$FORM EXECUTIVE OVERVIEW
FormFactor’s 2026 Investor Day materially strengthened the company’s strategic narrative by reframing FORM from a cyclical semiconductor test supplier into a leveraged beneficiary of 3 structural vectors: HBM, advanced logic/custom accelerator complexity, and co-packaged optics. The central claim is aggressive but internally coherent: revenue doubles from a 2025 base of approximately $785 million to $1.6 billion by 2030, non-GAAP gross margin expands to 55%, OpEx declines to 23% of revenue, and non-GAAP EPS rises from $1.27 to $5.00, implying approximately 15.3% revenue CAGR and 31.5% EPS CAGR over 2025-2030. The thesis is not merely end-market growth; it requires market share expansion, operating leverage, yield/cycle-time improvement, and defensible pricing in highly customized probe-card architectures. Management’s stated TAM rises from $3.1 billion to $4.5 billion by 2030, implying approximately 8% market CAGR, while FORM targets roughly 2x that growth rate. On a simple TAM share basis, FORM’s revenue target implies share rising from approximately 25.3% of its served market in 2025 to approximately 35.6% in 2030. This implies that roughly 43.5% of the incremental revenue can be attributed to market growth at constant share, while roughly 56.5% depends on share gain, new vector penetration, or category expansion. That is the critical underwriting point: the plan is not just a beta trade on AI infrastructure; it is an alpha thesis around proprietary MEMS, HBM test intensity, advanced packaging yield economics, CPO optical test leadership, and manufacturing scale execution. (FormFactor, Inc.)
The near-term evidence is favorable. FORM delivered Q1 2026 revenue of $226.1 million, up 5.1% q/q and 32.0% y/y, with non-GAAP gross margin of 49.0% versus 43.9% in Q4 2025 and 39.2% in Q1 2025. Management guided Q2 2026 revenue to $240 million +/- $5 million and non-GAAP gross margin to 49.5% +/- 1.5%, implying an annualized run-rate materially above the 2025 base and already within approximately 550 basis points of the 55% 2030 margin target. The Q1 mix is also highly instructive: Probe Cards generated $198.3 million of revenue, or approximately 87.7% of total revenue, and 50.5% segment gross margin, while Systems generated $27.9 million and 38.0% segment gross margin. The current earnings inflection is therefore overwhelmingly Probe Cards-led, not yet CPO-led. CPO remains an important medium-term option, but the investable near-term momentum is HBM, networking, and advanced logic probe-card demand combined with manufacturing yield/cycle-time improvement. (FormFactor, Inc.)
STRATEGIC REPOSITIONING
The Investor Day message was unusually explicit: FORM is positioned at the intersection of high-performance compute and advanced packaging, where test intensity and test complexity are both rising. The core technological argument is economically sound. Advanced packaging increases the value at risk before final assembly because multiple expensive die, interposers, substrates, and sometimes optical components are combined into 1 package. As stack heights and chiplet counts rise, the probability-adjusted cost of a latent defect rises non-linearly. Management’s HBM example is directionally compelling: a 1% incoming die defect rate is tolerable in a 4-high HBM2 stack, but becomes economically much more damaging in a 16-high HBM4 stack, where composite defect probability rises to roughly 15%. The practical implication is that wafer-level screening must become more comprehensive, more parallel, more thermally controlled, and more device-specific before packaging. That moves probe cards from a back-end consumable into an enabling technology for AI system yield.
This point matters because FORM’s products are not generic capital equipment. Probe cards are design-specific consumables tied to customer die layouts, mask changes, pad pitches, power envelopes, and test vectors. Every new GPU, ASIC, HBM generation, networking die, or CPO optical engine creates new customization requirements. This gives FORM a demand profile that is partly indexed to design activity and ramp complexity, not only wafer starts. The company’s economics therefore can decouple favorably from conventional semiconductor capital equipment cycles when AI accelerator cadence accelerates, even though the business remains cyclical at the customer-capacity level. The strategic significance is that a faster accelerator cadence, more reticle-limit die, more HBM stacks, and more chiplets per package all increase the value of early known-good-die assurance.
