CEVA earnings up next: Can AI royalties offset revenue dip?

CEVA is expected to report Q1 EPS of 2 cents on revenue of $26.14 million, up 114% and 8% year over year, but revenue is still projected to fall 16% sequentially from Q4’s $31.29 million due to seasonality. Investors are focused on whether the company’s AI licensing wins can convert into royalty revenue, with 10 NPU agreements signed last year and six customers expected to have silicon by end-2026. The setup is mixed: UBS raised its target to $42, while JPMorgan initiated at Hold and the stock already trades near its 52-week high.

Analysis

The setup is less about this quarter and more about whether the market is paying up too early for a royalty inflection that is still mostly forward-declared. In licensing-heavy semiconductor IP businesses, the first meaningful monetization step usually comes well after design wins, so the current debate is really a timing mismatch: equity is discounting AI royalty leverage now, while cash flow remains anchored to the slower-install base of design activity. That creates a fragile valuation regime where any guide-down in near-term licensing can compress the multiple quickly, even if the long-run thesis remains intact.

The second-order winners are likely the downstream customers and broader ecosystem names that can use CEVA-style IP to accelerate product roadmaps without building proprietary silicon teams from scratch. That can support smaller-edge-AI device launches at firms like MCHP, but it also means CEVA’s moat depends on breadth of deployment rather than one or two headline wins. If adoption broadens, the mix shift toward royalties should improve margin quality meaningfully over the next 4-8 quarters; if not, the company remains trapped in a low-visibility, lumpy pre-royalty phase.

The market is also underestimating how asymmetric the miss risk is versus the upside. Because expectations are already built around strong AI optionality, a normal seasonal dip or any softness in new bookings could trigger a larger de-rating than the small revenue delta implies, especially with the stock near highs and consensus target below spot. Conversely, the upside path likely requires explicit evidence that 2026 guidance is conservative and that multiple customer silicon tape-outs are converting on schedule, which is a months-long catalyst, not a days-long one.

Contrarian view: this may be one of the cleaner examples of narrative running ahead of monetization, but that does not make it a short in isolation. The better framing is that the stock is likely range-bound until the company proves royalty conversion, so the edge comes from trading volatility around the print rather than making a large outright bet on directional fundamentals.

Key Decisions for Investors

• Sell short-dated CEVA puts into the earnings event or use a put spread to express disappointment risk over the next 1-3 weeks; risk/reward favors downside convexity because expectations embed AI execution before royalties scale.
• For a lower-beta expression, pair long MCHP / short CEVA over 1-3 months to isolate broader edge-AI design-win exposure versus the more timing-sensitive IP monetization model.
• If CEVA gaps lower on any in-line print, consider buying the dip only after management confirms customer silicon timing and 2026 royalty visibility; use a staged entry over 4-8 weeks rather than catching the first move.
• Avoid chasing CEVA into the report at elevated levels; the setup offers poor asymmetry unless the company explicitly raises medium-term growth or margin guidance, which is the real catalyst needed to justify the premium.