3 Incredible Growth Stocks to Buy Now

Three small-cap plays on 'physical AI' are highlighted: Ondas reported Q3 2025 revenue of $10.1m (up 582% YoY), raised 2025 guidance to $36m and targets $110m in 2026 with positive EBITDA expected H2, holds $433m cash and won a large government autonomous-border drone tender with first PO expected Jan 2026 (but has raised ~$855m since June). Serve Robotics posted ~210% Q3 revenue growth, has completed >100,000 deliveries from 2,500+ restaurants, struck a DoorDash deal breaking exclusivity with Uber, and aims to deploy 2,000 robots U.S.-wide by year-end amid opaque unit economics and regulatory risk. Rigetti posted Q3 revenue of $1.9m versus a $20.5m operating loss, has two on‑premise orders totaling $5.7m plus a $5.8m Air Force Research Lab contract, but trades at a frothy ~$9.2bn market cap (~5,000x annualized revenue), leaving valuation and execution risk high.

Analysis

Market structure: Physical-AI winners are niche hardware integrators (ONDS, SERV) and semiconductor suppliers (NVDA, specialty analog/psu vendors) while legacy logistics layers and incumbents that rely on cheap human labor face margin squeeze. Ondas' government win (PO expected Jan 2026) can create multi-year recurring revenue if executed; Serve's DoorDash deal shifts demand concentration from Uber to a two-hub model that amplifies scale benefits for a single robot fleet operator. Supply constraints will center on GPUs, Li-ion cells and RF components — expect tighter spot markets for semiconductors and batteries into 2026 if deployments accelerate.

Risk assessment: Tail risks include cancelled government procurements, rapid regulatory bans on sidewalk robots in major metros, or a competitor quantum breakthrough that re-rates RGTI negatively; probability low but impact >50% equity loss for single names. Near-term (weeks) volatility will spike around contract/PO dates and quarter-ends; medium-term (3–12 months) risks are dilution (ONDS raised ~$855M since June), unit-economics revelations (SERV) and cash burn vs milestone delivery (RGTI). Hidden dependencies: prime-contractor status exposes ONDS to subcontractor supply chain failure and political procurement shifts.

Trade implications: Tactical exposures should be asymmetric, option-like bets sized small. Prefer a measured long exposure to ONDS ahead of Jan 2026 PO (1–2% portfolio risk) and a calibrated long to SERV (1%–1.5%) via call spreads to limit downside; express negative view on RGTI via limited-risk put spreads (0.5%–1%). Rotate 0.5–1% from pure software AI into NVDA and specialty materials (Li, Cu) to capture hardware demand tailwinds.

Contrarian angles: Consensus underestimates operational complexity and insurance/maintenance economics — fleet scale may take 2–4x longer to reach attractive unit economics. RGTI’s valuation appears frothy (≈5,000x revenue) — the market is pricing a near-certain quantum win rather than optionality; that is likely overdone. Conversely, ONDS may be underpriced if PO converts to multi-year recurring revenues; downside from dilution is real but capped if orders materialize.

Key Decisions for Investors

• Establish a 1.5% portfolio long position in ONDS ahead of the expected Jan 2026 purchase order: use common shares or a 12-month call spread to limit downside. If the PO is announced and revenue guidance to 2026 remains ~ $110M, add to 3–4%; if no PO by Feb 15, 2026, cut position to zero.
• Build a 1.0–1.5% long in SERV via 9–12 month call spreads (defined risk) to capture DoorDash volume ramp; tranche in now (50%) and add if Serve reports >150% YoY deliveries next quarter or robot fleet reaches 2,000 by Dec 31, 2025. If city-level bans or adverse regulatory rulings hit top-5 metros, exit immediately and cap loss at 35%.
• Express a cautious negative on RGTI with a 6–9 month put spread sized 0.5–1% (buy puts, sell lower strike puts) to limit capital at risk; target 40–60% downside if next two quarters show continued >$15M monthly cash burn or no material revenue >$10M/quarter. Avoid outright long exposure unless buying tiny (0.25%) long-dated OTM calls as lottery-ticket optionality.
• Reallocate 0.5–1.0% from long-only software AI into NVDA and materials (lithium, copper suppliers) to capture hardware demand; hedge increased equity volatility by selling 1–2 week ATM straddles on SERV/RGTI around major catalyst dates (earnings, PO announcements) to monetize spikes.