AIP · Arteris, Inc.

Current Thesis

Arteris licenses the network-on-chip (NoC) interconnect IP that sits inside SoCs, and collects a variable royalty every time a licensed chip ships. Royalties lag design wins by ~18–24 months, so the 2023–24 custom-silicon and ADAS design-win wave is only now hitting the P&L. The Q1 2026 print (reported 2026-05-12) is the proof: revenue $22.9M (+39% YoY), variable royalties +67% YoY, ACV+royalties $92.8M (+39%), and a raised FY26 guide. The stock did what an accelerating royalty story does it re-rated from ~$19.56 in early April to a $38.99 all-time high, with Rosenblatt and Jefferies hiking targets.

That is the catch: the easy leg (the undiscovered-small-cap re-rate) has already happened, and sell-side has caught up. Fundamentals are ACCELERATING, but the trade is now a quality-momentum continuation at ~2x off the lows, not a hidden-gem entry. A fresh long here rates MEDIUM the cleaner setups are a pullback that holds the rising 20-EMA in the low $30s, or a clean breakout back above the $38.99 ATH on expanding volume.

Bull Case

• Q1 2026 (reported 2026-05-12): revenue $22.9M, +39% YoY, beating consensus ~$21M by ~8% acceleration from the FY25 +22% pace, and above the $21.5M bull line set pre-print.
• Variable royalties +67% YoY (TTM $7.9M) this is the lead KPI, and it accelerated hard. Royalties are the cleanest read on prior design wins converting to volume; +67% confirms the design-win cycle turned.
• ACV + royalties $92.8M (+39%), RPO $118.3M (+33%) contracted backlog growing roughly in line with revenue gives forward visibility, not a one-quarter spike.
• FY26 guide raised to $91–95M with positive free cash flow; Q2 guide $23–24M a beat-and-raise quarter, the setup momentum books want.
• Design wins disclosed on the call (2026-05-12): a leading global hyperscaler, a top HBM memory supplier, a MIPS collaboration for physical-AI, and NoC IP flying in NASA Artemis lunar missions breadth across data-center, memory, and automotive/edge.
• Analyst re-rate: Rosenblatt to $38 from $20 (2026-05-13, Buy); Jefferies to $35 from $16 the kind of cluster of upward revisions that confirms a narrative going mainstream.
• Path to profitability: ~90%+ non-GAAP gross margin, non-GAAP operating loss narrowed to $2.5M, with the 4B+ deployed-chips milestone (2026-02-12) underpinning the royalty annuity.

Bear Case

• Up ~2x in roughly seven weeks (~$19.56 early April → $38.99 ATH), now $34.28 (2026-06-05) after a -7.8% session the price is extended and mean-reversion has started.
• The "undiscovered" edge is spent. Record-high headlines, a near-120% PT hike, and rising StockTwits chatter mean the narrative is going public. That is late-stage for the easy money, not early.
• Valuation re-stretched. At ~$34 on ~44M shares (~$1.5B cap), the name trades near ~16x FY26 revenue midpoint expanded from ~9.6x pre-rerate. Any guide wobble re-rates it hard in both directions.
• GAAP still loss-making: Q1 GAAP operating loss $9.3M (inflated by Cycuity acquisition items). No durable pricing-power premium yet.
• Concentration and lumpiness: license revenue is lumpy quarter-to-quarter, and variable royalties lean on a handful of auto Tier-1s and hyperscaler reporters a single slow customer swings the lead KPI.
• Competitive overhang: Synopsys (ARC, foundation IP) and hyperscalers' internal fabrics can pressure NoC mindshare on the highest-value sockets.
• Management transition: CFO Nicholas Hawkins retires 2026-08-31 (announced 2026-05-08), right as the company targets non-GAAP operating profit.

Setup & Price Structure

• Closed $34.28 on 2026-06-05, -7.82% on the day; after-hours $33.62; day range $33.24–$35.94 on ~862K shares. 52-week range $7.56–$38.99.
• Price is ~12% off the $38.99 ATH and ~2x off the early-April ~$19.56 level. The rising 20-EMA sits in the low $30s; the post-Q1 breakout shelf and rising 50-day cluster around ~$28; the $38.99 ATH is the breakout line.
• Read: fundamentals ACCELERATING, price extended and digesting. The high-quality fresh entries are (a) a pullback that holds the rising 20-EMA in the low $30s with the breakout base intact, or (b) a clean reclaim above $38.99 on expanding volume. Buying the gap straight into the prior ATH after a double is the lower-quality version of this trade.
• Small float means ±8% candles are routine the June 5 drop is noise within the trend, not yet a structural break. Size for that volatility.

Catalyst Calendar (next 30 days)

• No confirmed hard catalyst inside the 2026-06-06 → 2026-07-06 window. The name is between prints; near-term moves will be flow- and sympathy-driven off the AI-semi complex.
• ~2026-06 (est.): summer investor-conference season small-cap semis routinely present at June tech conferences; any appearance is incremental, not binary.
• Ongoing: Arteris issues design-win / partnership press releases frequently; on a ~44M-share float these can move the tape intra-quarter.
• ~2026-08-11 (est.): Q2 2026 print the next true binary, outside this 30-day window. Watch revenue vs. the $23–24M guide and variable-royalty YoY growth vs. the +67% just posted.
• 2026-08-31: CFO retirement effective; a successor announcement could land any time before and is a management overhang to track.

What Would Change Our Mind

• Variable-royalty YoY growth decelerating below ~+25% on a future print (vs. +67% in Q1) the lead KPI breaking is the cleanest thesis-kill.
• Q2 revenue below the $23M low end of guide, or a guide cut at ~16x sales the punishment is asymmetric.
• Weekly close below the ~$28 post-Q1 breakout shelf / rising 50-day that signals the momentum leg is finished; re-engage only on a fresh higher-low or a >$39 — reclaim, never by adding into the decline.
• A customer-loss or peer-failure event Synopsys taking a flagged hyperscaler socket, or a hyperscaler moving NoC in-house.
• Theme flip to SATURATED mainstream CNBC coverage plus broad retail ownership with price >50% above the 50-day and no new design-win catalyst to feed it.

Correlation Notes

• AIP is a second-derivative play on the same custom-silicon / AI-accelerator buildout that drives the large-cap AI-chip complex (TSMC, NVIDIA, Broadcom, Marvell, the HBM names). Its royalties literally accrue from chips those companies ship, so it carries high beta to the group and, given the thin float, amplifies the cluster's moves in both directions.
• Strongest read-through is from custom-silicon/ASIC names (Broadcom, Marvell) and from memory the disclosed top-HBM-supplier design win ties part of the royalty stream to the memory cycle.
• Automotive-AI/ADAS exposure (Renesas R-Car, NXP) is a partly-decorrelated royalty driver versus pure data-center, which slightly softens the single-factor risk.
• For a book already loaded with large-cap AI-semi exposure, AIP concentrates the same factor rather than diversifying it it should be sized as correlated cluster risk, not an independent bet.