Milk Road AI

Milk Road AI

@MilkRoadAI

Apr 16, 2026, 14:58 UTC

If you’re hearing about it now, you’re late. Milk Road PRO spotted $CRDO, $NBIS, $AAOI, and $BE months ago, now they’re up big. See the next setups in real time, join us before the next run starts. t.co/IkfaZC5eOt

Milk Road AI

@MilkRoadAI

May 11, 2026, 16:32 UTC

AAOI reported Q1 revenue of $151.1 million, up 51% year over year and right in line with management's own guidance and the stock sold off 7% the next day on a $6 million miss versus Street estimates (Save this). That reaction was a gift and the earnings call itself told a completely different story than the stock price. CEO Thompson Lin raised full year 2026 revenue guidance to over $1.1 billion, while simultaneously disclosing that actual customer demand is running at $1.4 to $1.5 billion meaning demand already exceeds current capacity by $300–400 million. Management guided for 60–80% sequential revenue growth in Q3 alone, raised non-GAAP operating income guidance from $120 million to $140 million, and put a 40%+ gross margin target on the board for Q4 2026, up from 29% today. Lin also revised the mid-2027 monthly transceiver run rate upward from $378 million to $471 million adding over $1.1 billion in annualized revenue to the model in a single quarter. The product execution behind that guidance is already in motion. AAOI completed its first volume shipment of 800G transceivers to a large hyperscale customer in Q1, landed $124 million in new 800G orders from a single customer across March and April. Also, they shipped the first 10,000 units of a separate 800G order to a second hyperscale customer all before the major volume ramp even begins in Q3. Manufacturing capacity is ramping from 100,000 units per month today toward 650,000 by year end. While retail was still selling, the Milk Road Pro analysts added shares at $167.05, right into the fear because the fundamental setup had not changed at all. The thesis was simple, a company with $1.4–1.5 billion in real demand, a 60–80% Q3 revenue ramp guided by management, and a path to a $5 billion-plus annual run rate by mid-2027 had just been handed to the market at a discount. As of this morning, AAOI is trading at $184.82, a 24% gain in under three trading days from that entry. Buying into a selloff requires conviction that most people don't have in the moment. Knowing what to look past and what actually matters is exactly what Milk Road Pro is built for. Come join us before the Q3 ramp begins, link below!

Milk Road AI

@MilkRoadAI

Jun 26, 2026, 21:30 UTC

Market was bleeding, but Milk Road PRO was buying. While everyone else was trying to figure out if the AI trade was over, we were adding new positions. $MU, $AMD, $AAOI, and $NBIS were only the beginning. The next runners are already forming. Sign up for Milk Road PRO using the link below for just $1 and get positioned before the market wakes up.

