NVTS › analyze
NVTS · Analyze
2026-07-10 | Sector: Semiconductors (Power Semiconductors — GaN/SiC) | Price: $13.37 | Mkt Cap: ~$3.10B
Data note: Yahoo Finance MCP (Tier 1) was not connected this session. This analysis uses roic.ai (Tier 2) for financials, Alpha Vantage for insider/institutional data, and WebSearch for qualitative/news research — full tool hierarchy substitution, not a degraded-data shortcut. Also: roic.ai's annual financial history for NVTS only returns FY2024–FY2025 (2 periods) despite a 2021 SPAC listing — the 5–8yr FCF CAGR the framework calls for cannot be computed; flagged throughout as a data gap.
0. Snapshot
Navitas designs GaN (gallium nitride) and SiC (silicon carbide) power semiconductors — GaNFast/GaNSlim/GaNSafe power ICs and GeneSiC MOSFETs/modules. Historically sold into mobile fast-chargers and consumer electronics (commoditizing, low-margin); now executing a strategic pivot ("Navitas 2.0") toward AI data-center power (NVIDIA 800V HVDC ecosystem), EV, and grid/industrial — funded by a $100M Nov-2025 private placement. IPO'd via SPAC Oct 2021. 190 employees. CEO Chris Allexandre (new).
This is not a mature, cash-generative business — it is an unprofitable, revenue-declining small-cap mid-pivot, richly valued on a multi-year AI narrative. Framework §4 (Small/Mid-Cap overlay) applies heavily; a full /analyze-smallcap run would add formal red/green-flag scoring and is recommended as a follow-up (offered below).
1. Fundamentals (Fundamentals Analyst)
Revenue & Profitability
| Metric | FY2024 | FY2025 | YoY |
|---|---|---|---|
| Revenue | $83.30M | $45.92M | -44.9% |
| Gross Profit | $28.34M | $14.25M | -49.7% |
| Gross Margin | 34.0% | 31.0% | -300bps |
| Operating Income | -$129.45M | -$89.72M | +31% (loss narrowed) |
| Net Income | -$84.60M | -$116.95M | loss widened (non-cash driven, see below) |
| EPS (diluted) | -$0.46 | -$0.57 | — |
Quarterly trend (revenue actively shrinking, then stabilizing):
| Q4'24 | Q1'25 | Q2'25 | Q3'25 | Q4'25 | Q1'26 | |
|---|---|---|---|---|---|---|
| Revenue | $17.98M | $14.02M | $14.49M | $10.11M | $7.30M | $8.60M |
| Gross Margin | 12.4% | 37.9% | 16.1%⚠️ | 37.9% | 38.1% | 37.6% |
Revenue nearly halved in four quarters as the legacy mobile-charger business is deliberately wound down (low-margin, commoditizing, contested by Innoscience/Infineon pricing). Q2'25's gross margin crater to 16% looks like a one-time inventory/write-down event tied to that wind-down, not the new trend — margin has since normalized to high-30s%, which is the more representative run-rate. TTM revenue: $40.5M.
Net loss is noisy, not just operating. FY2025's headline loss (-$117M) is worse than FY2024 (-$85M) despite operating losses improving 31% — the swing is a non-operating item (+$26.9M gain in FY25 vs -$40.8M loss in FY24), almost certainly fair-value marks on warrants/earnout liabilities from the GeneSiC deal. Core operating trend is improving; GAAP net income is not a clean signal here.
Capital Allocation
Opex fell from $157.8M (FY24) to $104.0M (FY25), -34%, faster than the framework's usual "growth investment" pattern — this reads as cost discipline (R&D -34%, SG&A -44%) through most of 2025, but opex has re-accelerated in Q4'25/Q1'26 ($27.6M → $30.6M) as R&D reinvestment ramps into the AI-datacenter pivot. No M&A, no buybacks, no dividends. 100% of capital allocation is growth investment + survival (equity raises funding R&D/opex, not returned to shareholders).