EXTERNAL DEMAND CONTEXT
External data broadly supports the direction of the FormFactor thesis. TSMC reported robust AI-related demand throughout 2025, 35.9% y/y revenue growth in US dollar terms, and record revenue and EPS, while Q1 2026 revenue reached $35.90 billion and Q2 2026 guidance was $39.0 billion to $40.2 billion with 65.5%-67.5% gross margin. This is relevant because TSMC is the central foundry and advanced packaging bottleneck for leading-edge AI accelerators, custom ASICs, and CPO-related platforms. TSMC also announced that its COUPE optical engine package is expected to begin production in 2026 and claims 2x power efficiency and 10x latency reduction versus a pluggable implementation, which externally validates FORM’s assertion that CPO is moving from lab activity toward production infrastructure. (Taiwan Semiconductor Manufacturing)
HBM demand also appears structurally supportive. SK hynix announced completion of HBM4 development and readiness for mass production in September 2025, with bandwidth doubled and power efficiency improved by more than 40% versus the prior generation. Micron announced in March 2026 that HBM4 36GB 12H was in high-volume production for NVIDIA Vera Rubin, with greater than 2.8 TB/s bandwidth and more than 20% power-efficiency improvement versus HBM3E, and also disclosed customer samples of HBM4 48GB 16H. Reuters separately reported in May 2026 that large technology customers were making unprecedented offers to secure SK hynix memory supply, including potential funding of production lines and EUV tools, while a source described available capacity as essentially zero. These external indicators are consistent with FORM’s HBM-driven probe card demand, although they also underscore customer concentration risk. (SK hynix Newsroom -)
The CPO backdrop is more nuanced. NVIDIA’s official photonics materials state that CPO switches with integrated silicon photonics provide 5x better power efficiency and 10x higher network resiliency versus pluggable transceivers, with Spectrum-X Ethernet Photonics available in H2 2026 and Quantum-X InfiniBand Photonics already positioned around 800Gb/s ports and large-scale AI fabrics. NVIDIA’s 2025 launch also identified TSMC as part of the silicon photonics ecosystem and described Quantum-X and Spectrum-X photonics switches as products intended to scale AI factories to millions of GPUs. However, Reuters reported in March 2025 that NVIDIA did not yet plan to use CPO directly in flagship GPUs because copper remained “orders of magnitude” more reliable for that use case, limiting CPO initially to networking switch chips rather than the main accelerator complex. For FORM, this distinction matters: CPO can still become a meaningful Systems growth vector through switch and optical-engine test, but the timing and size of the opportunity should not be extrapolated as if optical I/O were already broadly embedded across every GPU-to-GPU scale-up link. (NVIDIA)
REVENUE TARGET QUALITY
The $1.6 billion 2030 revenue target is credible enough to underwrite as a serious scenario, but not conservative. The company’s 2025 revenue was $785.0 million, with Probe Cards at $637.9 million and Systems at $147.1 million. By market, 2025 revenue included $369.9 million of Foundry & Logic, $247.4 million of DRAM, $20.6 million of Flash, and $147.1 million of Systems. That base is already meaningfully exposed to AI through HBM and advanced logic. The company’s 2030 plan depends on 4 principal growth vectors: HBM/DRAM, GPUs, custom ASICs, and CPO, with additional growth in the base business. Management’s target is not overly dependent on 1 product class, but the near-term contribution mix is clearly concentrated in Probe Cards, while Systems/CPO is still more forward-looking. (FormFactor, Inc.)
The strongest revenue argument is HBM test intensity. Management stated that HBM3/3E to HBM4/4E increases test intensity by approximately 25%-30%, driven by higher stack heights, more insertions, longer test times, and higher known-good-die requirements. HBM5 and hybrid bonding could create additional insertions, and custom HBM introduces logic IP in the base die, requiring a combined memory-plus-logic test capability. FORM’s competitive claim is particularly important here: it argues that it is the only supplier with production-proven platforms in both Foundry & Logic through Apollo and DRAM/HBM through SmartMatrix, enabling differentiation as custom HBM fuses logic and memory test requirements. If accurate, this could preserve share and pricing even if SK hynix/Samsung/Micron HBM market shares shift over time.
