
Sell, $33.33 target: the InP shortage is real, but the price already pays for it
We initiate at Sell with a $33.33 target, 14.3% below the $38.88 close. AXT makes the indium phosphide (InP) wafers under AI optical lasers, and demand exceeds supply. Our FY27E EPS of $2.63 matches consensus, but the price pays for a 54% mid-cycle gross margin where we model 40%.
- Mid-cycle gross margin is 40%, not the 54% the price implies. History runs 12.7% to 43.0%; worth $3.41 a share (Debate 1).
- Pricing falls from 1.05x to 0.73x by FY31E as JX, Sumitomo and PRC capacity arrive; worth $3.25 a share (Debate 2).
- The Tongmei round costs $1.12 a share if no listing follows, as investors can redeem at cost plus 8% a year (Debate 3).
What the market is missing. FY26E revenue of $237.9m is in line with $239.4m consensus. The price solves to a 53.8% mid-cycle margin or a 9.5% cost of equity; we use 40% and 11.9%.
Catalysts. 27 Nov 2026 expiry of China’s US-specific gallium and germanium suspension; Feb 2027 4Q26 results (guided ~$90m); mid-2027 InP capacity doubling.
Valuation and target. DCF base $31.05, mid-cycle P/E cross-check $27.93. Bear $10.70 (30%), base $31.05 (50%), bull $55.27 (20%) weight to $29.79, rolled to $33.33.
Risks. Upside: capacity beating plan (+$2.27 a share), Tongmei listing (+$1.12). A 1pt lower cost of equity gives $35.84 (−7.8%, Hold).
| Year end 31 Dec, $m | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Revenue | 99.4 | 88.3 | 237.9 | 509.9 | 717.0 |
| Growth y/y | n/a | (11.1%) | 169.4% | 114.3% | 40.6% |
| EBITDA | (5.8) | (12.9) | 91.7 | 261.3 | 360.3 |
| Margin | (5.9%) | (14.6%) | 38.5% | 51.2% | 50.2% |
| Net income to AXT | (11.6) | (21.3) | 72.5 | 169.2 | 223.2 |
| EPS, GAAP ($) | (0.27) | (0.49) | 1.13 | 2.56 | 3.38 |
| EPS growth | n/a | n/a | n/a | 125.7% | 32.0% |
| FCF incl. prepayments | (17.9) | (18.8) | n/a | 84.7 | 114.3 |
| At $38.88 | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| P/E, non-GAAP | n/m | n/m | 32.4x | 14.8x | 11.3x |
| P/E, GAAP | n/m | n/m | 34.3x | 15.2x | 11.5x |
| EV/EBITDA | n/m | n/m | 19.8x | 7.0x | 5.0x |
| EV/sales | 18.3x | 20.6x | 7.6x | 3.6x | 2.5x |
| FCF yield | (0.7%) | (0.7%) | n/a | 3.3% | 4.4% |
| Dividend yield | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| P/B, total equity | n/m | 7.7x | 1.6x | 1.4x | 1.2x |
| P/E at our target | n/m | n/m | 27.8x | 12.7x | 9.7x |
Source: Company filings (A); Penmark Research model (E). EV before Tongmei round. FCF incl. customer prepayments; FY26E n/a (3Q26 cash flow not filed). Shaded = estimates.
Base value is $31.05 and the weighted target $33.33: upside is $22.97 against $26.90 of downside
Source: Yahoo Finance closes; Penmark Research model. Scenario targets are model values rolled forward 12 months at the 11.9% cost of equity.
InP stays short into FY29, capacity 2.2x in FY27, US permits granted, Tongmei listed; mid-cycle margin 46%.
Capacity 2.0x in FY27, utilisation 97% to 82%, price index 1.05 to 0.73, margin 60% to 40% mid-cycle.
Export permits squeezed, FY27 utilisation 75%, price index 0.55 by FY30, mid-cycle margin 29.8%, 60% no listing.
| Income statement, $m | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Revenue | 99.4 | 88.3 | 237.9 | 509.9 | 717.0 |
| Gross profit | 23.8 | 11.2 | 128.5 | 305.9 | 401.5 |
| EBITDA | (5.8) | (12.9) | 91.7 | 261.3 | 360.3 |
| EBIT | (14.8) | (22.0) | 81.3 | 243.9 | 331.5 |
| Net financial items | 4.1 | 0.4 | 17.3 | 28.4 | 29.7 |
| Tax | (1.1) | (1.7) | (13.9) | (40.9) | (54.2) |
| Net income, GAAP | (11.6) | (21.3) | 72.5 | 169.2 | 223.2 |
| Net inc., non-GAAP | n/a | n/a | 76.5 | 173.7 | 228.0 |
| Cash flow, $m | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Net income, consol. | (11.8) | (23.2) | 84.8 | 231.5 | 307.1 |
| D&A and SBC | 12.1 | 12.4 | 14.3 | 21.8 | 33.5 |
| Working capital, other | (12.4) | (2.0) | n/a | (8.7) | (76.3) |
| Operating cash flow | (12.1) | (12.8) | n/a | 244.7 | 264.3 |
| Capex | (5.8) | (6.0) | n/a | (160.0) | (150.0) |
| FCF incl. prepay. | (17.9) | (18.8) | n/a | 84.7 | 114.3 |
| Customer prepay. | n/a | n/a | n/a | 20.0 | (15.0) |
| FCF excl. prepay. | n/a | n/a | n/a | 64.7 | 129.3 |
| Balance sheet, $m | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Cash, investments | 23 | 128 | 1,384 | 1,468 | 1,583 |
| Receivables | 26 | 27 | 48 | 77 | 108 |
| Inventory | 85 | 82 | 104 | 112 | 156 |
| PP&E, net | 160 | 162 | 217 | 359 | 480 |
| Total assets | 339 | 434 | 1,807 | 2,071 | 2,381 |
| PRC bank loans | 47 | 63 | 72 | 72 | 72 |
| Net cash | (24) | 66 | 1,311 | 1,396 | 1,510 |
| Total equity | 255 | 335 | 1,572 | 1,808 | 2,120 |
| Per share and KPIs | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Diluted shares, m | 43.2 | 43.9 | 62.8 | 66.1 | 66.1 |
| EPS, non-GAAP ($) | n/a | (0.42) | 1.20 | 2.63 | 3.45 |
| EPS, GAAP ($) | (0.27) | (0.49) | 1.13 | 2.56 | 3.38 |
| Dividend/share ($) | 0.00 | 0.00 | 0.00 | 0.00 | 0.00 |
| FCF/share ($) | (0.41) | (0.43) | n/a | 1.28 | 1.73 |
| Equity/share ($) | 5.91 | 7.62 | 23.79 | 27.36 | 32.08 |
| InP revenue ($m) | n/a | n/a | 169.5 | 433.2 | 633.6 |
| InP utilisation | n/a | n/a | n/a | 97% | 95% |
Source: Company filings (A); Penmark Research model (E). $m unless stated; FY ends 31 December. Cash includes restricted cash and investments; total equity includes non-controlling and redeemable interests. FY26E cash flow n/a: 3Q26 cash flow not filed at the model date. Bold = totals and thesis lines.
