After a 100x surge in just one year, $AXT Inc(AXTI)$ reached its all-time high of $143.
Barely a few months later, it had fallen back to around $60 — a maximum drawdown of over 56%.
A lot of people started saying “AI is done” and “the bubble is bursting,” so of course optical names had to crash with it. But has the AI bubble actually burst? I think it’s way too early to call that. Most of the current judgment is pure emotion, completely detached from fundamentals. Everyone’s just arguing their own narrative.The real questions we should be asking are:
Why did money rush for the exits right now?
And more importantly for AXTI — is this just collateral damage, or is it actually a chance to reassess real value?Why did AXTI drop more than 70% while the company keeps landing orders? Why isn’t the stock getting sustained positive feedback? And if this is a potential opportunity, what signals should we actually wait for?
Latest earnings (Q2 2026 reported July 30):
EPS: $0.19 vs $0.07 expected (massive beat)
Revenue: record $47.6 million (up 164.8% YoY, up 77% sequentially)
Non-GAAP gross margin: 45.0% (up from 29.9% in Q1 and just 8.2% a year ago)
Indium Phosphide (InP) sales: record $30.7 million, driven by AI data-center demand for 800G and 1.6T optical transceivers
Guidance & outlook: Q3 2026 non-GAAP EPS expected $0.30–$0.32
Management targeting $60 million quarterly InP revenue run-rate capacity by end of 2026
This is a clear step-change. Revenue more than doubled sequentially in the key growth area, margins expanded sharply, and the company is guiding for another strong sequential EPS jump. So why the earlier collapse and why the ongoing volatility? A lot of people will blame the pullback on “AI cooling off” or broad market weakness.
That logic doesn’t fully hold up. If the entire AI infrastructure investment cycle was turning, you’d see the whole chain get hit — optical modules, servers, switches, data centers, power equipment, the works. Instead, AI infrastructure build-out is still moving forward and many segments remain strong.
A more accurate read is this:
the market is re-evaluating whether AXTI’s future growth speed actually matches its valuation.During the big run-up the market treated AXTI as more than just another semiconductor materials company. It became viewed as a critical link in the AI infrastructure supply chain.
The logic was simple — the real value in AI isn’t only the chip makers. Companies that control key materials, processes, and bottlenecks can benefit too.AXTI drew attention because of its position in indium phosphide (InP), which has real scarcity in high-speed optical communications. Money wasn’t buying current revenue or profits.
It was betting that when high-speed optical demand explodes, AXTI would become an important supplier.That’s also where the problem started. Once the stock ran hard and valuation climbed, the market’s focus shifted. At low prices people cared about growth potential. At high prices they started asking whether the growth speed could actually justify the valuation.The volatility wasn’t because the business deteriorated.
It was because the market adjusted its expectations and pricing.The real debate has already shiftedIt’s no longer “is there demand?”
It’s “can they deliver on the demand?”
By earlier 2026 AXTI’s InP backlog already exceeded $100 million. For a company whose annual revenue was still relatively modest, that number is impressive — it proves real customer demand. But orders ≠ revenue.
Revenue ≠ profit.
The missing piece is delivery and conversion. AXTI’s delivery capability is tightly tied to its supply-chain structure.
It’s a U.S.-listed company, but core production capacity is concentrated in China.
That brings cost and efficiency advantages, yet it also creates external approval and supply-chain uncertainty — something the market clearly dislikes.
Of the earlier $100M+ backlog, a portion was already permitted/unrestricted while a larger portion still needed further approvals. Demand is confirmed; timing of conversion is the key variable. Over the next few quarters, what actually matters isn’t how good the AI story sounds. It’s whether orders convert into revenue, revenue drives better profits, and profits start generating consistent free cash flow.
Q2 showed meaningful progress on all three fronts: revenue conversion accelerated, margins jumped, and guidance points higher. If that continues, AXTI can shift from concept-driven to performance-driven. If conversion slows again or supply-chain issues linger, pressure can return.What is AXTI actually worth?
I won’t throw out a single target price.
Method 1 – Normalized future earnings
Assume by 2027 capacity expands reasonably well, revenue reaches ~$220 million, and net margin recovers to around 15% → roughly $0.50 EPS.
Give a growth-oriented semiconductor materials company a 40–70x multiple + net cash → reasonable value roughly in the high-$20s to mid-$40s range in a more conservative framework.
Stronger execution and higher margins (like the 45% just reported) push the upper end higher.
Method 2 – Reverse-engineering current price
At elevated prices the implied long-term growth expectations are demanding — well above the broader indium phosphide industry’s long-term growth rate of ~13%. The market is pricing in continued outperformance and successful capacity ramp.
Method 3 – Probability-weighted scenarios
• Optimistic: Capacity expands smoothly, AI optical demand stays strong, pricing holds, restrictions ease → significantly higher
• Base case: Capacity improves, AI demand grows solidly, but competitors expand and pricing normalizes → mid-range
• Pessimistic: Further export tightening, faster silicon photonics progress, or industry overbuild → much lower The cluster of reasonable outcomes still sits in a zone that requires a margin of safety relative to where the stock has traded at peaks.A good industry does not mean every price is a good buy. We invest in price relative to value, not stories. The direction is solid, but entry price still matters a lot.
Bottom line after the latest report
The direction is right.
The value of indium phosphide in AI high-speed optical is real. Q2 delivered clear evidence that demand is converting, margins are expanding, and the company is executing better.
Guidance for Q3 is strong.But the most important question remains: can AXTI turn the opportunity into sustained revenue growth, rising profits, and free cash flow over multiple quarters?
Key things to keep watching:
Continued revenue growth and backlog conversion
Gross margin sustainability near or above the new elevated levels
Management commentary on export licenses, capacity ramp, and competitive landscape
Progress toward the $60M quarterly InP capacity target
AXTI is worth watching closely. The latest earnings give it a stronger fundamental footing. Now it needs to keep proving it quarter after quarter.
Comments