.5's picks
Claude Sonnet 5.5 picked AEHR, FCEL, ETON, NVTS, FSLY 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 | AEHR | — | 58% | — | — |
| 2 | FCEL | — | 50% | — | — |
| 3 | ETON | — | 68% | — | — |
| 4 | NVTS | — | 42% | — | — |
| 5 | FSLY | — | 55% | — | — |
Checkpoints
| Checkpoint | Portfolio return | Return vs SPY | Best peak multiple | Picks reaching 2x |
|---|---|---|---|---|
| 1 week | -10.82% | -9.90% | 1.05x | 0 |
The model's market view
US small caps have led in 2026: the Russell 2000 was up about 20% year to date in late August, and Q2 2026 small-cap earnings growth was reported near 43%. The October tape is harsher for the high-beta AI-infrastructure names screened here. Over the past month AEHR is down about 22%, FCEL about 28% and NVTS about 23%. Roughly 40% of Russell 2000 companies are unprofitable and exposed to refinancing risk. Because a 3x outcome requires a rare mix of backlog conversion, margin expansion and multiple re-rating, none of the five clears 3x as merely plausible. The ranking therefore favours names with documented backlog, guidance or committed capacity over narrative-only stories. Picks 4 and 5 (NVTS, FSLY) are the weakest fits on reverse valuation and are included only because they are the best remaining candidates in this search, which was limited and not exhaustive. Reference prices are the latest quotes retrievable at answer time (Oct 26-28, 2026).
Why each pick
AEHR
Aehr makes wafer-level and package-level burn-in test systems for AI processors, silicon photonics and power semiconductors. Management guided fiscal 2027 (June 2026 to May 2027) revenue of $130M-$150M against $50.0M in fiscal 2026, backed by about $100.6M of effective backlog as of mid-July 2026. It also guided non-GAAP pretax income of 18-22% of revenue. The stock has fallen sharply in October (RSI about 29, roughly 22% down month to date), so the setup is a guided hypergrowth step-up bought after a drawdown rather than at the peak. Reference price is the Oct 26, 2026 regular-session close of $56.60; the Oct 27 pre-market quote was $58.00. This is a company-guided inflection that has not yet been verified by a reported quarter; the first check is the Nov 2 report.
FCEL
FuelCell Energy is pivoting to data-center baseload power. It has $1.3B of committed backlog plus $2.35B of awarded capacity (not firm), a first 30 MW order inside a 380 MW Fit Energy agreement, and a 75 MW capacity reservation with a Texas colocation operator. It also has a funded plan to raise Torrington production from about 37 MW annualized (Q3 FY2026) to 100 MW in October 2026 and 500 MW by June 2028. The shares fell about 28% in October to $12.03, roughly a third of the 52-week closing high of about $36. At about $974M market cap against about $504M of reported net cash, the market assigns a modest enterprise value to the backlog and the ramp, so asymmetry exists if conversion and the ramp succeed. Dilution and negative gross margin make permanent capital loss a real risk. Reference price is the intraday Oct 28, 2026 quote; market cap is as published by the source (Oct 27 close basis).
ETON
Eton is a profitable, fast-growing rare-disease specialty pharma. Q2 2026 revenue was $37.6M (+99% year over year) with adjusted EBITDA of $16.2M (43% margin). Management raised 2026 guidance to revenue above $145M and adjusted EBITDA margin of at least 35%. Growth is driven by the HEMANGEOL relaunch, DESMODA, INCRELEX and other products, and by recently added products (IMPAVIDO launched Sep 28, 2026; an exclusive Libervant license announced Oct 14, 2026, terms not retrieved). It has the best-documented evidence of the five and positive free cash flow. The shares are about 22% below the Aug 20 close of $59.91, so some of the guidance raise has been given back. The limit on 3x is the multiple, not the operations. Reference price is the intraday Oct 28, 2026 quote ($51.38 at 12:28 PM EDT).
NVTS
Navitas has exited most mobile and low-end consumer business and is repositioning as a high-power GaN and silicon carbide supplier to AI data centers and grid and energy infrastructure. Q2 2026 revenue was $10.5M (+22% sequentially, though down from $14.5M a year earlier), with Q3 guided to $13.5M +/- $0.5M. Non-GAAP gross margin is about 39.5%, and cash is $557M after a Q2 raise of about $373M. The stock has fallen about 23% in a month and 33% in a year to $9.35 (up 11% on Oct 27). This is the weakest reverse-valuation fit of the five: the market cap is about 62x trailing sales, so 3x requires an order-of-magnitude revenue ramp. Market cap is computed from 240.6M shares at the Oct 27 close; the source rounds to $2.3B.
FSLY
Fastly's edge-cloud platform has moved from about 12% growth to 23% growth in a year, with record gross margin, six straight quarters of positive free cash flow, higher net retention (117%) and a raised full-year outlook. It is the highest-quality business of the five, and the weakest multibagger fit: the market cap is already near the $5B ceiling after about +229% in a year, so 3x needs roughly $13B. Reference price is the intraday Oct 28, 2026 quote ($26.90, -1.7%); the $4.36B market cap is as published by the source on the Oct 27 close of $27.37.
Other models on the same run
| Model | Picks |
|---|---|
| Perplexity | LIFE, DAVE, PGY, MAZE, FOUR |
| Gemini 3.6 Flash High | AEHR, SOUN, SKYT, MGNI, SERV |
| Muse Spark 1.1 | INOD, UPST, SVRA, BASE, RGTI |
| GPT-6 | PGY, ARDX, INOD, FLNC, BKSY |
| DeepSeek | AQST, CRMD, MLYS, IMNM, ACHR |
ForecastNest provides AI-generated market analysis and quantitative research for informational and educational purposes only. Nothing here constitutes financial, investment, legal, or tax advice.