Tower Semiconductor: After a 200% Rally, Is the Silicon Photonics Pick-and-Shovel Play Still Just Beginning?

What is Tower Semiconductor and why is it suddenly on the radar? The Israel-based foundry, listed on Nasdaq and the Tel Aviv Stock Exchange as TSEM, does not design chips — it manufactures what others design, printing them onto silicon wafers sliced from ultra-pure ingots. That material is the gatekeeper that controls how much information or current moves through a chip. While names like Coherent give you direct exposure to photonics products, Tower sits one layer below, selling the manufacturing floor that the transition must stand on.

That positioning makes it a textbook pick-and-shovel play. Every silicon photonics package needs a silicon wafer, so you benefit from the common substrate without having to crown a single transceiver or networking winner. The clip leans into exactly that logic, calling Tower the lesser-known firm that makes the foundational layer — not the end device, but the ground every device has to be printed on.

The most talked-about fundamental is the earnings ramp. The estimate cited in the show puts earnings per share around $3.86 by year-end, doubling to $6.64 the following year and then nearly doubling again the year after — a projection already revised higher this year. The host calls it a massive ramp, contrasting weak-looking technicals with what he sees as phenomenal fundamentals: the price, in his view, has not yet fully priced the profitability inflection.

How does the market corroborate that thesis? Two signals are highlighted. First, the steady green institutional flow that the team nicknames the stairway to heaven on Money Flows, lifting the stock for months. Second, a clean sweep of earnings beats — all green — and an upward-sloping revenue trend. The late-July hedge-fund blow-up is presented as the exception that proves the rule: someone was forced to sell, the chart wobbled, but the beat streak and revenue slope stayed intact. The takeaway is explicit: the best opportunities often hide in smaller, less covered names.

The core photonics argument is supply, not demand. Only a handful of firms can grow ultra-pure silicon boules, slice them into wafers and print chips at scale, and demand is exploding. The show invokes Jensen Huang’s framing as a vote of confidence: push copper as far as it will go, then you need light. Light here means the whole photonics stack — fiber, lasers and related optics — integrated directly onto the silicon die. If that shift goes mainstream, the runway for the wafer layer is described as enormous precisely because the technology is still in its infancy.

That runway claim now has Q2 2026 receipts. Tower’s August 4, 2026 press release reports record second-quarter revenue of $460 million, up 24% year over year, with third-quarter guidance at $520 million — up 31% year over year and 13% sequentially. Gross profit was $138 million, up 72%, operating profit was $90 million, 2.3 times a year ago, and net profit was $91 million, up 95%, for diluted EPS of $0.79. The standout is SiPho: the annualized run rate jumped from $180 million a year ago to $680 million, triple-digit growth. CEO Russell Ellwanger guides to crossing a $1 billion SiPho run rate in Q4 2026 and, for 2028, a business model of $3.6 billion in revenue with $1.2 billion in net profit, described as already spoken for by customers.

So is there still runway after more than 200%? The second guest in the clip argues that pinpointing the slope is hard this early in a growth phase, and that the constraint is supply. When the bottleneck is not whether customers want more but whether anyone can make enough wafers, the few scalable foundries get reordered on every wave of photonics adoption. Even in the risk discussion that follows, the logic holds: knowing planes will not collide makes people more willing to board, not less — safety rules, in this telling, expand adoption rather than choke it.