FormFactor is the gatekeeper of wafer test contact for the highest-performance chips on earth, and the company has finally converted that position into margin. The proof is a margin structure that the market has spent two years doubting, and the quarter that settled it has just closed.
The most recent quarter shows the proof, and the print is the clearest evidence of the inflection yet. GAAP gross margin of 50.7 percent more than tripled the prior-year 37.3 percent. Revenue grew 31.9 percent to 258.2 million.
The open question is whether the Texas factory ramp and the multi-year restructuring can hold that margin line through a capital-intensive year.
FormFactor designs and manufactures electrical and optical test tools used at every stage of a semiconductor's life, from early characterization through high-volume production test. The business runs two reportable segments: Probe Cards, which sells contact cards and analytical probes to memory and logic manufacturers, and Systems, which sells probe stations, thermal chucks, and cryogenic rigs to developers and fabs. Together they make the company one of the few suppliers that touches a chip before it leaves the wafer and again when it is packaged.
The relevant competitive set is thin. In MEMS probe cards, FormFactor's main rivals are Technoprobe, the probe card unit inside ASMPT, and a smaller field of Asian suppliers, and no rival has the combined scale and contact-count capability of the Livermore-based company. On the Systems side the company competes against Advantest, Tokyo Electron, and the thermal-chuck specialists that sit inside those houses. The strategic position is distinctive: probe cards are consumable, design-specific, and bought repeatedly as wafer volumes grow, which makes them a share-of-test spend business rather than a one-time equipment sale.
Three structural shifts are re-rating that position right now. The first is the migration to high-bandwidth memory, where HBM stacks require dense, low-impedance probing across hundreds of thousands of contacts per card. The second is co-packaged optics, where silicon photonics moves chip-to-chip interconnect onto the package and creates an entirely new test step. The third is the reshoring of advanced manufacturing, which raises the value of North American capacity and qualified supply. FormFactor's revenue mix already shows the shift, with South Korea plus Taiwan now generating over 60 percent of sales as HBM and foundry logic demand concentrate in those two markets.
The consequence for shareholders is that the company's revenue is now tied to the two fastest-growing corners of the semiconductor industry, and the margin structure finally reflects it. The risk is concentration, and the company's own disclosure names SK hynix as a customer above the 10 percent threshold in multiple recent quarters.
The core product is the MEMS probe card, a wafer-sized contact plate that presses microscopic composite pins against every bond pad, bump, and through-silicon-via on a chip during test. Density and frequency reach are the practical moat, and a foundry or memory maker cannot swap suppliers without re-qualifying an entire card architecture. FormFactor's cards carry over 150,000 contact elements at spacings of just 40 microns, and the company probes at millimeter-wave frequencies for radio-frequency and automotive radar devices. Qualification runs through the full design cycle of a chip, which typically spans many months, so a win today is a multi-year revenue stream.
The second moat is design automation. Probe cards are custom to each wafer layout, and FormFactor has invested heavily in proprietary routing and trace-length adjustment tooling that compresses the design-to-fab cycle. When a customer's next-gen memory design spins up, the company can deliver a matched card quickly, and speed becomes a share win in a market where test capacity is the bottleneck of a chip ramp.
The newest product line is Triton, a high-volume co-packaged optics test solution launched into the Systems segment. The commercial significance is that silicon photonics is moving from lab prototypes into high-volume data-center production, and co-packaged optics is the test step that almost no other supplier can serve at volume today. The most recent quarter showed Systems revenue up 44.1 percent year over year, which the company attributed primarily to Triton sales, and once Triton is qualified into a customer's CPO line it becomes a recurring part of that line's test cost structure, so first-mover qualification compounds.
On the memory side, the moat is process depth. HBM cards require thermal matching across extreme test temperatures and stable contact force over millions of compression cycles, and FormFactor's thermal chuck technology vertically integrates with its own probe stations. That integration is why the DRAM probe card franchise is concentrated with the HBM leaders, and why 75 percent of the year-over-year DRAM growth in the latest quarter was driven by HBM. The structural takeaway is that FormFactor holds the highest-friction position in the wafer-test supply chain, and each of the three growth engines, HBM, co-packaged optics, and advanced packaging, increases the test complexity rather than reducing it.
The latest quarter, ended June 27, 2026, is the defining datapoint of the current cycle. The print is the strongest of the current cycle, and it marks the arrival of a cost structure that matches the revenue mix. Revenue came in at 258.2 million, up 31.9 percent from the prior-year quarter. It also rose 14.2 percent sequentially.
