The fact that only ALAB is falling while NASDAQ is rebounding means that the market is not stupid, but that there is a real problem with ALAB. However, the key to this report is to determine whether the problem is a structural collapse or a cyclical adjustment.
FACT ALAB is trading at $106.33 as of April 1, 2026, down -59.6% from its 52-week high ($262.90). YTD -36.1%, underperforming QQQ (-1.1%), NVDA (-2%), AVGO (+1.6%), and MRVL (+12.8%). (Source: Yahoo Finance, 2026-04-01)
FACT FY2025 revenue was $852.5M, a +115% YoY growth, but GPM is declining from 75.7% to Q1'26 guidance of 74%, and is expected to fall further to 72-73% due to Amazon warrant headwinds. (Source: ALAB Q4 FY2025 Earnings Release, 2026-02-10)
INFERENCE Market confidence has collapsed as it relies on Amazon for 70%+ of its revenue, coupled with CFO replacement and insider selling. This is not just a valuation adjustment, but closer to a narrative destruction.
NARRATIVE However, the structural tailwinds of 50%+ PCIe retimer market share, NVIDIA NVLink Fusion's only third-party partner, and $6,600-6,900B hyperscaler 2026 CAPEX have not been damaged. The problem is timing and margin, not technological competitiveness.
FACT Non-GAAP GPM Trend: Q2'25 77.0% → Q3'25 76.4% → Q4'25 75.7% → Q1'26E ~74%. It has been steadily declining by 70-130bp every quarter. (Source: ALAB Earnings Reports, each quarter)
FACT The Amazon $6.5B warrant deal results in approximately 200bp of non-cash GPM deduction per quarter. When fully reflected from Q2'26, the actual GPM will fall to 72-73%. (Source: CFO remarks, Q4 2025 Earnings Call)
INFERENCE The signal that the margin direction has turned is fatal to the market. This is because the core basis for the 30x+ Forward P/E premium that investors gave to ALAB was the software-level margin superiority. It is not the absolute drop of 3pp, but the "change of direction" itself that is the derating trigger.
FACT Amazon (AWS) accounts for more than 70% of total revenue as of FY2025. The $6.5B multi-year warrant agreement signed in February 2026 guarantees stable orders, but at the same time solidifies dependence. (Source: ALAB 8-K, 2026-02-05)
NARRATIVE The existence of the warrant deal itself proves ALAB's technology. Amazon would not have entered into a $6.5B long-term partnership if it did not intend to. The problem is that ALAB has entrusted its fate to the CAPEX priorities of one customer, and will be immediately impacted if Amazon changes its connectivity strategy.
FACT CFO Mike Tate → Desmond Lynch replacement (2026-03-02). CEO Jitendra Mohan sold 157,500 shares ($20.7M) under a 10b5-1 plan. Chief Legal Officer and Director also sold at the same time. (Source: SEC Form 4, 2026-03)
INFERENCE Since this is a 10b5-1 pre-planned sale, it is not a direct bearish signal, but if all executives sell at a time when the stock price has fallen by -59.6%, the market imagines the worst. The CEO's remaining holdings of ~6M shares ($650M+) are evidence of management's commitment, but psychological damage has already occurred.
FACT Forward P/E of approximately 30x is still a premium compared to NVDA (21x), AVGO (23x), and MRVL (25x). EV/EBITDA 94x is extreme compared to the peer average (22-60x). PEG 0.83 is good, but SBC dilution is not reflected. (Source: Yahoo Finance, StockAnalysis, 2026-04-01)
NARRATIVE The AI premium party immediately after the IPO is over. The decline from 60-80x Forward P/E in 2024-2025 to 30x is a normalization process. However, it is still uncertain whether this process is over. In an environment with interest rates of 4%+, the appropriate Forward P/E is 20-25x, which corresponds to a stock price of $80-95.
FACT FFR 3.5-3.75%, 10-year Treasury yield 4.30%, VIX 25. Oil prices $100+ due to Iran war risk. "Great Rotation" — Funds are moving from small AI growth stocks to small value stocks. (Source: Federal Reserve, Bloomberg