Milk Road AI

@MilkRoadAI

Aug 14, 2026, 20:15 UTC

The GPU is not the only bottleneck inside an AI data center. (Save this if you want to understand how to play the AI supercycle.) The real constraint is optical and photonic interconnect, the lasers, transceivers, DSPs, and test systems that move data between chips, between racks, and between buildings at the speed of light. Without high speed optical connectivity, every NVIDIA GPU cluster in the world becomes a traffic jam, compute waiting on data that cannot move fast enough. As AI clusters scale from 800G to 1.6T to 3.2T and beyond each generation requiring exponentially more bandwidth, lower latency and higher power efficiency the optical photonics supply chain is now the limiting factor in the entire AI buildout. NVIDIA recognized this before almost anyone else, and has been systematically locking in strategic partnerships, direct investments, and supply agreements across every layer of the optical stack from lasers to test equipment. The seven companies above represent the full photonic food chain that makes AI data centers possible and here is a breakdown of each. AAOI, Applied Optoelectronics is the highest leverage pure play on the optical photonics bottleneck, a vertically integrated US manufacturer of InP lasers, EML chips, and finished transceivers that sits inside the supply chain constraint rather than being subject to it. In Q1 2026, data center revenue surged from $32 million to $81.4 million year over year, and the company just received its first volume order for 1.6T transceivers, a $200 million+ contract from a long-term hyperscale customer with shipments beginning Q3 2026. AAOI is targeting 500,000+ combined 800G and 1.6T units per month by year end, full year 2026 revenue tracking toward $1 billion, a 107.6% year over year increase with 2027 consensus projecting another 90.1% on top of that. CRDO, Credo Technology is the connectivity intelligence layer of the optical stack providing Active Electrical Cables (AECs), ZeroFlap optical transceivers, optical DSPs, retimers, and SerDes chiplets that make high speed AI rack-to-rack connections reliable and power-efficient. Credo's ZeroFlap technology solves one of the most persistent and underappreciated problems in AI data centers, link flaps, repeated connect disconnect cycling that causes silent performance degradation and cluster instability at scale. The company recently acquired DustPhotonics, expanding its portfolio from 800G to 1.6T and 3.2T, and partnered with Rebellions to integrate ZeroFlap AECs into RebelPOD AI inference clusters placing Credo at the heart of the inference infrastructure buildout. MRVL, Marvell Technology is the platform semiconductor giant of the optical interconnect world providing 800G/1.6T scale out optics, optical DSPs, custom XPU silicon, and through its acquisition of Celestial AI, next generation optical I/O for scale-up connectivity. Q1 FY2027 revenue hit a record $2.418 billion, up 28% year over year with Q2 guided to $2.7 billion. The Celestial AI acquisition is expected to contribute $500 million in annualized revenue by Q4 FY2028, doubling to $1 billion by Q4 FY2029, as optical scale up interconnects become the dominant architecture for next generation AI clusters. COHR, Coherent Corp is the most vertically integrated optical components company in the world spanning materials science, InP laser fabrication, silicon photonics, EML devices, transceivers, and full co-packaged optics in a single platform. NVIDIA made its position clear with a multi billion-dollar long term strategic supply agreement and a $2 billion direct equity investment confirming Jensen Huang views Coherent's optical engines as critical national AI infrastructure, not a commodity vendor. LITE, Lumentum is the laser chip backbone of the entire AI optical ecosystem manufacturing the EML laser chips and photonic components that go inside transceivers from almost every major supplier in the industry. Q3 FY2026 revenue of $808.4 million up 90% year over year crushed consensus, and Q4 guidance of $960 million to $1.01 billion represents yet another record, with management projecting a trajectory toward $2 billion in quarterly revenue. The Rosenblatt InP supply model projects Lumentum's InP Datacom laser revenues growing from $600 million in 2025 to $9 billion by 2030, the largest absolute revenue expansion of any single company in the photonics supply chain. VIAV, VIAVI Solutions is the quality assurance and test infrastructure layer of the optical buildout, the company that validates every transceiver, every silicon photonic device, and every 1.6T Ethernet link before it goes into production inside an AI data center. VIAVI unveiled the industry's first high density OSFP test platform for 1.6T Ethernet interoperability at OFC 2026, and is already at the forefront of 3.2T testing positioning it to benefit from each successive speed generation transition. Because every optical component must be tested before deployment, VIAVI's revenue is directly correlated to total optical hardware production volume making it a leveraged proxy on the entire photonics buildout. AEHR, Aehr Test Systems is the wafer level burn in and test specialist for silicon photonics and optical I/O devices, the last quality gate before photonic chips ship at scale into AI data centers and optical interconnect systems. Aehr's FOX-XP and FOX-NP systems perform full-wafer contact test and burn-in for the exact photonic and integrated optical devices used in CPO and NPO architectures ensuring silicon photonics chips perform reliably under sustained load before any hyperscaler touches them. As silicon photonics production volumes scale by orders of magnitude to meet the 12x InP capacity ramp identified by Rosenblatt, Aehr becomes the mandatory infrastructure partner for every foundry trying to yield chips at commercial scale. Of all seven companies across the photonic stack, AAOI is Milk Road's highest conviction position and the reason is vertical integration. Every other company in this stack depends on someone else for a critical component. AAOI designs, grows, fabricates, and ships its own InP lasers, EML chips, and finished transceivers from a single vertically integrated manufacturing platform capturing margin at every layer and avoiding the supply dependencies that constrain its peers. A $200 million+ first 1.6T volume order already booked, $1 billion in revenue targeted for 2026, 107% year over year growth consensus, a new Texas facility coming online, and 500,000 units per month capacity target by year end,. At Milk Road, we’re tracking the photonics bottleneck layer by layer, from lasers and transceivers to DSPs, silicon photonics, and the test equipment behind it all. AAOI is our highest conviction name in this stack, and if you want to see the trades we’re making around photonics and the rest of the AI infrastructure buildout, come join us for just $1 using the link. t.co/thIhK9ZH4E