Poet Stops the Bleeding, Gets Back To Shipping

If you owned POET Technologies (NASDAQ: POET) over the last twelve months and you still have hair, congratulations — you outperformed the chart.

Two registered direct offerings. A short report. A confrontational CFO rebuttal. A Q4 miss. A geopolitical shock that put oil on a tear and dragged every growth name in the semiconductor universe through a wood chipper. The kind of year that turns IR teams into therapists and traders into day drinkers. And here is the punchline: POET is back to business, the optical trade is the trade on Wall Street, and the noise has finally burned off enough to see what the company is actually doing.

The Wild Ride, Abridged

Two $150 million raises in three months took the share count past 152 million — a roughly 53% YoY expansion. Long-time holders did not throw a parade. Management's argument was the unsexy correct one: raise from strength, not from a margin call.

Then Wolfpack Research published the obligatory short hit — PFIC tax claims, undisclosed promotion history, the usual greatest-hits album. CFO Thomas Mika's response was not a polished IR statement. It read like a man who had been waiting for the call. Retail piled in, short interest stayed elevated, and the chart did its best impression of an EKG attached to an espresso machine.

On March 31, Q4 2025 missed both lines: EPS of -$0.32 against -$0.09, revenue of $341K against $400K. The bear case wrote itself in crayon — sub-billion market cap, hundreds-multiple price-to-sales, no earnings, balance sheet still tied to the equity tap.

Then Iran happened.

And Then a War Broke Out

Just as photonics was finally getting its long-awaited promotion to a real Wall Street narrative, the Strait of Hormuz lit up. JPMorgan went on record warning that the market was complacent on the conflict, that energy-driven inflation could hand cyclical and growth-sensitive names a beating, and that the rate-cut script everyone was reading from was about to need a rewrite.

Markets agreed, briefly and violently. Oil ripped, multiples got compressed, and small-cap semiconductors — POET very much included — got dragged into the alley with everything else that trades on duration. There is nothing quite as helpful for a story stock fighting a short report and an earnings miss as a war risk premium getting bolted onto its discount rate.

Except, of course, that semiconductors are also where you go to buy the dip. NVIDIA, Micron, AMD, TSMC — every name in the AI infrastructure stack got marked down on the Iran shock and then promptly bid back up by anyone who had been waiting for an excuse. POET was not exempt from the selling. It was also not exempt from the bounce. By the time the smoke cleared, the chart had absorbed a war and was still pricing the same underlying question: does this company ship product or not?

What the Headlines Missed

Three things changed in POET's favor while the tape was busy panicking, and none of them showed up in a quarterly EPS line.

The product is real. At OFC 2026 in Los Angeles, POET demonstrated Starlight Gen 2 — eight high-power lasers in a compact external light source with wavelength-division multiplexing built in. WDM is the only way the math works at 800G and above. This was not a roadmap slide. It was working silicon in front of the people who decide what gets specified into next-generation transceivers.

The supply-chain access is real. Foxconn and LITEON are not press-release wallpaper. They are the routes through which integrated photonic engines actually reach hyperscale racks. You do not get qualified into those programs by sending a cold email.

The unit guidance is real. Management is targeting more than 30,000 unit shipments in 2026, with high-volume production beginning in Q2. A photonics company crossing from samples into volume crosses a different river than one that misses on $400K of revenue. The first is execution. The second is a rounding error inside a $550 million market cap.

The Sector Caught Up

Here is what matters more than any single quarter: optical interconnects are no longer a niche thesis defended on Twitter by a handful of photonics nerds. They are the trade.

Jim Cramer went on Mad Money in early April and called it directly — hardware up, software down, and the war inside tech, in his words, more captivating than the actual war in Iran. Salesforce and Adobe got steamrolled in the same session. Sell-side desks have rewritten their AI infrastructure decks. Optics now leads the conversation; general-purpose compute follows.

The proof points are everywhere. Coherent (COHR) and Lumentum (LITE) earned S&P 500 additions out of OFC, which means passive flows now reprice these names mechanically. Applied Optoelectronics (AAOI) has run from roughly $10 in early 2025 to all-time highs near $155 in April 2026 — gains north of 1,140% — on the back of an alleged $53M order for 800G and a $200M+ order for 1.6T modules, with management projecting full-year 2026 revenue above $1 billion against fiscal 2025's $455 million. Lightwave Logic (LWLG), long treated by Wall Street as a science fair project, went parabolic in March on a definitive agreement with Tower Semiconductor and crossed a 939% one-year total return by mid-April.

Three different business models. Three different risk profiles. One unifying force: the AI buildout has pushed bandwidth, power, and reach past what copper can carry. Operators reported in early 2025 that interconnects were already eating nearly 30% of cluster power. At 224 Gbps signaling speeds, passive copper reach has shrunk to under one meter. The Copper Wall is not a slide. It is a load-bearing problem hyperscalers are spending billions to engineer around.

Where POET Sits in the Stack

Each of those names attacks a different piece of the buildout. AAOI assembles transceivers from discrete components. LWLG works at the modulator material level. COHR and LITE sell the broadest lineup of lasers and optical components.

POET is doing something architecturally different. The Optical Interposer integrates lasers, modulators, and electronics on a single chip-scale platform — eliminating the multi-step packaging and alignment work that has historically driven cost and power consumption in optical modules. The rest of the field is iterating on better components. POET is iterating on a better way to assemble them in the first place.

That difference matters as the industry moves toward co-packaged optics, where the optical engine sits adjacent to or directly on the switch ASIC. An architecture built for chip-scale integration from day one has a structural edge in that transition. It is a different bet on a different timeline, and it is the bet POET has been making for years.

Back to Business

Every small-cap technology story eventually arrives at the moment where the noise burns off and the company either delivers or it does not. POET is in that moment now.

The dilution debate is settled — over $300 million in cash is multi-year runway, not a near-term overhang. The short report has been answered, and the answer is shipping product, not a press release. The Q4 miss is in the rearview, with high-volume production landing inside the very next quarter the company will report on. The Iran shock has been digested. The sector backdrop has rotated from sympathetic to actively bullish. Passive flows, large-cap upgrades, and parabolic peer moves are all pointing the same direction.

None of which guarantees a smooth chart. The float is bigger now. Volatility is the rent you pay for living in a sub-billion semiconductor name with a hundreds-multiple revenue ratio. There will be drawdowns on macro pullbacks, on profit-taking after big peer-group runs, on every quarter where revenue ramps slower than the most caffeinated bull modeled.

But the bear case has narrowed. The bull case has widened. And the company is finally focused on the only metric that ends every photonics debate: units shipped. The wild ride was the prologue. The chapter where the photonics company actually delivers photonics starts on May 12.