To The Moon

Six weeks ago, $AXT Inc(AXTI)$ was a quiet, niche materials supplier trading in the high-$30s. Then a blowout earnings print sent it as high as $96, before the stock gave back a meaningful chunk of that move and now trades in the high-$60s to low-$70s.

Before SPR's technical coverage on AXTI updates next week, here's the fundamental case underneath the volatility: what actually drove the surge, what Wall Street thinks it's worth now, and the export-permit risk that's been the company's recurring headache for three years running.

The News Behind the Move

Picture a small semiconductor-materials company that most investors had never heard of, quietly supplying the wafers that go into fiber-optic lasers — and then AI data center demand shows up at its door all at once. That's the AXT story in one image. The company's second-quarter 2026 results, reported in late July, showed revenue of $47.6 million against a Wall Street estimate of $34.1 million, more than doubling from $18.0 million a year earlier. Adjusted earnings per share came in at $0.19 versus a $0.07 consensus, flipping the company from a prior-year loss into solid profitability. The driver was record indium phosphide (InP) revenue of $30.7 million — InP being the material used in the high-speed optical transceivers that move data between AI servers.

Management didn't treat it as a one-off. Guidance for the current quarter calls for EPS of $0.30–$0.32 on roughly $66 million in revenue, and the company has set a target of reaching $140 million in annualized InP production capacity by the end of 2026. Backing that expansion is a long-term supply agreement with Lumentum running through 2031, secured with $87 million in customer deposits — the kind of commitment that suggests this isn't just a spot-market spike. The expansion itself was funded in April, when AXT raised roughly $550 million in a stock offering priced at $64.25 per share specifically earmarked for its Chinese subsidiary Tongmei's InP capacity buildout.

What Wall Street Analysts Are Saying

The Street moved quickly once the numbers came in, and it moved almost entirely in one direction.

• Wedbush — reiterated Outperform, $93 target

• Needham — upgraded to Buy, $90 target

• B. Riley — raised its target to $55 from $52, the most conservative of the group

• Street average (per S&P Global, 5 analysts) — consensus rating of "Buy," average target near $91, with estimates ranging from a low of $55 to a high of $125

That's still meaningfully above where the stock trades today even after its pullback from the mid-$90s — but it's worth noting the range itself: a $70 gap between the most cautious and most optimistic targets on a stock that's already moved this fast is a signal that analysts are still calibrating how durable the AI-driven demand really is.

Is AXTI Overvalued or Undervalued?

This is genuinely a harder call than it looks. On one hand, the company just posted a quarter with 32.2% gross margin and positive free cash flow of about $3.6 million — real, cash-generating profitability, not just a revenue headline. The balance sheet is unusually strong for a company this size: roughly $412 million in cash, a current ratio near 4.8, and minimal long-term debt, largely thanks to the spring capital raise.

On the other hand, a stock that ran from the high-$30s to the mid-$90s in a matter of weeks has, by definition, priced in a great deal of future growth already. Even management's own bullish framing — reaching $140 million in InP capacity by year-end — implies the market is already looking well past this year's results toward 2027 and beyond. Whether that's justified depends heavily on two things: how consistently China's export-permit process behaves, and whether gross margins can climb from the low-30s toward the "40s" management has been targeting as capacity utilization improves.

The Risk Variables Investors Should Watch

Three threads worth following past next week:

Export permits remain the company's oldest and most persistent risk. Every shipment of gallium arsenide, germanium, and indium phosphide substrates out of China requires individual approval from China's Ministry of Commerce — a process AXT has been navigating since 2023. As recently as the fourth quarter of 2025, revenue came in below guidance specifically because fewer permits were issued than expected. The current growth story assumes that process stays smooth; it hasn't always.

The Tongmei IPO overhang is still unresolved. AXT's Chinese subsidiary Tongmei first applied to list on Shanghai's STAR Market back in January 2022, and as of the company's most recent disclosures, that listing remains under regulatory review with no assurance of approval. A resolution either way — approval or a formal withdrawal — would be a structural event for how investors value AXT's stake in its own manufacturing base.

Dilution has already happened once this year and could happen again. The April equity raise added over 8.5 million new shares to fund expansion. A company scaling capacity this fast, even while generating positive free cash flow, tends to keep capital markets as an option on the table rather than off it.

If the Story Shifts From Here

Suppose Tongmei's export permits keep flowing smoothly and the Lumentum ramp accelerates faster than guided — that's the scenario where Wedbush's and Needham's targets start to look conservative rather than aggressive, and gross margins climb toward the "40s" management has flagged. Now suppose instead permit delays resurface the way they did in late 2025, or the Tongmei IPO stalls in a way that unsettles the ownership picture — that's the scenario where B. Riley's more cautious $55 target, and the stock's sharp pullback from its August high, look like the more accurate read on how much of this growth is durable. Where the stock actually settles between those two outcomes is exactly what next week's technical coverage is built to capture.

Conclusion

One-line takeaway: AXTI's earnings turnaround is real and cash-generative, but the stock's round-trip from the high-$30s to $96 and back into the $70s shows the market is still working out how much of this AI-driven demand is durable versus how much was a re-rating that ran ahead of itself.