The profit build underneath is what changed the story. Cost of revenue was 127.3 million against gross profit of 130.9 million. The GAAP gross margin of 50.7 percent more than tripled the prior-year quarter's level.
Operating expense totaled 73.1 million, and the composition of that spend is the real story. The largest component was research and development, which shows the company still investing through the margin build. Selling and administrative was 37.2 million, and factory start-up costs were 4.9 million. The operating result was income of 57.8 million, or 22.4 percent of revenue. Net income landed at 56.2 million. That is 0.71 per diluted share, against 0.12 a year earlier. The per-share print is the line to watch going forward, because it strips out the working-capital noise in the absolute numbers.
The first-half picture is even more striking, and the cash generation is the part that convinces. Revenue of 484.4 million was up 31.9 percent year over year. Non-GAAP net income reached 109.5 million against 39.2 million a year ago. Free cash flow of 83.3 million came after 24.8 million of capex. The balance sheet supports the expansion. Cash and marketable securities stood at 345.6 million, and the revolving facility was undrawn at quarter end. Total debt remains modest under the building term loan. The cost of the shift is the restructuring, and the reward is a margin base the market has not yet fully credited.
Management guided the third quarter of fiscal 2026 to revenue of 270 million, plus or minus a small band. The midpoint implies continued sequential growth after a record second quarter. GAAP gross margin was guided to 52 percent plus or minus 1.5 points. Diluted EPS was guided to 0.75 plus or minus 0.09 on a GAAP basis. The non-GAAP figure is 0.86. The guidance embeds the assumption that the restructuring charges stay mostly non-cash and concentrated in the current fiscal year.
The largest execution risk is the Farmers Branch, Texas ramp. The company purchased the site in June 2025. The build-out capex is committed at 140 to 170 million. Production is expected to begin in the fourth quarter of fiscal 2026, with a ramp through the following year. Factory start-up costs are already visible at 4.9 million in the latest quarter. Management has guided the full-year start-up cost to the 20 to 25 million range. The risk is that ramp timing slips, or that utilization on the new line does not reach the level the margin model assumes, in which case the fixed cost base grows faster than revenue and the margin trajectory flattens.
The second risk is the restructuring itself, which is the largest one-time item in the current fiscal year. It consolidates the Carlsbad and Baldwin Park, California plants, with Baldwin Park closed in January 2026 and Carlsbad expected to run through December. The expected total charges of 32 to 40 million are led by property and equipment impairment. That component alone is 17.5 to 20 million. Severance is expected to run 10 to 12.5 million, with the rest covering lease impairments and other costs. The company has already booked 27.8 million of that total in the first half, and the execution risk is the transition: moving volume out of two California plants while standing up Texas is a multi-quarter capacity shuffle, and any quality or yield event during the move hits gross margin directly.
The remaining risks are concentration and the new product line. SK hynix has been a 10 percent or larger customer in each of the eight most recent quarters disclosed. One customer represented 22.9 percent of revenue in fiscal 2025, so a single HBM design change or pricing concession at that account would move consolidated results materially. Triton, meanwhile, shows early traction with Systems revenue up 44.1 percent year over year, but the installed base is still small. With 70.9 million of repurchase authorization still available, management paused buybacks in the first half of 2026 to fund the Texas ramp, which is a signal that the factory, not the share count, is the priority over the next year.
The bear case runs on three legs. The first is an HBM demand plateau. If the two HBM leaders slow their capacity plans for 2027, DRAM probe card revenue, now the largest growth driver, decelerates. The mix then reverts toward the lower-margin foundry and logic business that dragged margins through the prior downcycle. The second is a Texas ramp slip that pushes factory start-up costs above the 25 million guidance, compressing the margin bridge. The third is a transition-quality event at the California consolidations that forces a yield loss in the probe card segment during the move.
In a bear scenario, non-GAAP EPS lands in the low single digits for fiscal 2026. The valuation math on that outcome is straightforward and unambiguous. The multiple compresses from the current re-rated level back toward the historical 10 to 12 times operating earnings range, and the stock reverts toward its pre-HBM cycle valuation. The downside is real but bounded, because the company has 345.6 million in cash and marketable securities. With no meaningful debt, the balance sheet can absorb a 12 month demand pause without financing.
The most important mitigant is the diversification of the growth story. HBM is the largest driver, but foundry and logic probe card revenue grew 22.4 percent year over year in the latest quarter, and Systems grew 44.1 percent. Even if HBM growth normalizes, the co-packaged optics and advanced packaging test lines provide a second growth vector that is not tied to a single customer's memory roadmap.
