| Metric | COHR | MRVL | LITE | AAOI |
|---|---|---|---|---|
| Alpha v2 | 78 (#1) | 76 (#2) | 70 (#3) | 62 (#4) |
| Price | $248.53 | $89.61 | $653.17 | $112.23 |
| Market Cap | $39.4B | $78.3B | $46.6B | $8.4B |
| Rev Growth (latest Q) | +17% YoY | +22% YoY | +65.5% YoY | +34% YoY |
| Revenue Run Rate | ~$6.8B | ~$8.9B | ~$3.2B | ~$600M |
| GAAP Profitable? | Yes | Yes | Yes (recent) | No |
| NVDA Strategic Deal | $2B ✓ | Celestial AI acq | $2B ✓ | None |
| S&P 500 | Mar 23 ✓ | Already in | Mar 23 ✓ | No |
| Product Differentiator | 6" InP wafers + OCS | DSP + custom ASIC + photonic fabric | Lasers + OCS + CPO | In-house InP lasers + CATV |
| Competitive Position | Vertically integrated | Platform (DSP brains) | Component supplier | Cost-competitive 2nd source |
| Fwd P/S | ~5.5x | ~8.8x | ~14x | ~8.4x |
| % Below 52W High | -17.2% | -12.9% | -16.7% | -13.0% |
| Beta | 2.11 | 1.85 | 1.41 | 3.25 |
| Dilution Risk | NVDA placement + Bain | Low | Low | $250M ATM active |
| 1-Year Return | +273% | +90% | +854% | +1,056% |
| SIGNAL | HOLD | RE-ENTER | WAIT $550-600 | AVOID |
COHR leads on vertical integration — they make the InP lasers, the optical engines, AND the transceivers. The 6-inch InP wafer fab (vs industry standard 3-inch) produces 4x more chips at less than half the cost. This is a structural manufacturing advantage that takes 3+ years and $500M+ to replicate. The NVDA $2B equity investment converts a customer relationship into a strategic partnership.
MRVL owns the DSP brains (Electra coherent DSP on 2nm) that go inside every coherent transceiver — including COHR's and LITE's products. The Celestial AI acquisition adds photonic fabric for intra-rack optical interconnect. Custom ASIC design wins (Trainium, Maia) create multi-year revenue lock-in. Platform moat, not component moat.
LITE makes the lasers and optical circuit switches. The NVDA deal is identical in structure to COHR's. Strong technology but competing against COHR, Broadcom, and Intel's silicon photonics efforts. Less vertically integrated than COHR — more of a component supplier than a system integrator.
AAOI is a cost-competitive second source. In-house InP lasers are real, but the company lacks the breadth, the hyperscaler design wins at COHR/LITE's scale, and critically, NVDA didn't invest in AAOI — they chose COHR and LITE. Jensen's capital allocation is the market telling you who the long-term partners are.
COHR has the densest catalyst stack: OFC Innovation Briefing (Mar 17), S&P 500 inclusion (Mar 23), GTC product mentions, May 13 earnings, 1.6T cycle ramp, Bain overhang resolution. Multiple new-information events front-loaded in the next 60 days.
MRVL just printed a record quarter and guided above consensus. COLORZ 1600 launch at OFC, Celestial AI integration milestones, custom ASIC ramp updates, next ER ~June. The Celestial AI revenue trajectory ($500M run-rate by end FY28, $1B by FY29) is the multi-year catalyst queue.
LITE has the same S&P 500 + NVDA + OFC stack as COHR, plus the Rosenblatt $900 PT generating attention. But the biggest catalyst (Q3 ER showing $780-830M revenue) is ~2 months out. The near-term catalysts are known and partially priced after the +11% S&P inclusion day.
AAOI has the $200M 1.6T order as a confirmed catalyst, but the 800G firmware qualification delay (pushed from Q4 to Q1 to Q2) is a negative catalyst risk. Next ER April 30. The 1.6T order is priced in after the $54→$129 run.