The second revenue argument is expansion into GPU and custom ASIC probe cards. Historically, custom ASICs were 2-3 nodes behind leading GPUs, reducing the need for the highest-performance MEMS probe cards. Management stated that custom ASICs are now moving toward 3nm-class nodes with higher core density, higher thermal load, and more advanced power/speed requirements, pushing them into the advanced MEMS category. This is strategically important because custom silicon demand from hyperscalers broadens the customer opportunity beyond the traditional GPU leader and reduces single-platform risk. However, the share-gain component is large. FORM disclosed in Q&A that it historically did not participate meaningfully in GPU test and expects major GPU producer demand to ramp in H2 2026. That means the 2030 revenue plan embeds successful penetration of markets where incumbents and customer qualification cycles matter.
The third revenue argument is CPO Systems. Management framed silicon photonics and CPO as a multi-hundred-million-dollar opportunity by 2030, with CPO wafer-level test SAM around $300 million-$400 million for the addressable wafer-level insertion set. FORM stated that it has installed more than 160 silicon photonics test systems globally, acquired Keystone Photonics in 2025, deployed multiple TRITON systems, and expects 2026 to be an inflection year as CPO moves from design/niche production toward high-volume manufacturing. The Keystone acquisition was small at $20.6 million net of cash acquired, but strategically relevant because it brings optical probing technology for SiPh/CPO wafer testing and expands FORM’s capability as SiPh and CPO transition into high-volume manufacturing. The opportunity is real, but it remains early, and the Systems segment’s Q1 2026 revenue and margin profile show that CPO is not yet carrying the P&L. (FormFactor, Inc.)
GROSS MARGIN AND EPS BRIDGE
The most important earnings issue is not whether revenue can grow; it is whether 55% non-GAAP gross margin can be achieved and defended. Management’s bridge from the 2025 baseline to 55% consists of volume leverage, operational excellence, and innovation. The stated components are approximately 600 basis points from volume, 500 basis points from operational excellence, and 400 basis points from innovation, while the stated headline gross-margin uplift is approximately 1,400 basis points. The component bridge therefore appears rounded and should be treated as directional rather than mathematically precise. The underlying claim is that 2025 margin pressure from manufacturing cost, tariffs, mix, and factory inefficiency is being replaced by better yields, lower cycle times, stronger utilization, higher-value product architecture, and a lower-cost Texas footprint.
There is clear evidence of operational momentum. Q1 2026 non-GAAP gross margin was 49.0%, and Q2 guidance implies 49.5% at the midpoint. Probe Cards gross margin was already 50.5% in Q1 2026, compared with 40.5% for Probe Cards in FY2025. This supports the claim that the business can move beyond the historical 40%-mid-40% gross margin regime when volume, yield, and mix align. However, Systems gross margin was only 38.0% in Q1 2026, and management acknowledged that some recent gross margin improvement included temporary elements. The next 500-600 basis points of gross-margin expansion will be harder than the 1st leg because low-hanging fruit, under-absorption recovery, and mix normalization have already contributed. (FormFactor, Inc.)
The 2030 EPS target is highly sensitive to gross margin. At $1.6 billion revenue, 55% gross margin, and 23% OpEx, implied operating income is approximately $512 million before interest/tax and other below-the-line items. Using the current share count implied by the market capitalization and current price, $5.00 EPS implies roughly $396 million of non-GAAP net income, consistent with an approximate 22%-23% tax/other drag from operating income. If gross margin were 50% instead of 55% at the same revenue and OpEx ratio, EPS would fall to roughly $4.20 under the same simplified assumptions. If OpEx remained 25% of revenue instead of 23%, EPS would fall to roughly $4.70. Therefore, the underwriting debate should focus heavily on the durability of >50% gross margin and OpEx leverage, not merely the revenue target.
FARMERS BRANCH EXECUTION
Farmers Branch is the operational fulcrum of the 2030 model. Management described the site as a brownfield expansion with more than 50,000 square feet of existing clean room, intended to expand MEMS capacity from Livermore and consolidate Carlsbad A&T operations into Texas. The strategy is “copy smart” rather than “copy exact,” meaning FORM plans to transfer proven technologies while incorporating more advanced toolsets, industry best practices, digital controls, automation, and AI-enabled inspection. The claimed benefits are higher revenue per clean-room square foot, lower COGS, better process control, and a more fungible manufacturing network. The modular ramp approach is financially prudent because capacity is added against customer demand rather than speculatively.