Returns turn positive in FY26E and we are within 2% of consensus on EPS; the debate is margin, not volume
| Growth, margins (%) | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Growth y/y | |||||
| Sales | n/a | −11.1% | 169.4% | 114.3% | 40.6% |
| EBITDA | n/a | n/a | n/a | 185.1% | 37.9% |
| EBIT | n/a | n/a | n/a | 200.0% | 35.9% |
| EPS, GAAP | n/a | n/a | n/a | 125.7% | 32.0% |
| DPS | n/a | n/a | n/a | n/a | n/a |
| Margins | |||||
| Gross margin | 24.0% | 12.7% | 54.0% | 60.0% | 56.0% |
| EBITDA margin | −5.9% | −14.6% | 38.5% | 51.2% | 50.2% |
| EBIT margin | −14.9% | −24.9% | 34.2% | 47.8% | 46.2% |
| FCF margin | −18.0% | −21.3% | n/a | 16.6% | 15.9% |
| Capex / sales | 5.8% | 6.8% | n/a | 31.4% | 20.9% |
| Ratios and valuation | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Returns and leverage | |||||
| ROE | n/a | n/a | 8.9% | 13.7% | 15.6% |
| ROA | n/a | −6.0% | 7.6% | 11.9% | 13.8% |
| ROCE | n/a | n/a | 8.0% | 13.8% | 16.3% |
| Net debt / equity | n/a | −19.6% | −83.4% | −77.2% | −71.2% |
| Net debt / EBITDA | n/m | n/m | −14.3x | −5.3x | −4.2x |
| Valuation at $38.88 | |||||
| P/E, non-GAAP | n/m | n/m | 32.4x | 14.8x | 11.3x |
| EV/EBITDA | n/m | n/m | 19.8x | 7.0x | 5.0x |
| P/B, total equity | n/m | 7.7x | 1.6x | 1.4x | 1.2x |
| FCF yield | (0.7%) | (0.7%) | n/a | 3.3% | 4.4% |
| Dividend yield | 0.0% | 0.0% | 0.0% | 0.0% | 0.0% |
| P/E at our target | n/m | n/m | 27.8x | 12.7x | 9.7x |
Source: Penmark Research model; company filings. Ratios derived from model statements: ROE, ROA, ROCE on average total equity, assets and equity plus bank debt; net debt negative = net cash. Multiples at the 6 Nov 2026 price of $38.88 except the last row (at our $33.33 target). P/E non-GAAP excludes SBC; FY27E EV/EBITDA uses EV before the Tongmei round.
| $m; EPS $ | 4Q26E | FY26E | FY27E |
|---|---|---|---|
| Revenue ($m) | |||
| Penmark estimate | 90.0 | 237.9 | 509.9 |
| Consensus | 91.5 | 239.4 | 516.7 |
| Prior-year period | 23.0 | 88.3 | 237.9 |
| Diff. vs consensus | (1.7%) | (0.6%) | (1.3%) |
| Growth y/y (Penmark) | 290.6% | 169.4% | 114.3% |
| Adjusted EPS ($) | |||
| Penmark estimate | 0.58 | 1.20 | 2.63 |
| Consensus | 0.57 | 1.18 | 2.65 |
| Prior-year period | (0.05) | (0.42) | 1.20 |
| Diff. vs consensus | 1.8% | 1.7% | (0.9%) |
| Growth y/y (Penmark) | n/m | n/m | 118.9% |
Source: Penmark Research model; consensus LSEG via Yahoo Finance, Nov 2026; 4Q25 and FY25 from company filings. 4Q26E guided ~$90m revenue and $0.57–0.59 non-GAAP EPS.
Source: Company filings (A); Penmark Research model (E); consensus LSEG via Yahoo Finance, Nov 2026 (amber tick).
Next 12 months hinge on permits and capacity; AXTI is not cheap on history, and only fairly priced against peers
| 6 Nov 2026 | AXTI | Peers | Diff. |
|---|---|---|---|
| Valuation | |||
| Fwd P/E, at price | 14.8x | 16.5x | −10.3% |
| Fwd P/E, at target | 12.7x | 16.5x | −23.0% |
| Mid-cycle P/E (anchor) | 15.6x | n/a | n/a |
| Growth, returns and risk | |||
| Rev. growth, 3Y fcst | 101% | 43% | +58pp |
| Gross margin, TTM | 44.0% | 38.2% | +5.8pp |
| Beta, 5Y | 2.01 | 1.65 | +0.36 |
| Debt / equity | 0.08 | 0.31 | −0.23 |
Source: Penmark Research model; peers Coherent, Lumentum, Photronics median (S&P Global Market Intelligence data, 6 Nov 2026). AXTI FY27E non-GAAP P/E; gross margin is trailing twelve months. Cost-of-equity beta 1.41 is the peer median relevered (Exhibit 70). Difference in % for multiples, absolute or pp for ratios.
- 18 NovAPEC Shenzhen, 18–19 Nov; AI dialogue due by end-Nov. A first US-bound InP permit: +$0.07 a share (25%); we assume none.
- 27 NovChina’s US-specific gallium and germanium ban suspension expires. A squeeze on permits takes FY27 utilisation from 97% to 75% (−$0.52).
- Feb 20274Q26 results: we expect $90.0m revenue and $0.58 non-GAAP EPS; consensus $91.5m and $0.57.
- Mid-2027InP capacity update: base case already doubles capacity; 2.2x adds $2.27 a share (40%); a miss costs $0.12.
Source: Company filings and calls; Penmark Research model catalyst re-runs.
- Capacity beats the 2.0x plan. InP capacity of 2.2x instead of 2.0x in FY27 lifts value. Probability 40%; +$2.27/share.
- Tongmei listing completes. No-IPO odds fall from 40% to 0%, removing the 8% redemption floor. Probability 60%; +$1.12/share.
- Shortage runs to FY29. Utilisation 5 to 10 points higher in FY28 to FY29. Probability 35%; +$0.40/share.
- Supply wave cuts prices harder. A further 0.10 off the price index from FY29 supports the Sell. Probability 55%; −$2.60/share.
Source: Penmark Research model; value = base-case re-run less base value, probability as set in the model.
Source: Company filings FY09–FY25; Penmark Research model. Forward P/E history is not meaningful: AXT lost money in FY23–FY25.
Source: Yahoo Finance (peer data as of 10 Nov 2026); Penmark Research (AXTI at the 6 Nov price of $38.88 and FY27E EPS). Includes uncovered Sumitomo Electric.
The stock follows gross margin, and the price already pays for a 54% mid-cycle level
Source: Yahoo Finance quarter-end closes; company releases and calls (non-GAAP gross margin). 4Q24 to 3Q26; the 6 Nov price of $38.88 is outside the series.
Source: Company earnings calls, CFO prepared remarks; 4Q25 stated approximately.
Source: Penmark Research model, Revenue build, base case. Index is relative to the 4Q26 price and mix.