Balance Sheet & Debt
- Total debt: $6.3M (capital leases only) vs. total liabilities $61.3M vs. total assets $481.4M → debt-to-assets ~1.3%, effectively unlevered.
- Cash: $221.0M (Q1'26), boosted by the $100M Nov-2025 private placement.
- Cash runway: at the current ~$16-17M/quarter burn rate, roughly 3+ years — comfortable, not urgent, but not indefinite if the AI reinvestment ramp accelerates burn further.
- ⚠️ Data inconsistency flagged: roic.ai's
book_value_per_sharefield reports -$2.44/share, which contradicts its own balance sheet data (total equity +$420.0M ÷ 232.0M shares = +$1.81/share, tangible book value ≈ +$0.90/share). This looks like a provider calculation bug (possibly mis-including a liability line as negative equity). Used the balance-sheet-derived figures throughout this report instead. - ⚠️ Altman Z-Score (9.6–16.8, "safe zone") is misleading here — it's a $3B market cap company with only $61M total liabilities, so the score is dominated by the market-value-of-equity term, not operating health. Not a meaningful distress signal in either direction for a story stock; disregarded.
Shares Outstanding — Real Dilution Headwind
| FY2024 | FY2025 | Q1'26 | |
|---|---|---|---|
| Shares out | 188.1M | 230.5M | 232.0M |
| YoY change | — | +22.5% | +23.3% (TTM) |
Revenue/share fell even faster than revenue itself: $0.453 → $0.223 (-50.7%), because share count grew alongside the top-line decline. This is growth-by-dilution risk the framework flags explicitly — funded losses, not funded growth, for now.
Scorecard: 🟡 Mixed. Balance sheet is genuinely clean (no debt, ample cash). But revenue is in outright decline, GAAP profitability is likely years away (2029E per outside analyst models), and per-share value has been diluted meaningfully. Cost discipline in 2025 is a real positive; re-accelerating opex in 2026 raises the stakes on the AI pivot actually converting to revenue.
2. Moat & Competitive Advantage (Moat Analyst)
Quantitative base: ROIC is deeply negative (pre-profit) — not a usable moat signal yet. Gross margin (~31-38% normalized) is respectable for a fabless semi but well below scaled peers (Infineon, onsemi run high-30s to 40s% at much larger volume) — no visible cost-advantage moat.
Adversarial stress-test — "how easily could a well-funded rival attack this?" Very easily, and it's already happening: - Infineon (acquired GaN Systems for $830M in 2023) and onsemi/STMicro (SiC market leaders, 22% and 29% share respectively) are multi-billion-revenue incumbents that can out-invest and out-price Navitas in both GaN and SiC simultaneously. - Innoscience (China) is scaling GaN aggressively on cost — a structural pricing threat to exactly the legacy consumer/mobile business Navitas is exiting. - Wolfspeed filed a patent infringement suit against Navitas on 2026-07-07 (Delaware) asserting five patents (8,169,005; 10,998,418; 10,886,396; 10,749,443; 11,888,392) are infringed by GaNFast, GaNSlim, GaNSafe, GeneSiC, and SiCPAK — i.e., essentially the entire current product portfolio, not one product line. Navitas disputes the claims but this is now live litigation risk sitting directly on top of the moat thesis: if Wolfspeed prevails, outcomes range from damages to forced redesigns to injunctive relief on core products. (Wolfspeed investor release, Navitas response)
Revenue-stream map: - Legacy mobile/consumer GaN (declining, being deliberately wound down, commoditizing) — durability: low. - EV/industrial SiC (GeneSiC, automotive-qualified up to 3300V) — durability: moderate, longer design-cycle lock-in once qualified, but early-stage revenue. - AI data-center power (NVIDIA MGX/800V HVDC ecosystem, Kyber/Rubin Ultra architectures) — durability: unproven, real TAM (announced May 2025, GTC demos March 2026, MGX ecosystem collaboration June 2026), but Navitas is one of several vendors in NVIDIA's multi-supplier ecosystem, not an exclusive or sole-source partner — Infineon, onsemi, and STMicro are pursuing the same design wins.
