Looking at the raw quarterlies first: Cerus posted $53.7M in Q1 2026 vs $43.2M Q1 2025 — that's 24.3% YoY, an acceleration from the 14% trailing CAGR the momentum module flagged. Q4 2025 hit $57.8M (+13.8% YoY), Q3 was effectively breakeven at -$19K, and the full-year 2025 came in at $206.1M vs $180.3M (+14.3%). Gross margin held at 53.5% — flat, not expanding meaningfully (2024 gross was 55.2%, actually slightly down). The synthesis claim of "margins improving 400+ basis points annually" doesn't square with the gross line; it must be referencing operating leverage, but operating loss actually *widened* from -$14.2M in 2024 to -$36.3M in 2025. That's a meaningful deterioration the synthesis glosses over. Net loss did improve ($15.6M vs $20.9M) but only because of non-operating items. So the "path to profitability" narrative is shakier than the prior models present.

The cash flow picture is the genuinely interesting fact and where I partially side with the bulls: $11.4M operating CF and $8.5M FCF in 2025, first time meaningfully positive. But cash on hand is just $20.0M with no disclosed debt figure (a red flag in itself — Cerus historically carries ~$60-70M in term debt; absence here means the data is incomplete, not that the balance sheet is clean). At a $36M operating loss run-rate with only $20M cash, the gap between GAAP operating loss and positive FCF is being bridged by working capital and stock-based comp. That's exactly what the Market Forces module flagged as "poor cash conversion" and a "binary bet" before dilution — and I think that module is closer to right than the synthesis verdict of "Reasonable Premium." A capital raise at $2.72 with 200M+ shares outstanding would be dilutive and ugly.

Insider activity tells a clearer story than "Neutral." On 2026-03-12 alone, insiders dumped roughly 943,000 shares across five separate sales, plus another ~116,000 on 03-05. That's over a million shares sold in a one-week window against just 463,854 in awards. Calling that "neutral" is generous — it's net selling at depressed prices ($2.72, near 52-week lows per the pre-flight note), which insiders typically avoid unless they have to. Either liquidity needs are pressing or they don't see a near-term catalyst. The contrarian counter is that 10b5-1 plans and tax-related sales on award vesting can mechanically produce this pattern, and the absolute dollar value (~$3M) is small for a company this size. But the pattern doesn't support the bull case.

I dissent from the "Reasonable Premium" synthesis verdict and lean toward Market Forces' framing. The 1.9x P/S looks cheap for medtech only if you believe the 24% Q1 acceleration is sustainable AND that operating losses compress without a raise. Both are unproven. INTERCEPT platelets is a 20-year-old product; the RBC program is the optionality, but it's been "coming soon" for a decade. At $2.72 and $0.55B market cap, the stock is pricing in execution risk fairly — it's neither obviously cheap nor obviously expensive. Fair value in a no-raise, sustained-20%-growth scenario is probably $3.50-4.00; in a dilutive-raise scenario, $2.00-2.25. The asymmetry isn't compelling at current price given $20M cash against $36M operating losses. I'd want to see Q2 2026 confirm the acceleration AND a financing event (or clear evidence one isn't needed) before committing capital. The prior models are too sanguine about the balance sheet fragility.