Source: Penmark Research model, Market Check, in % (price-implied solves at $38.88, base case).
Source: Penmark Research model (Valuation, Thesis, Market Check); bear to bull range of the DCF, thesis range of the P/E cross-check, and cost of equity of 10.9% to 12.9%.
Source: Penmark Research model, Market Check debate re-runs (each driver set to the market view, base case).
Three debates decide the stock, and the price sits on the market side of all three
The market capitalises shortage-era margins and prices as permanent and expects Tongmei to list. We model a supply response from FY28 and a 40% chance that the listing fails.
| Debate | Market view | Penmark view | Evidence | What would change our view | Resolves |
|---|---|---|---|---|---|
| 1. How high is mid-cycle gross margin? | Shortage pricing becomes the norm: 46% (bull case); the price implies 53.8%. | 40%: a supply wave from FY28 pulls margin back; value $31.05 against $34.46. | AXT ran 12.7% to 43.0% from FY09 to FY25; InP is now ~74% of sales; JX, Sumitomo and Yunnan Germanium add capacity. | Gross margin still above 50% in FY29 results. | FY29 to FY30 results |
| 2. How far does InP pricing fall? | Price index holds at 0.88 from FY31 (bull case). | 0.73 by FY31E; value $31.05 against $34.29. | JX plans ~7x to 10x capacity by ~FY30; Sumitomo 3.1x FY24 level by FY28; AXT and Sumitomo are ~80% of supply today. | Lead times stay above nine months and allocation continues through FY28. | JX and Sumitomo ramps, FY28 |
| 3. Does Tongmei list within three years? | Listing is certain: 0% chance of no IPO. | 40% chance of no qualified IPO; redemption at cost plus 8% a year costs $1.12 a share. | STAR listing pending 2021 to 2026, then withdrawn; investor backs overseas listing (8-K, 28 Oct 2026). | A filed Hong Kong prospectus before the late 2029 deadline. | Late 2029 |
Source: Company filings and calls; Penmark Research model (Market Check). Value per share: base-case DCF with one driver at the market view.
Source: Penmark Research model, Market Check: full re-runs with one driver at the market view.
How the debates map to the thesis
Each debate has one chapter. Debate 1 (margin, $3.41) is chapter 1; debate 2 (price, $3.25) is chapter 2; debate 3 (Tongmei, $1.12) is chapter 3. Section 9 links drivers to debates.
Why the market is wrong
The price capitalises shortage margins as permanent. At $38.88 the market needs a 53.8% mid-cycle gross margin; AXT has not earned more than 43.0% in any fiscal year FY09 to FY25.
Source: Company filings and calls; Penmark Research. February 2027 results date is an estimate, not company-confirmed.
Quarterly sales leapt from $28m to $73m in a year, and the price assumes the shortage never ends
AXT grows the crystals that sit under lasers and radio chips, all in China. Demand for one, indium phosphide, has outrun supply.
The scene
On 29 October 2026, AXT told investors that its order backlog had passed $500m, up from a little over $100m three months earlier. That is more than five years of FY25 sales ($88.3m). Every customer was on allocation, management said, with lead times past nine months. AXT is a Fremont, California company, listed since 1998.
What it actually does
Chips are built on a thin crystal disc called a wafer. AXT grows indium phosphide (InP) crystals, used in optical lasers, and gallium arsenide (GaAs), used in radio and sensor chips, then slices and polishes them. It does not design chips. About a third of sales are raw materials such as gallium.
How the money comes in
Customers pay per wafer, and the price rises when wafers are scarce. In 3Q26, InP brought in $54.3m of $73.4m revenue (74%), GaAs $9.1m (12%), raw materials $9.7m (13%) and germanium $0.3m. Of each dollar of FY27E revenue we model, 48 cents is operating profit and 17 cents is free cash flow after capex.
Where the story turns
In February 2025 China put InP substrates on its export-control list, so every InP wafer needs a Chinese permit. InP revenue then rose from $8.0m to $54.3m in three quarters and non-GAAP gross margin from 21.5% to 56.1%. On 28 October 2026 AXT agreed to invest $400m in its subsidiary Tongmei alongside up to $545m from outsiders, leaving AXT with 73.4%.
Why it works, and why it might not
Management says only one other supplier is at parity in 6-inch InP. The loop breaks if rivals add capacity: JX plans roughly 7x to 10x and Sumitomo 3.1x by FY28 to FY30 (debate 2, section 4).
Exhibit 26: One customer, start to finish. Source: Company call and release, 29 Oct 2026; Penmark Research model.
InP now brings in 74% of revenue and 56 cents of every dollar becomes gross profit
Source: Company 3Q26 results release (8-K, 29 Oct 2026), GAAP; product split from CFO remarks. 3Q26 (quarter to 30 Sep 2026).
Source: Company releases and calls, 29 Oct 2026 and 30 Jul 2026; Signals KPI series.
What it sells, and why customers buy
InP wafers are what the market is short of. AXT and Sumitomo supply about 80% of InP; all customers are on allocation.
How it is priced and paid for
Customers pay per wafer. Lumentum prepaid $43.5m in 3Q26; AXT is pulling back from long-term agreements to keep pricing flexible.
Cash versus profit
We model FY27E operating cash flow of $244.7m against $160.0m of capex: free cash flow is $84.7m.
Source: Company filings and releases (A), 4Q24 to 3Q26; quarter ends 31 December to 30 September.
| $m | 4Q25 | 1Q26 | 2Q26 | 3Q26 | 3Q26 share |
|---|---|---|---|---|---|
| InP wafers | 8.0 | 13.6 | 30.7 | 54.3 | 74% |
| GaAs wafers | 7.0 | 5.4 | 6.6 | 9.1 | 12% |
| Germanium, raw materials, other | 8.0 | 7.9 | 10.3 | 10.0 | 14% |
| Total revenue | 23.0 | 26.9 | 47.6 | 73.4 | 100% |
| Non-GAAP gross margin | 21.5% | 29.9% | 45.0% | 56.1% |
Source: Company calls, CFO prepared remarks, 4Q25 to 3Q26; Signals KPI series. 4Q25 InP and GaAs are stated approximately; other = total less InP and GaAs.
Revenue is capacity times utilisation times price, and price is the input we expect to fall
Source: Penmark Research model, Revenue build, base case. Capacity in $m of quarterly revenue at 4Q26 prices; index relative to 4Q26.
Source: Penmark Research model, Revenue build (E). Revenue by product, $m.
Source: Company 10-Q 2Q26 (revenue by destination, via Signals); top-five share from Signals KPI series. Other includes North America $0.2m.
Source: Penmark Research model, Assumptions (gross margin by scenario, %); FY31E bear = historical median 29.8%. FY09 to FY25 range 12.7% to 43.0%.
Fixed and variable cost to deliver
Operating costs are mostly fixed, so margin follows price. Operating expenses rise only from $47m in FY26E to $62m in FY27E while revenue grows 114%; gross margin does the work (60% FY27E, 40% from FY31E).
