Third quarter 2026 earnings call, October 29, 2026

AXT reported third quarter revenue of $73.4 million, up 54% sequentially and 162% from $28.0 million a year earlier, and guided the fourth quarter to roughly $90 million of revenue it has either already permitted or does not need permits to ship. The more important disclosure was that backlog has moved above $500 million, from a figure management had only described as "more than $100 million" a quarter ago. Non-GAAP EPS of $0.41 compares with $0.19 in the second quarter and a $0.03 loss a year ago, and fourth quarter non-GAAP EPS guidance of $0.57 to $0.59 implies another step up of roughly 40%.

Backlog Quintuples, but Not Because of Long-Term Agreements

The backlog figure includes long-term multiyear agreements, contracts of 12 months or less and spot purchase orders running through 2027. Northland's Tim Savageaux pressed on how much of the $300 million to $400 million sequential increase came from long-term supply agreements. VP of Business Development Tim Bettles was direct: it is "run rate business," and long-term agreements are "a relatively small part" of the total. The Lumentum agreement, which carried a $43.5 million wafer prepayment recorded as a liability, is the one signed in the third quarter and not a follow-on.

CFO Gary Fischer said this is deliberate. Management has pulled back on formal long-term agreements because it wants to protect price and avoid being locked into volume commitments. "We're not going to be announcing multiple long-term agreements," he said, even though "plenty of customers want us to sign up for them." For investors, this cuts both ways. The backlog is shorter-dated and more reliant on spot and annual pricing than a contract-heavy book would be, which leaves AXT exposed if the market turns, but it also lets the company capture further price increases in a supply-constrained market. Bettles said lead times have stretched beyond nine months, all customers are on allocation, and customers with existing long-term agreements are asking to raise volumes two to three times.

Gross Margin Resets Above 56%, With More Guided

Non-GAAP gross margin reached 56.1%, from 45.0% in the second quarter and 22.4% a year ago. Savageaux calculated incremental gross margin of 77% in the quarter. Management attributed the gain to volume, a richer indium phosphide mix, higher market prices, the shift from 3-inch to 4-inch wafers and manufacturing efficiency gains from higher volume. Fischer said he did not think margins had peaked and that the fourth quarter could start with a 6, "not a high 6." He added that "it's pretty obvious that we're going to cross over 60%" but declined to give a longer-term target given how fast conditions are changing.

The drivers beyond near-term mix are worth understanding. A new generation of proprietary crystal growth furnaces, originally purchased for 6-inch development, is also delivering large productivity gains on 4-inch laser material. CEO Morris Young pointed to demand for tighter etch pit density, or EPD, specifications on high-power lasers as another pricing lever, noting that "low EPD is our forte" and that tighter requirements mean "less competition and that's a premium to the price." Young said that when AXT first qualified with a major U.S. laser maker the requirement was below 500, and the company now delivers less than one-tenth of that. Raw material costs are rising significantly, but Young said efficiency, productivity and low-EPD gains "overwhelm" the cost increase. Early 6-inch yields are expected to be low, which is a modest headwind until more furnaces come online.

Capacity: Tripling in 2026, Doubling Again in 2027, and Still Short

AXT said it surpassed its goal of doubling indium phosphide capacity in August. Fischer indicated that, indexed to end-2025 at 1.0, capacity should reach about 2.75 to 3.0 by the end of 2026, with a goal of doubling again in 2027. A newly acquired adjacent building is in the design phase for conversion to indium phosphide production, with work starting within months. Young said the company's indium phosphide revenue opportunity has "more than quadrupled" in 2026 relative to where it began the year. Bettles said potential revenue capacity is "a lot higher" than the figures management offered last quarter, which an analyst recalled as roughly $60 million and $130 million for successive periods, helped by larger wafers and higher average selling prices. He said essentially all of the incremental fourth quarter revenue will be indium phosphide, implying quarterly indium phosphide revenue of well over $54.3 million.

Asked by Needham's Charles Shi about investor concern that AXT is building toward overcapacity in 2027, management rejected the premise. Bettles said that in 2027 "I'm not sure we'll even get our head above water," and Fischer said the CFO's view is that 2027 and 2028 capital spending is "less than is going to be needed." Young said a major customer told him the addressable market could expand "1,000x" and not 20x across scale-up, scale-out and scale-across architectures, though he admitted he did not know how that figure was derived. He was more cautious beyond the near term, saying he is confident for the next two years and that "three years out, it's hard to say." Fischer said capex would be reassessed as 2029 approaches. Investors should note that the argument rests on management's read of customer forecasts and allocation, and that new Chinese entrants are pursuing indium phosphide substrates, though Young argued 40 years of crystal growth experience cannot be quickly replicated....