Execution risk remains meaningful. The 10-K states that Farmers Branch start-up costs are expected to be $20 million-$25 million in fiscal 2026 and that the site is expected to begin revenue-generating production in late Q4 fiscal 2026. The Investor Day transcript describes MEMS fab line and A&T qualification by December 2026 with initial production ramps starting in 2027. These are effectively consistent, but the schedule leaves limited room for slippage if customer demand is accelerating now. The company’s Q&A commentary indicates that existing industry probe card supply is constrained and that FORM is relying on yield, cycle-time, and asset-utilization improvements to bridge demand before Texas capacity becomes productive. A delayed qualification, weaker-than-expected yields after transfer, or higher-than-expected depreciation/start-up absorption could pressure both share-gain timing and gross margin. (FormFactor, Inc.)
CUSTOMER CONCENTRATION AND CUSTOMER QUALITY
The customer base is high quality but concentrated. In Q1 2026, SK hynix represented 29.5% of revenue, NVIDIA represented 10.2%, and total 10%+ customers represented 39.7% of revenue. In FY2025, SK hynix represented 22.9% of revenue, while Intel was below 10%; in FY2024, Intel represented 14.6%. This shift captures the company’s transition from a more diversified logic/memory profile toward AI memory and advanced compute/networking exposure. The positive interpretation is that FORM is becoming more embedded with the most strategically important AI supply-chain customers. The negative interpretation is that near-term earnings are increasingly levered to HBM allocation, SK hynix product cycles, and a small number of hyperscaler/GPU-related programs. (FormFactor, Inc.)
This concentration is partly inherent to the advanced semiconductor supply chain. The customers that matter for HBM, leading-edge GPU, custom ASIC, advanced packaging, and CPO are few, technically demanding, and globally scaled. FORM’s “customer intimacy” model, with early R&D engagement and lab-to-fab pull-through, is strategically appropriate for this market structure. However, it also creates dependency on customer roadmaps, qualification windows, purchasing behavior, and negotiating leverage. Large customers can value supplier reliability highly during capacity shortages, but they can also demand price concessions once supply normalizes. Management’s 55% gross margin target assumes that FORM captures a portion of the value created by higher yield and faster ramps rather than returning most productivity gains to customers through lower pricing. That assumption is plausible in constrained, high-complexity ramps, but less certain once applications mature.
COMPETITIVE POSITIONING
FORM’s competitive advantage appears strongest where custom design, MEMS performance, current-carrying capability, thermal control, high parallelism, low pitch, and signal integrity converge. The 10-K lists competitive factors in production probe cards including delivery time, probe card lifetime, chip damage prevention, touch-down accuracy, electrical speed, current-carrying capability, parallelism, probe tip density, pitch, signal integrity, and cleaning frequency. These map well to the HBM4/HBM5, custom ASIC, networking, and GPU challenges discussed at Investor Day. FORM also states that high capital investment, MEMS development cost, and lengthy customer evaluation cycles represent barriers to entry for MEMS probe cards. This is a credible moat, particularly when customers need first-time-right execution on compressed AI accelerator launch cadences. (FormFactor, Inc.)
The competitive landscape is not static. FORM identifies Japan Electronic Materials, Korea Instrument, Micronics Japan, MPI, STAr, Max One, Technoprobe, TSE, and others as probe-card competitors, with JEM, Micronics Japan, and Technoprobe offering MEMS-based probe cards similar in type to FORM’s technology. In Systems, TEL and MPI are among probe-station competitors, and CPO could attract adjacent test-equipment, optical-subsystem, or ATE ecosystem entrants. The company’s differentiation is therefore not “no competition,” but rather breadth across DRAM/HBM and Foundry & Logic, proprietary MEMS, early customer co-development, global support, and production scale. The moat is likely strongest in advanced probe cards and most uncertain in emerging CPO test architectures, where industry standards, insertion economics, and supplier boundaries are still forming. (FormFactor, Inc.)
CPO ANALYSIS
The CPO strategy is attractive because it offers a new Systems growth vector with potentially higher strategic relevance than traditional probe stations and thermal systems. FORM’s lab-to-fab positioning is logical: photonics test requires precise optical coupling, electrical test integration, automation, thermal stability, and wafer-level throughput. Early insertion testing can reduce cost of good die by catching defects before expensive package assembly. Management’s claim that insertion 1 testing can reduce cost of good die by up to 5x relative to later insertions is important, because it anchors the value proposition in manufacturing economics rather than theoretical test capability.