Concentration
Every wafer is made in China and sold by permit. InP, GaAs and germanium all need Chinese export permits; none for US-bound InP has been approved. The top five customers were 29% of 2025 revenue and 35% in 3Q26.
Two suppliers hold about 80% of InP today, and three rivals are adding 3x to 10x capacity
Source: Company 10-K FY25 and 3Q26 call; Penmark Research. Chain simplified to the steps AXT sells into.
Source: Company 10-K FY25; Coherent and Lumentum FY26 10-Ks; management 3Q26 call; Penmark Research model.
Source: Management 3Q26 call; JX release 16 Jun 2026; Nikkei; Reuters (as logged in the Penmark model). Bases differ by company; markers are mid-points.
| Advantage | Evidence | Lasting? |
|---|---|---|
| Low defect density | Barrier at 8-inch GaAs; premium for low-EPD InP (10-K FY25) | Medium |
| Own raw materials | Gallium and germanium made in-house; 13% of sales | Medium |
| Recent capacity | Only supplier to add significant facilities recently (10-K) | Fades by FY28 |
| 6-inch InP | One other supplier at parity (CEO, 3Q26 call) | Medium |
| China cost position | All output made in the PRC under export permits | Cuts both ways |
Source: Company 10-K FY25; 3Q26 call; Penmark Research assessment.
Where the margin sits and who threatens it
AXT and Sumitomo supply about 80% of InP today, so price is set by scarcity, not cost. New supply arrives from JX (7x to 10x over four years), Sumitomo (3.1x by FY28) and Yunnan Germanium, while AXT itself doubles capacity again in FY27. Demand is unusually concentrated: Coherent has $11.8bn of vendor purchase commitments and Lumentum $2.35bn of purchase obligations (FY26 10-Ks), so double ordering is a cycle risk as well as a demand signal.
Source: Company filings; Signals key metrics (Photronics FYE Oct, Entegris FYE Dec); Penmark Research model, Normalisation and Market Check. Gross margin in %.
Management has beaten its last two revenue guides by 11% or more, but insiders sold $29m in June to August
| Director / officer | Since | Role |
|---|---|---|
| Morris S. Young | 1989 | Chairman, CEO; not independent |
| Jesse Chen | 1998 | Lead independent director |
| David C. Chang | 2000 | Independent director |
| Leonard J. LeBlanc | 2025 | Independent director |
| Tracy Liu | Jun 2026 | Independent; ACM Research audit chair |
| Jia-Bin Duh | Jul 2026 | Independent; ex-Microsoft China |
| Gary L. Fischer | 2014 | CFO, Secretary (not a director) |
Source: 2026 proxy statement (DEF 14A); company 8-K filings of 22 Jun and 22 Jul 2026. Board expanded to five on 17 Jun and six on 16 Jul 2026.
Management record
Management beat its last two revenue guides by 11% and 40%. In January 2026 it cut 4Q25 guidance on fewer export permits and then landed inside the cut range.
Pay and independence
Half of the annual equity award is tied to a financial metric. Awards vest over four years; say-on-pay passed on 4 Jun 2026 with 97.4% of votes cast (18.26m for, 0.41m against, 0.08m abstaining). Five of six directors are independent; the four directors in the 2026 proxy are aged 68 to 85.
Source: SEC Form 4 filings via Signals, 1 Jun to 19 Aug 2026; proceeds = shares x price per filing. Sales only; CEO figure includes option exercises.
Source: Company filings (A); Penmark Research model (E). FY26E is the weighted average; 4Q26E run-rate 66.1m.
| Guided | Guide | Reported | Result |
|---|---|---|---|
| Revenue, 4Q25 | $22.5m to $23.5m | $23.0m | In range |
| Revenue, 2Q26 | ~$34m | $47.6m | +40% |
| Non-GAAP EPS, 2Q26 | $0.06 to $0.08 | $0.19 | Above |
| Revenue, 3Q26 | ~$66m | $73.4m | +11% |
| Non-GAAP EPS, 3Q26 | $0.30 to $0.32 | $0.41 | Above |
Source: Company releases and calls; Penmark Research model, Guidance. Guides given 8 Jan, 30 Apr and 30 Jul 2026; ranges or approximate figures.
Capital allocation
AXT spends on capacity and funds it with Tongmei equity, not buybacks. We model $160m of capex in FY27E and $150m in FY28E against $25m in 4Q26E; the Tongmei round adds up to $545m of outside capital and $400m from AXT. Governance does not hold back the rating; the concern is the $545m of outside capital with an 8% redemption floor (debate 3, $1.12 a share).
The stock trades on InP margins and optics sentiment, and the market has priced both above our value
Three drivers pass at least two of three tests: gross margin and InP pricing, capacity and backlog, and optics sentiment. The stock trades on the InP supply story and on its peers, not on the market.
| Driver | Tests | Market reaction | Model value | Debate | Next number |
|---|---|---|---|---|---|
| 1. Gross margin and InP pricing | Events, Q&A | +21% to +29% on three beats; −9.6% on the fourth | $3.41 a share | 1, 2 | 4Q26 GM: 64.5% (ours) |
| 2. InP capacity and backlog | Q&A, events | Backlog >$500m; 39 of 91 questions | $2.27 a share | 2 | Mid-2027: 2.0x capacity |
| 3. Optics and AI sentiment | Factors, events | R² 0.44; −54% in a month | $2.53 a share | 1 | Coherent and Lumentum prints |
Source: Penmark Research model (Market Check, Sensitivities), Yahoo Finance, Signals call transcripts. Model value = difference between value at the market view and at our view, base case; driver 3 = cost of equity 1pt lower.
Source: Yahoo Finance daily closes, next-session reaction in % to the last eight results (20 Feb 2025 to 29 Oct 2026); Penmark Research calculation.
Source: Yahoo Finance daily returns, 1 Oct 2025 to 6 Nov 2026; OLS R-squared of AXTI on each factor, Penmark Research calculation.
Source: Signals call transcripts, 4Q25 to 3Q26; 91 analyst-question paragraphs coded by topic, Penmark Research.
Dropped candidates
The market and Tongmei do not move the stock. The S&P 500 explains 17% of daily moves and the chip index 28%. Tongmei and capital raising drew 2 of 91 questions, though it is worth $1.12 a share to us (debate 3). Export permits drew 12 questions but have no measured event-day reaction, so they stay in the risk register.
Source: Yahoo Finance, Signals call transcripts and company releases; Penmark Research calculations.
Each driver is worth $2 to $4 a share to us; none supports the price on its own
What the market does. Non-GAAP gross margin rose from 21.5% (4Q25) to 56.1% (3Q26) while the price rose from $16.35 to $77.41; the correlation is 0.93 over eight quarters. At $38.88 the price implies a 53.8% mid-cycle margin.
What our model says. A 40.0% mid-cycle margin is worth $31.05 a share; the market view (46%) is worth $34.46. The gap is $3.41, or 8.8% of the price.