Disruption forecast: The 800V AI-datacenter power shift is a genuine, industry-wide tailwind (not narrative-only) — but it's a rising tide that lifts several boats, and Navitas's boat is by far the smallest ($40M TTM revenue vs. Infineon's ~€15B, onsemi's ~$7B). The company's edge is being an early, nimble mover with real IP — but that same IP is now contested in court by a direct competitor.
Evergreen assessment: Not yet a moat, an unproven pivot. The company has no demonstrated ability to convert design wins into durable, profitable revenue at scale; independent analyst models (SeekingAlpha) don't expect profitability until 2029.
Moat-defense evaluation: Management is litigating the Wolfspeed claim, doubling down on 800V ecosystem relationships, and cutting legacy-market costs. Reasonable moves given the hand dealt, but a $49.8M annual R&D budget cannot out-innovate Infineon or onsemi on a dollar basis — Navitas's only realistic edge is speed and specialization, both of which are now under direct legal challenge.
3. Valuation (Valuation Analyst)
Non-dividend payer → DYT and DDM are N/A. Negative EPS → Graham's formula is undefined (√ of a negative number) — explicitly N/A, not force-fit. Bogle's model needs positive earnings growth inputs — also not meaningfully applicable pre-profit.
For a pre-profit, pre-revenue-inflection name, the framework calls for multiples-vs-peers and reverse-DCF instead:
| Metric (current) | Value |
|---|---|
| Price | $13.37 |
| Market Cap | ~$3.10B |
| Enterprise Value | ~$2.89B (net cash position) |
| EV / TTM Sales | ~71x |
| P / TTM Sales | ~77x |
| P / Book | 7.4x |
| P / Tangible Book | 14.9x |
| P / FCF, FCF yield | negative (FCF burning) |
This is priced like a hyper-growth story stock, not a semiconductor company with declining trailing revenue. For comparison, established SiC/GaN players (Wolfspeed, onsemi, STMicro) trade at low-single-digit to low-teens EV/Sales even with real, larger revenue bases.
Reverse-DCF framing: to normalize toward a still-generous 8-10x forward sales multiple (typical ceiling for a scaling, profitable-adjacent semi story), revenue would need to reach roughly $290-360M — a 7-9x scale-up from today's $40.5M TTM base. That is the size of the bet embedded in the current price.
Scenario-based fair value range (no clean point estimate is honest here): - Bear (AI ramp stalls, legacy decline continues, litigation goes badly): re-rates toward net-cash-plus-optionality floor, roughly $5-7/share. - Base (modest AI/EV revenue ramp over 2-3 years, legacy stabilizes, litigation resolved without material damages): $10-16/share — broadly where it trades today. - Bull (800V design wins convert at scale, litigation immaterial): SeekingAlpha's $19 PT (2026-07-07, "Buy") models ~$0.36 incremental EPS by 2029 from AI/grid/industrial, implying the stock is a bet on a 2029 profitability story. (Seeking Alpha)
Current price sits roughly mid-range between bear and bull — the market has already priced in a decent probability of success, so there's no obvious margin of safety at $13.37. Conservative bias per the framework says: this is not a "cheap, great company" — it's a "not-yet-proven company, not particularly cheap for what's proven so far."
4. Sentiment & Intelligence (Sentiment Analyst)
- Extreme volatility: +151% in H1 2026, then -41% in the trailing month as revenue-decline/valuation concerns resurfaced, then another leg down on the Wolfspeed suit (-9% on 2026-07-08). Stock is currently well off its ~$17 recent highs.
- Analyst consensus is "Hold," with unusually wide target dispersion ($9.49–$21 depending on source/coverage set) — genuine disagreement, not false precision. No source shows strong conviction in either direction.