MRVL has the most remaining edge. The Celestial AI photonic fabric thesis is genuinely non-consensus — most analysts model MRVL as a custom ASIC company, not a photonic interconnect platform. The $2.4B guide beat repriced the near-term, but the Celestial AI revenue ramp ($0 now → $1B/yr by FY29) is not in most models.
COHR has moderate remaining edge. The 6-inch InP advantage and >4x data center book-to-bill are known to the sell-side, but the magnitude of the capacity ramp and the margin expansion trajectory from 39% to potentially 45%+ as InP scales are not fully modeled. Bain selling has masked the fundamental re-rating.
LITE has limited remaining edge. Up 854% in a year. Every sell-side shop has published the "LITE is the next photonics winner" thesis. Rosenblatt's $900 PT is the most aggressive, but the consensus move has already happened. Retail sentiment is +41, smart money is -7 per the Signal Engine. Classic edge exhaustion.
AAOI has the least remaining edge. Up 1,056% in a year. The short squeeze is spent. The $1B revenue target is the consensus bull case. The $250M ATM dilution signals management agrees the stock is fully valued. At $112 on negative earnings, you're paying for a future that everyone can see.
COHR & MRVL benefit from moderate beta (2.1 and 1.85) relative to AAOI's 3.25. In a VIX 22+ risk-off tape, lower-beta names hold better. Both have S&P 500 passive buying (COHR upcoming, MRVL already in) as structural demand floor. Iran conflict drives distributed datacenter buildout which requires more optical interconnects.
LITE & AAOI suffer in this regime. LITE at $653 and 164x trailing P/E is the most valuation-exposed to a rate-driven multiple compression. AAOI at beta 3.25 with no profitability and an active ATM offering is the highest-risk profile in a risk-off environment. A broader market correction would hit AAOI 2-3x harder than COHR.
COHR (#1, Alpha 78) is the best risk-adjusted position in the photonics complex. Vertically integrated, NVIDIA-validated, S&P 500 inclusion in 12 days, OFC briefing in 5 days, and the stock is 17% below its ATH after a pullback that created the entry the framework demands. You already own 1,000 shares at $207. Hold through the catalyst stack. Do not sell.
MRVL (#2, Alpha 76) is the re-entry candidate you should have paired with a trigger when you sold. The Celestial AI photonic fabric thesis is the most differentiated moat in the group — nobody else can put optical interconnects INSIDE the processor package. At $89.61, the post-earnings gap has been partially absorbed. The COLORZ 1600 launch at OFC next week is the near-term catalyst. A 2-3% position (~$150-200K) re-establishes Ph1 exposure in the DSP/ASIC layer that no other holding covers.
LITE (#3, Alpha 70) is the right thesis at the wrong price. Same NVIDIA deal as COHR, same S&P 500 inclusion, same photonics bottleneck thesis. But at 14x forward P/S and 164x trailing P/E with 854% 1-year return, the edge has decayed to the point where R/R is symmetric (1.07:1 per the Signal Engine). Wait for $550-600 pullback post-S&P inclusion fade, then initiate a 2% starter.
AAOI (#4, Alpha 62) is the stock that already moved. From $9.71 to $128.96 in 12 months, now pulling back 13% in a single session. No NVIDIA strategic deal. No S&P 500 inclusion. $250M ATM dilution active. 17.8% short interest with the squeeze spent. The $200M 1.6T order is real, but it's priced in after a 10x run. At $65-70, this becomes interesting. At $112, the asymmetry is gone.
Portfolio Action Summary:
COHR: Hold 1,000 shares through OFC + S&P inclusion. Do not trim.
MRVL: Re-enter 1,500-2,000 shares at $88-92. Define paired exit trigger this time.
LITE: Set limit order at $575. Patient.
AAOI: No action. Revisit at $65-70 if macro correction creates the entry.