Debate and next number. Debates 1 and 2 (section 4). Next: 4Q26 results in February 2027 (date unconfirmed). We expect revenue of $90.0m and gross margin of 64.5%; consensus revenue is $91.5m. The last four results moved the stock −9.6% to +28.7%.
What the market does. Backlog above $500m and capacity at 2.75x to 3.0x FY25 exit. Capacity and demand were 39 of 91 analyst questions; management had no worry about overcapacity for about two years.
What our model says. Doubling capacity in FY27 is already in our base case; beating it (2.0x to 2.2x) adds $2.27 a share (40% probability), and a 5pt lower FY27 utilisation under the Coherent clause costs $0.12.
Debate and next number. Debate 2 (section 4). Next: mid-2027 capacity update; we model $141.8m of quarterly InP capacity exiting FY27 against $70.9m exiting FY26.
What the market does. Daily R-squared against Coherent and Lumentum is 0.44 (correlation 0.53 and 0.52). From the 6 Oct close of $84.06 to $38.88 on 6 Nov the stock fell 53.7%, a month that included sector weakness, the 28 Oct Tongmei round and the 29 Oct 3Q26 results (−9.6% next session); SOXX fell 30.5% in the same month.
What our model says. The price implies a 9.5% cost of equity against our 11.9%. A one-point lower cost of equity lifts the weighted value to $32.32 and the target to $35.84 (−7.8%, Hold), the only robustness test that changes the rating.
Debate and next number. Debate 1 (section 4). Next: Coherent and Lumentum quarterly results (dates unconfirmed); beta has no model estimate, so we do not forecast the reaction.
Source: Penmark Research model, Market Check and Sensitivities (base case; optics driver is the change in probability-weighted value, $29.79 to $32.32).
Chapter 1: the price needs a 54% mid-cycle gross margin; we model 40%
What the market believes
The price of $38.88 needs a 53.8% mid-cycle gross margin; our bull case uses 46%. The price therefore requires peak margins to persist through the cycle.
The evidence
- Reported: AXT earned 12.7% to 43.0% across FY09 to FY25, median 29.8%.
- Industry: Photronics and Entegris ran 20.3% to 46.4%, median 40.1%.
- Management: the CFO expects FY27 gross margin above 60%, and the CEO sees no overcapacity for about two years.
What it is worth
We take the shortage to FY28 (60%, 56%) and fade to 40% from FY31E. At the market view of 46% the base value is $34.46; the difference is $3.41 a share. Using the historical median in every case gives $23.91.
Signposts
- FY29 results: margin still above 50% would break our view.
- 3Q27 to 4Q27 prints: margin below 60% confirms fade; above 64.5% breaks our view.
Source: Penmark Research model, Sensitivities (full model re-runs at the 6 Nov 2026 price and inputs, base case).
Source: Company releases and calls (A), 4Q24 to 3Q26; Penmark Research model (reference lines).
Chapter 2: InP pricing falls from FY29, worth $3.25 a share against the market view
What the market believes
The bull case holds the price index at 0.88 from FY31. The price pays for a durable shortage.
The evidence
- Company release: JX plans roughly 7x to 10x InP capacity over four years (16 Jun 2026).
- Press: Sumitomo targets 3.1x its FY24 level by FY28 (Nikkei); Yunnan Germanium plans 450k wafers a year (Reuters).
- Filings: Coherent has $11.8bn of vendor purchase commitments and Lumentum $2.35bn of obligations: double ordering can unwind.
What it is worth
Our index falls to 0.73 by FY31E (market view 0.88): $3.25 a share. A further 0.10 off the index from FY29 costs $2.60 (55% probability); a double-ordering unwind costs $0.42.
Signposts
- FY28: JX and Sumitomo ramp dates and wafer lead times.
- Lead times shortening from 9+ months signals the turn.
Source: Penmark Research model, Revenue build, base case (E). Index relative to 4Q26 price and mix.
Source: Penmark Research model, Sensitivities (catalyst re-runs, base case, $ a share).
Chapter 3: the Tongmei round dilutes AXT to 73.4% and carries an 8% redemption floor
What the market believes
The round is a clean positive: $400m from AXT and up to $545m from outside investors build $947.6m of Tongmei net cash, and a listing follows.
The evidence
- 8-K, 28 Oct 2026: AXT subscribes $400m for 342,998,797 shares and ends with 73.4% of Tongmei.
- 10-K FY25: the STAR Market IPO process ran from 2021 and was then withdrawn.
- Terms: investors can redeem at cost plus 8% a year if no qualified IPO within three years.
What it is worth
We put 40% on no qualified IPO in three years (bear 60%, bull 20%). The redemption floor costs $1.12 a share against a certain listing, and new investors hold 16.2% of Tongmei equity; with existing minorities, non-controlling interests take 27% of FY27E net income.
Signposts
- Late 2029: IPO deadline; an earlier Hong Kong filing lifts value.
- 3Q26 10-Q: deposits and cash by entity.
Source: Penmark Research model, Valuation (equity bridge, DCF, base case, $m); 66.1m shares gives $31.05. Deposits $57.5m (30 Sep estimate); JV minorities $27.2m; "Other holders" = existing Tongmei minorities.
Source: Penmark Research model, Scenarios (value at price date, $ a share). Bear assumes 60% no-IPO probability, bull 20%.
AXT lost money in three of five years; GAAP and non-GAAP EPS differ by only $0.07
Source: Company 10-K filings FY21 to FY25 via Signals; gross margin in %.
Source: Company 10-K filings via Signals; operating cash flow less capital expenditure.
| $m unless stated | FY21A | FY22A | FY23A | FY24A | FY25A |
|---|---|---|---|---|---|
| Revenue | 137.4 | 141.1 | 75.8 | 99.4 | 88.3 |
| Revenue growth | n/a | 2.7% | −46.3% | 31.1% | −11.1% |
| Gross margin | 34.5% | 36.9% | 17.6% | 24.0% | 12.7% |
| Operating income (GAAP) | 12.9 | 12.6 | −21.6 | −14.8 | −22.0 |
| Diluted EPS, GAAP ($) | 0.34 | 0.37 | −0.42 | −0.27 | −0.49 |
| Stock-based compensation | 4.5 | 4.0 | 3.5 | 3.1 | 3.3 |
| Capital expenditure | 29.6 | 28.5 | 10.5 | 5.8 | 6.0 |
Source: Company 10-K filings FY21 to FY25 via Signals; Penmark Research. FY22 growth from FY21 revenue of $137.4m.
Adjusted versus reported
The gap between GAAP and non-GAAP EPS is small: $0.07 in FY25A and $0.07 in FY27E. Only stock-based compensation is excluded ($3.3m in FY25A, $4.5m in FY27E). The DCF treats it as a cash cost.
Profit to cash
Cash conversion has not been tested in an upcycle. FY27E operating cash flow is $244.7m against $231.5m consolidated net income, after $28.7m of working-capital build; capex takes $160.0m.
What flatters the numbers
Customer prepayments and minorities need watching. Customer deposits are estimated at $57.5m at 30 Sep 2026 (the 10-Q was not filed at the model date); we deduct them in the equity bridge. Minority interests take $62.4m of FY27E consolidated net income, so net income to AXT is $169.2m of $231.5m.