- Institutional positioning is a mild positive: 343 institutional holders, 54% institutional ownership, 190 holders increased positions vs. 103 decreased in the latest period — net accumulation despite the volatility.
- Legal overhang is the dominant near-term risk factor: the Wolfspeed suit (filed 2026-07-07) targets essentially the full product line and will likely be a multi-quarter-to-multi-year distraction regardless of merits.
- Genuine tailwind, not pure hype: NVIDIA MGX/800V ecosystem collaboration (announced June 2026), GTC demos (March 2026), and a $1-1.8B SiC grid-upgrade opportunity flagged by Zacks (2026-06-23) are real, sourced developments — but retail-facing headlines ("beating Nvidia," "beating Broadcom" in 2026 performance) for a $3B/$40M-revenue company are a valuation-disconnect flag worth naming explicitly.
- Near-term catalyst: Q2 2026 earnings report on 2026-07-27 — first real read on whether AI/data-center/EV revenue is offsetting the legacy decline, and management commentary on the Wolfspeed suit.
Sources: Wolfspeed lawsuit, Navitas response, NVIDIA MGX collaboration, Q2 earnings date, Motley Fool — H1 2026 performance, Zacks — grid opportunity, Seeking Alpha — price target
5. Synthesis — Weighted Verdict
Is this a good business? Not yet demonstrated. Revenue is actively declining (-44.9% YoY), profitability is likely 3+ years away, and the moat thesis (proprietary GaN/SiC IP) is under direct legal challenge covering nearly the full product line. The one clear positive is balance-sheet discipline: near-zero debt, $221M cash, comfortable multi-year runway, and net institutional accumulation.
Has the market priced that in? No — if anything the market is pricing a successful, multi-year pivot (71x EV/Sales) that hasn't happened yet. There's no valuation cushion for the bear case (litigation loss, AI ramp disappoints, legacy decline outpaces new-market growth).
Per framework §0: this reads as Poor-to-unproven business quality + Expensive pricing — the least attractive quadrant, tempered only by the real (not narrative-only) AI-datacenter tailwind and clean balance sheet giving it optionality rather than distress risk.
Verdict: Speculative / High-Risk. Conviction: [4.0] — comparable tier to other speculative watchlist names (EVGO [4.5], QBTS [5]), weighted slightly lower here because revenue is already declining (not merely pre-scale) and there's live, portfolio-wide patent litigation, partially offset by materially cleaner balance sheet/liquidity than EVGO.
Key risks (named, not buried): 1. Wolfspeed patent litigation — targets the full product line; outcome/timeline unknown. 2. Revenue still declining as of the most recent quarter; AI/EV revenue hasn't yet visibly offset the legacy wind-down. 3. Extreme valuation (71x EV/Sales) leaves no margin of safety if the pivot stalls. 4. Navitas is the smallest player chasing the same NVIDIA-ecosystem design wins as Infineon, onsemi, and STMicro — all far better capitalized. 5. 23%+ YoY share dilution funding the pivot.
Not a buy at $13.37. Recommended posture: Watch, not own. Reassess after (1) Q2 2026 earnings (2026-07-27) show whether AI/EV revenue is offsetting legacy decline, (2) clarity emerges on the Wolfspeed suit (dismissal, settlement, or damages/injunction risk becomes quantifiable), (3) gross margin holds a stable mid-30s%+ for 2+ consecutive quarters. A pullback toward the $5-8 bear-case floor (net-cash-plus-optionality) would be a far more asymmetric entry than current levels.
A portfolio-specific passage was removed from the public build.
Recommendation: Adding to Watchlist.md as a new speculative Semiconductors watch entry at [4.0], entry zone flagged for the bear-case floor rather than current price. A follow-up /analyze-smallcap NVTS would formalize the red/green-flag scoring (dilution trajectory, burn trajectory, catalyst likelihood×impact) this report only partially covers — worth running before any real capital commitment.