The risk that matters most to the thesis is therefore not a demand shock but an execution one: whether the Texas ramp and the California consolidation convert the 50 percent margin level into a durable cost structure. The restructuring charge level is a proxy for how much of the old cost base is being retired, and the factory start-up line is a proxy for how much of the new one is being built. Both are visible in the quarterly print, which makes this a testable thesis rather than a narrative one.
The valuation anchor is the earnings bridge from the pre-HBM cycle to the current one. On a first-half annualized basis, non-GAAP net income of 109.5 million in the first half implies roughly 221 million for the full year if the second half mirrors the first. The Q3 guide of 0.86 non-GAAP EPS on about 80 million diluted shares implies a second-half net income in the low hundreds of millions. A reasonable full-year non-GAAP EPS is therefore in the range of 2.70 to 3.00.
Against that, the multiple framework has three reference points. The first is FormFactor's own history. In the memory trough of the prior downcycle the stock traded at a discount to 10 to 15 times non-GAAP earnings. The multiple sat in the upper part of that range during the normal phases of the cycle. The second is the probe card peer set, where Technoprobe and ASMPT's test-and-handling unit have historically traded at 12 to 18 times operating earnings in growth phases. The third is the broader semiconductor equipment group, which the AI cycle has pushed to the upper end of that range for leaders with visible growth.
The quantified scenarios are applied to a full-year non-GAAP EPS estimate of 2.80. The bear case, where the Texas ramp slips and HBM growth normalizes to the low double digits, carries fiscal 2027 EPS growth in the low single digits. It also carries a multiple reset toward 12 times, which prices the equity at roughly 34 per share against the current re-rated level. The base case, where the margin holds near 50 to 53 percent GAAP, supports a multiple in the mid-teens on higher EPS. The multiple in that case is in line with the current market pricing.
The bull case, where Triton becomes a genuine second growth vector and the Texas line ramps ahead of schedule, carries the multiple toward 18 to 20 times. The EPS growth in that scenario is 30 percent plus, and that is the scenario that creates additional upside. The framework conclusion is that the stock is priced for the base case to clear, with the restructuring and the Texas ramp as the two items that separate the base case from the bear case. The margin base, at 50.7 percent GAAP and 53.3 percent non-GAAP in the most recent quarter, is the item the market is most likely to underweight. That is because it arrived late in the cycle and came with 27.8 million of one-time charges that obscure the underlying cost curve.
The quarter that matters is the one that just closed. FormFactor delivered 258.2 million in revenue at a 50.7 percent GAAP gross margin. The first-half non-GAAP net income of 109.5 million is more than double the prior-year period. That is not a cyclical blip; it is the arrival of a cost structure that matches the revenue mix. The 1,500 basis point gross margin expansion in four quarters and the 31.9 percent revenue growth together describe a company that has converted a strategic position into a financial one. The 83.3 million of first-half free cash flow confirms it.
The named events of the cycle each carry weight. The Keystone Photonics acquisition in December 2025, at 20.6 million, bought the company the optical probing technology that Triton now commercializes, and it is the reason co-packaged optics test is a first-mover line rather than a follow. The Keystone Microtech partnership expansion in July 2026 extends the company's probe card manufacturing and distribution footprint into the Taiwan supply chain, which is where 36.6 percent of the latest quarter's revenue shipped. The restructuring plans, at 32 to 40 million in aggregate charges, are the price of retiring the California cost base that capped margins for a decade. The Farmers Branch purchase is the capex commitment that makes the next margin step possible. Each of these is a visible, dated, and measurable item in the filing, and together they form the execution spine of the thesis.
The counterargument deserves direct treatment: the customer concentration, with one HBM customer above the 20 percent threshold, is a real fragility, and the restructuring means a significant portion of the margin gain is purchased with one-time charges that the market should discount. The response is that the concentration is a feature of the HBM market structure, not of FormFactor's execution. The non-HBM growth in foundry and logic, up 22.4 percent in the latest quarter, provides a genuine second driver that does not depend on the same account.
The judgment is that FormFactor is fairly valued at current prices for the base case, and the equity carries meaningful optionality toward the bull case if the Texas ramp clears on schedule and Triton converts into a franchise. The items to watch over the next twelve months, in order of importance, are the Farmers Branch ramp timing and its start-up cost trajectory, the HBM mix share of DRAM revenue, the Triton contribution to Systems growth, the completion of the Carlsbad consolidation, and the resumption of share repurchases once the ramp is funded. The thesis is a margin-structure story with an execution clock, and the next two prints are the test.