Inventory
Inventory was $103.7m at 30 Sep 2026, 117% of FY25 revenue; the FY27E build to $111.8m is modest.
Source: Company 10-K filings FY09 to FY25 via Signals; Penmark Research model, Normalisation (operating income divided by revenue).
Our forecasts match consensus on revenue and EPS and differ on how long the margin lasts
Source: Company filings (A); Penmark Research model (E), Financials. Revenue in $m; gross margin in %.
| $m unless stated | FY26E | FY27E | FY28E | FY29E | FY30E | FY31E |
|---|---|---|---|---|---|---|
| Revenue | 237.9 | 509.9 | 717.0 | 719.7 | 693.7 | 750.6 |
| Gross margin | 54.0% | 60.0% | 56.0% | 46.0% | 41.0% | 40.0% |
| EBITDA | 91.7 | 261.3 | 360.3 | 295.5 | 251.6 | 267.2 |
| Non-GAAP EPS ($) | 1.20 | 2.63 | 3.45 | 2.78 | 2.35 | 2.49 |
| Capex | n/a | 160.0 | 150.0 | 110.0 | 80.0 | 75.1 |
| FCF incl. prepayments | n/a | 84.7 | 114.3 | 141.8 | 151.0 | 149.8 |
Source: Penmark Research model, Financials (base case); FY26E free cash flow not modelled because 9M FY26 cash flow was not filed. Estimates shaded.
| 4Q26E | Guide | Penmark | Consensus |
|---|---|---|---|
| Revenue ($m) | 88.2 to 91.8 | 90.0 | 91.5 |
| Gross margin | 60% to 65% | 64.5% | n/a |
| Operating expenses, GAAP ($m) | 13.7 to 14.3 | 14.0 | n/a |
| EPS, GAAP ($) | 0.56 to 0.58 | 0.56 | n/a |
| EPS, non-GAAP ($) | 0.57 to 0.59 | 0.58 | 0.57 |
| Diluted shares (m) | 66.0 to 66.2 | 66.1 | n/a |
Source: Company 3Q26 call (29 Oct 2026); Penmark Research model; consensus LSEG via Yahoo Finance.
Volume, price and cost
Capacity doubles in FY27E; revenue growth then slows from 40.6% in FY28E to 0.4% in FY29E. Utilisation eases from 97% to 82% and the price index from 1.05x to 0.73x.
Fixed and variable cost
Operating costs rise from $62m in FY27E to $80m in FY31E; capex falls from $160m to $75m once capacity is built.
Against consensus
Revenue and EPS are within 2% of consensus. Our 60% FY27E gross margin matches the CFO’s “above 60%”; minorities absorb its effect on EPS.
Both valuation methods land at $28 to $31 a share, below the $38.88 price
| $m | 4Q26E | FY27E | FY28E | FY29E | FY30E | FY31E |
|---|---|---|---|---|---|---|
| EBIT | 44.1 | 243.9 | 331.5 | 257.1 | 207.4 | 220.1 |
| Less tax on EBIT | −5.9 | −36.5 | −49.7 | −38.6 | −31.1 | −33.0 |
| Plus D&A | 2.8 | 17.3 | 28.7 | 38.4 | 44.2 | 47.0 |
| Less capex | −25.0 | −160.0 | −150.0 | −110.0 | −80.0 | −75.1 |
| Less working capital | −9.5 | −28.7 | −61.3 | −16.8 | −3.2 | −22.6 |
| Unlevered FCF | 6.4 | 36.0 | 99.3 | 130.1 | 137.3 | 136.5 |
| Present value | 6.4 | 33.1 | 81.5 | 95.5 | 90.1 | 80.0 |
Source: Penmark Research model, Valuation. Unlevered FCF = EBIT less tax, plus D&A, less capex and working capital; excludes customer prepayments; stock compensation treated as cash; cost of equity 11.9%, terminal growth 3.0%.
Source: Penmark Research model, Scenarios (closed-form DCF grid, base case, $ a share).
| $m | DCF | P/E |
|---|---|---|
| EV, 30 Sep 2026 | 1,523.3 | 1,266.1 |
| Plus Tongmei net cash | 947.6 | 947.6 |
| Less deposits, JV minorities | −84.7 | −84.7 |
| Tongmei equity value | 2,386.2 | 2,129.0 |
| New investors’ claim | −469.6 | −444.7 |
| Existing Tongmei minorities | −238.6 | −209.6 |
| AXT share of Tongmei | 1,678.0 | 1,474.7 |
| Plus parent net cash | 350.8 | 350.8 |
| AXT equity, 30 Sep 2026 | 2,028.8 | 1,825.5 |
| Value per share ($) | 31.05 | 27.93 |
Source: Penmark Research model, Valuation (equity bridge: AXT holds 87.55% of Tongmei equity after the new investors’ claim; 66.1m shares; equity rolled to 6 Nov at 1.011x). DCF primary; mid-cycle P/E of 15.57x cross-check.
Source: Penmark Research model, Valuation (DCF at 30 Sep 2026); $m.
Method and why
A DCF fits because FY27E to FY28E earnings are far above mid-cycle. The cross-check applies a 15.57x through-cycle P/E (range 12.71x to 21.02x) to FY27E normalised NOPAT of $59.95m, plus $332.7m of above-mid-cycle profit. The 10.0% gap is under the 15% threshold.
The $33.33 target implies −14.3%, and only a lower cost of equity moves the rating
| 6 Nov 2026 | Price | Fwd P/E | Beta | Gross margin | Growth, 3Y | Through-cycle P/E |
|---|---|---|---|---|---|---|
| Coherent | $192.80 | 16.5x | 2.23 | 38.2% | 43.0% | n/a |
| Lumentum | $538.60 | 18.5x | 1.65 | 48.0% | 69.3% | n/a |
| Photronics | $27.36 | 13.3x | 1.32 | 33.7% | 3.7% | 12.7x |
| SUMCO* | ¥3,117 | 72.4x | 1.12 | 9.3% | 10.4% | 21.0x |
| AXT, at price | $38.88 | 14.8x | 2.01 | 44.0% | 101% | n/a |
Source: Penmark Research model, Peers (S&P Global Market Intelligence data, Signals, 6 Nov 2026; *SUMCO 25 Sep). Peer multiples at price; AXT non-GAAP FY27E EPS $2.63. AXT growth is the FY25A to FY28E revenue CAGR and gross margin is trailing twelve months (56.1% in 3Q26); Siltronic (15.6x through-cycle) sets the anchor only; Entegris and MACOM are excluded.
| Case | Probability | Value | Target | Return | Rating |
|---|---|---|---|---|---|
| Bear | 30% | $10.70 | $11.98 | −69.2% | Sell |
| Base | 50% | $31.05 | $34.74 | −10.6% | Sell |
| Bull | 20% | $55.27 | $61.85 | +59.1% | Buy |
| Weighted | 100% | $29.79 | $33.33 | −14.3% | Sell |
Source: Penmark Research model, Scenarios. Target = value x (1 + 11.9% cost of equity); no dividend. Probabilities are analyst judgements. Unrounded return −14.27%.