The risk is that CPO insertion strategy may evolve. Management acknowledged that later insertions exist but do not yet scale economically because of complex systems, longer cycle times, and low throughput. FORM is focusing on insertion 1 while preserving optionality to extend its optical probing technology into later insertions as high-volume manufacturing matures. This is strategically sound, but it also means the 2030 Systems growth plan depends on a still-forming manufacturing flow. If CPO remains mostly switch-focused for longer, if optical-engine yields improve enough to reduce test intensity, if test moves toward integrated ATE/handler architectures dominated by larger suppliers, or if customer-specific flows fragment, FORM’s Systems growth could be lower or more volatile than management suggests.
CAPITAL ALLOCATION AND BALANCE SHEET
Capital allocation is appropriate for the strategic phase. Cash, cash equivalents, and marketable securities were $303.3 million at Q1 2026, operating cash flow was $45.0 million, and free cash flow was $30.7 million. The 10-K disclosed a $150 million undrawn revolver at year-end 2025 and a $75 million repurchase authorization intended primarily to offset stock-compensation dilution. The company is prioritizing Farmers Branch, selective tuck-in M&A, R&D, and limited share repurchases, which is rational given the magnitude of organic opportunity and capacity constraints. The Keystone Photonics acquisition is a good example of targeted capability acquisition rather than scale M&A. (FormFactor, Inc.)
The balance sheet does not appear to be the binding constraint. The more important constraint is execution bandwidth: tool qualification, engineering hiring, customer sign-offs, supply-chain availability for ceramic/organic substrates and complex PCBs, and retention of technical talent. The 10-K explicitly notes dependence on suppliers for materials and critical components, including ceramic and organic substrates and complex printed circuit boards, with some components supplied by a single vendor. It also notes that most manufacturing employees are in California, Oregon, and Germany and that recruiting/training skilled manufacturing and engineering personnel is critical to capacity ramps. These risks are especially relevant when the company is simultaneously expanding Farmers Branch, consolidating Carlsbad A&T, advancing Apollo/SmartMatrix roadmaps, and absorbing Keystone. (FormFactor, Inc.)
VALUATION AND INVESTMENT IMPLICATION
At a current price of $141.78 and market capitalization of approximately $11.24 billion, FORM trades at approximately 28.4x the 2030 non-GAAP EPS target of $5.00 and approximately 162.7x current reported EPS. This valuation already discounts a meaningful portion of the 2030 earnings-power narrative. On that basis, the stock is less a simple “AI test equipment is growing” story and more a high-multiple execution compounder where upside requires either visible acceleration above the $1.6 billion plan, confidence that 55% gross margin is conservative, CPO optionality larger than management’s explicit SAM, or sustained premium terminal multiples. Conversely, any evidence of HBM normalization, Farmers Branch delay, margin giveback, share-gain slippage, or CPO adoption delay could produce multiple compression because the current price already capitalizes much of the management case.
The most balanced interpretation is that the Investor Day was strategically credible and operationally more convincing than a generic AI TAM presentation. The company provided a coherent link between advanced packaging economics and test intensity, demonstrated near-term margin improvement in Q1/Q2 2026, disclosed specific capacity and margin levers, and tied growth to externally validated AI infrastructure vectors. The main caveat is that the plan is ambitious by construction. More than 50% of the incremental revenue bridge appears dependent on share gain or category expansion beyond constant-share TAM growth, while the EPS target depends heavily on retaining operating-efficiency gains and achieving/defending a 55% gross margin in a customer-concentrated supply chain. The long-term narrative is therefore high quality, but not de-risked enough to ignore valuation. The proper underwriting stance is constructive on the business quality and strategic direction, selective on incremental capital at current valuation, and focused on 5 proof points over the next 4-6 quarters: H2 2026 GPU/customer ASIC probe-card revenue ramp, Q3/Q4 2026 HBM4/4E probe-card trajectory, Farmers Branch qualification milestones, sustained Probe Cards gross margin above 50%, and CPO revenue moving from the high end of the $10 million-$20 million 2026 range into a visible multi-year backlog.