| One input solved at a time | Penmark | Price-implied |
|---|---|---|
| Mid-cycle gross margin | 40.0% | 53.8% |
| Cost of equity | 11.9% | 9.5% |
| Terminal growth | 3.0% | 9.6% |
Source: Penmark Research model, Market Check (each solve changes one input, base case, to reach $38.88).
| Robustness test | Weighted value | Target | Return | Rating |
|---|---|---|---|---|
| Cost of equity −1pt | $32.32 | $35.84 | −7.8% | Hold |
| Cost of equity +1pt | $27.79 | $31.37 | −19.3% | Sell |
| Beta 1.41 (peers) | $29.79 | $33.33 | −14.3% | Sell |
| Beta 2.01 (AXT own) | $24.81 | $28.51 | −26.7% | Sell |
| Hist. margin, all | $23.91 | $26.75 | −31.2% | Sell |
Source: Penmark Research model, Market Check (full model re-runs at the 6 Nov 2026 price and inputs).
Reconciliation to page 1
Target $33.33 = $29.79 x 1.119. Bear $10.70 (30%), base $31.05 (50%) and bull $55.27 (20%) tie to the values behind Exhibit 3 (targets $11.98, $34.74 and $61.85). At an unchanged $33.33 target, a Buy needs a price below $30.30 and the Sell is invalidated if the price falls below $37.04 (Hold between the two). At today’s price, a weighted value above $38.2 would be a Buy and one at or below $31.3 remains a Sell.
Limits
Probabilities rest on judgement; the bull case needs several things at once.
Source: Penmark Research model, Scenarios; values at 6 Nov 2026, target 12 months forward.
Index funds and trading firms hold the stock, and short interest is up 59% since April
Source: SEC 13F filings via Signals, period 30 Jun 2026; 13F filers reported 45.9m shares, 70.0% of 65.6m outstanding. Shares in millions.
Source: FINRA short interest via Signals, 30 Apr to 15 Oct 2026 (2026 settlement dates); percentage of 65.6m shares outstanding, Penmark Research calculation.
Source: Third-party Street target data (low $56, mean $94.46, high $125; Street ratings 3 Buy, 2 Hold, 1 Sell); Penmark Research model. Street targets were set before the October to November sector sell-off.
| Insider | Date | Transaction | Shares | Price |
|---|---|---|---|---|
| Morris S. Young, CEO | 1 to 2 Jun 2026 | Option exercises, sales | 197,498 sold | $112 to $113 |
| Jesse Chen, lead director | 3 to 15 Jun 2026 | Open-market sales | About 55,000 sold | $87 to $115 |
| David C. Chang, director | 15 Jun 2026 | Sale | 8,333 sold | $111.17 |
| Leonard J. LeBlanc, director | 7 to 17 Aug 2026 | Sales | 6,000 sold | $88 to $96 |
| Morris S. Young, CEO | 27 Oct 2026 | Equity grant | 23,350 granted | n/a |
| Gary L. Fischer, CFO | 27 Oct 2026 | Equity grant | 8,524 granted | n/a |
Source: SEC Form 4 filings via Signals, 1 Jun to 29 Oct 2026. Sales are selected filings; Form 4 data do not show why an insider sold.
Who owns it
Index funds and market makers are the largest filers. BlackRock (5.07m shares, $365.6m at 30 June) and Vanguard (5.00m, $360.3m) lead; Marex and Jane Street, both trading firms, follow. The 13F data cover only filers above the reporting threshold and say nothing about holders that are absent.
Trading and positioning
Short interest is up 59% since 30 April (7.55m to 11.98m shares) while the stock fell 54% in a month. Days to cover of 1.41 means shorts can exit within about a day and a half of average volume, so the position adds volatility rather than a squeeze risk.
The model takes revenue from $88.3m to $717.0m and EBITDA to $360.3m by FY28E
| $m unless stated | FY24A | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Revenue | 99.4 | 88.3 | 237.9 | 509.9 | 717.0 |
| Gross profit | 23.8 | 11.2 | 128.5 | 305.9 | 401.5 |
| EBITDA | −5.8 | −12.9 | 91.7 | 261.3 | 360.3 |
| EBIT | −14.8 | −22.0 | 81.3 | 243.9 | 331.5 |
| Net income to AXT | −11.6 | −21.3 | 72.5 | 169.2 | 223.2 |
| Diluted EPS, GAAP ($) | −0.27 | −0.49 | 1.13 | 2.56 | 3.38 |
| Operating cash flow | −12.1 | −12.8 | n/a | 244.7 | 264.3 |
| Capital expenditure | −5.8 | −6.0 | n/a | −160.0 | −150.0 |
| Free cash flow | −17.9 | −18.8 | n/a | 84.7 | 114.3 |
| Cash and investments | 22.8 | 128.4 | 1,383.8 | 1,468.5 | 1,582.8 |
| Total equity (incl. minorities) | 254.9 | 334.6 | 1,572.4 | 1,808.4 | 2,120.3 |
Source: Company 10-K filings via Signals (A); Penmark Research model, Financials (E). FY26E free cash flow not modelled because nine-month cash flow was not filed. Total assets at 30 Sep 2026 are model estimates; the 3Q26 10-Q was not filed at the model date.
| $m | 1Q25 | 2Q25 | 3Q25 | 4Q25 | 1Q26 | 2Q26 | 3Q26 | 4Q26E |
|---|---|---|---|---|---|---|---|---|
| Revenue | 19.4 | 18.0 | 28.0 | 23.0 | 26.9 | 47.6 | 73.4 | 90.0 |
| Gross margin, non-GAAP | 6.1% | 8.2% | 22.4% | 21.5% | 29.9% | 45.0% | 56.1% | 64.5% |
| Operating income, GAAP | −10.3 | −6.7 | −1.1 | −3.8 | −1.6 | 10.4 | 28.4 | 44.1 |
| Net income to AXT | −8.8 | −7.0 | −1.9 | −3.5 | −1.6 | 11.1 | 25.7 | 37.3 |
Source: Signals quarterly statements (10-Q and 10-K) and company releases (A); Penmark Research model (E). Gross margin non-GAAP for actual quarters; model gross margin for 4Q26E.
| Revenue by product, $m | FY25A | 3Q26A | FY26E | FY27E | FY28E |
|---|---|---|---|---|---|
| Indium phosphide | n/a | 54.3 | 169.5 | 433.2 | 633.6 |
| Gallium arsenide | n/a | 9.1 | 30.2 | 34.7 | 38.2 |
| Germanium | n/a | 0.3 | 1.2 | 1.2 | 1.2 |
| Substrates, total | 58.9 | 63.7 | 200.8 | 469.1 | 673.0 |
| Raw materials and other | 29.4 | 9.7 | 37.1 | 40.8 | 44.0 |
| Total revenue | 88.3 | 73.4 | 237.9 | 509.9 | 717.0 |
Source: Company 10-K FY25 (revenue by product type) and 3Q26 call, CFO remarks (A); Penmark Research model, Revenue build (E). Indium phosphide revenue is capacity x utilisation x price index; not split out in the FY25 10-K on this basis.
Source: Company 3Q26 call; Penmark Research model, Financials. Balance-sheet figure is a model estimate.
Reference: GAAP to non-GAAP, definitions, peer set and method
| $ a share | FY25A | FY26E | FY27E | FY28E |
|---|---|---|---|---|
| GAAP diluted EPS | −0.49 | 1.13 | 2.56 | 3.38 |
| Add back SBC | 0.07 | 0.07 | 0.07 | 0.07 |
| Non-GAAP diluted EPS | −0.42 | 1.20 | 2.63 | 3.45 |
| SBC, $m | 3.3 | 4.0 | 4.5 | 4.8 |
| Diluted shs, m | 43.9 | 62.8 | 66.1 | 66.1 |
Source: Company 10-K FY25 and 3Q26 release (A); Penmark Research model, Financials (E).
Source: Company 10-K FY25 (A); Penmark Research model, Financials (E).
Glossary
InP indium phosphide, the substrate for lasers and detectors in optical links. GaAs gallium arsenide, used in wireless, LEDs and sensors. Ge germanium, used in solar cells and infrared optics. EPD etch pit density, a measure of crystal defects. VGF vertical gradient freeze crystal growth. Tongmei Beijing Tongmei Xtal Technology, AXT’s main operating subsidiary. LTA long-term agreement. Qualified IPO a listing that releases the 8% redemption floor for new Tongmei investors.
Ratio definitions
Price return is the 12-month target divided by the price, minus one, excluding dividends. Target is probability-weighted value x (1 + cost of equity). Cost of equity is 4.9% + 1.41 x 5.0% = 11.9%. FCF is operating cash flow less capex; the DCF uses FCF excluding customer prepayments and treats stock compensation as a cash cost.
Peer set
Coherent, Lumentum, Photronics and SUMCO set beta and forward P/E; Photronics, SUMCO and Siltronic set the 15.57x through-cycle P/E anchor. Entegris (acquisition amortisation) and MACOM (chip designer) are excluded from the anchor.
Methodology
Primary method is a DCF; the cross-check is a mid-cycle P/E. The archetype is volume x price for a capacity-constrained materials maker, so we show gross margin, utilisation and backlog rather than a regression of P/E on growth. Forward P/E bands are not meaningful because AXT lost money in FY23 to FY25; Exhibit 15 shows a gross-margin band instead. Catalysts and risks are carried in Exhibits 13 and 14 on page 4. Price-date balance-sheet items at 30 Sep 2026 are model estimates, customer deposits are estimated at $57.5m, the cash split uses the 30 June 2026 disclosure, and the Tongmei investor price is assumed equal to AXT’s.
Data sources
Model date and price date 6 Nov 2026. Company 10-K FY25, 10-Q 2Q26, 8-K 29 Oct 2026, proxy statement 2026 and call transcripts via Signals; Yahoo Finance prices and consensus (LSEG); S&P Global Market Intelligence peer data.
Source: Penmark Research model, Methodology and Valuation.
Analyst certification and important disclosures
Ratings
Ratings are based on the upside from the share price to our 12-month price target, excluding dividends. The price target is weighted across bull, base and bear scenarios.
| Buy | Upside to price target above +10% |
| Hold | Upside to price target from −10% to +10%, inclusive |
| Sell | Upside to price target below −10% |
We review ratings at each set of results and whenever the facts change materially. A share price can move outside its band before the rating changes, and targets may not be reached.
Analyst certification
The Penmark Team certifies that (1) the views in this report accurately reflect its views about the subject company and its securities, and (2) no part of its compensation was, is or will be directly or indirectly related to the specific ratings, price targets or views in this report.
Contact
Penmark Research
Questions about our research:
research@penmarkresearch.com
All our research and coverage:
penmarkresearch.com
Basis of preparation
This report uses information publicly available up to its publication date. Share prices are closing prices on the trading day before the publication date, unless stated otherwise. Forecasts, valuations and price targets are Penmark Research’s own, and forecast periods are marked E. Sources are shown beneath each exhibit.
Valuation and price targets
Price targets are based on the valuation methods set out in each report and are weighted across bull, base and bear scenarios. The main risks to each rating and price target are described in the report. Penmark Research updates its coverage after each set of results but is not obliged to update a report at any other time.
Independence and conflicts of interest
Penmark Research is independent. It does not provide investment banking, brokerage, market-making or asset management services, does not trade securities, and does not accept payment from any company for research coverage. Companies do not see or review Penmark Research reports before publication.
No member of the Penmark Team holds a long or short position, directly or through derivatives, in the securities of any company Penmark Research covers.
Regulatory status
Penmark Research publishes general, impersonal investment research on a regular basis. It is not registered as an investment adviser with the US Securities and Exchange Commission or any state securities authority, is not a registered broker-dealer and is not a member of FINRA. Penmark Research does not give advice tailored to any person’s circumstances.
Not investment advice or an offer
This report is for information only. It is not personalised investment advice and is not an offer or solicitation to buy or sell any security or other financial instrument. It does not take into account any reader’s objectives, financial situation or needs. Readers should make their own investment decisions and seek independent professional advice where appropriate.
Forward-looking statements and risk
Forecasts, valuations and price targets are estimates based on assumptions that may prove wrong, and they may not be achieved. Opinions are current as of the publication date and may change without notice. Investing in equities involves risk, including the loss of capital. Share prices can be volatile, and past performance is not a reliable indicator of future results.
Data and sources
Company financials, operating data and filings come from company filings, earnings releases and earnings calls. Market data and any consensus estimates come from third-party sources named beneath the relevant exhibit; consensus figures from aggregators are unofficial. Penmark Research does not use other brokers’ research. Third-party data belong to their providers and may not be extracted from this report.
Accuracy and liability
Information is taken from sources believed to be reliable, but Penmark Research makes no representation or warranty as to its accuracy, completeness or timeliness. To the fullest extent permitted by law, Penmark Research accepts no liability for any loss arising from use of this report. Descriptions of litigation and regulatory matters reflect allegations and public records, not findings, unless stated. If a published report is corrected, the corrected version carries an “Updated” date at penmarkresearch.com.
Distribution and access
Penmark Research reports are available through penmarkresearch.com. Downloads and programmatic access are governed by the Penmark Research terms of use. This report is not directed at, or intended for, any person in Singapore or India, and access from those countries is blocked. It is not intended for distribution to, or use by, any person in any jurisdiction where that would be contrary to law or regulation or would subject Penmark Research to registration or licensing requirements there.
Copyright and permitted use
© Penmark Research. All rights reserved. Penmark Research is a trade name of Brightriver Labs. Readers may share a link to this report but may not copy, reproduce, redistribute, sell or use its content, including for text and data mining, without written consent. Contact: research@penmarkresearch.com.