Lattice’s inline EPS masks a demand inflection the guide makes harder to fade

The investable message from this print is not that Q4 was a clean upside quarter, because it was not: EPS was exactly $0.32 against the $0.32 estimate, and revenue of $145.8 million beat the $143.2 million estimate by only +1.8%. What matters is that the quarter converted a slow recovery into a guideable acceleration, with revenue rising +9.3% QoQ and +24.2% YoY in Q4 FY2025, followed by Q1 FY2026 guidance for $158 million to $172 million and non-GAAP EPS of $0.34 to $0.38. What was priced in was a recovery that had already become visible through Q1 FY2025 revenue of $120.2 million, Q2 FY2025 revenue of $124.0 million, and Q3 FY2025 revenue of $133.3 million. What actually surprised was the step-up in the slope: Q4 FY2025 revenue reached $145.8 million, and the midpoint language on the call moved the debate from stabilization to acceleration. Ford Tamer put a stake in the ground with a commitment-style sentence rather than a soft recovery comment: “Looking ahead, our Q1 revenue guidance of $165 million at the midpoint, representing over 37% year-over-year growth, reflects our confidence in a strong recovery and accelerated momentum.” The word “accelerated” matters because the reported sequence had already improved from +2.3% QoQ in Q1 FY2025 to +3.2% QoQ in Q2 FY2025 to +7.6% QoQ in Q3 FY2025, and the company then printed +9.3% QoQ in Q4 FY2025.

That acceleration matters because the stock’s debate had been distorted by the depth of the 2024 correction and the noise in margins. Revenue fell from $170.6 million in Q4 FY2023 to $117.4 million in Q4 FY2024, a -31.2% YoY decline, and the business did not convincingly turn until Q3 FY2025 revenue grew +4.9% YoY. The Q4 FY2025 result changed the base effect from argument to evidence, with +24.2% YoY revenue growth on $145.8 million, but the headline EPS of $0.32 matching consensus made the result easy to dismiss. That dismissal is too simple. The earnings line used for the street comparison was flat to expectations, but the operating data show that management is pushing revenue through the model while holding spending discipline: Lorenzo A. Flores said full-year 2025 revenue increased 2.7% to $523.3 million, non-GAAP operating expense was reduced approximately 1% to $213.5 million, non-GAAP operating margin expanded 340 basis points, and EBITDA margin increased 320 basis points to 35%. Those are not one-quarter optics; they frame the Q1 FY2026 guide as the first quarter where the recovery can show up in both revenue and EPS.

The margin question is where the bear case has its best number, and it should not be ignored. Q4 FY2025 gross margin was 62.4%, down from 67.9% in Q3 FY2025 and below the 68.0%, 68.4%, and 67.9% pattern from the first three quarters of FY2025. That single datapoint looks especially poor against Lattice’s own pre-downturn levels of 69.8% in Q1 FY2023, 69.7% in Q2 FY2023, 70.0% in Q3 FY2023, and 69.7% in Q4 FY2023. The reason not to make Q4 gross margin the center of the thesis is that the company’s forward margin framework contradicts the idea of structural degradation: Q1 FY2026 gross margin is expected to be 69.5% plus or minus 1% on a non-GAAP basis. The conflict is real because the historical gross-margin table reports Q4 FY2025 at 62.4%, while the call guide is on a non-GAAP basis; those are different reporting bases and should not be blended. But for investment purposes, the forward number is the one that tests whether Q4 was a trough or a new run-rate problem, and the company gave a concrete band rather than vague repair language.

The operating leverage argument also deserves more credit than the inline EPS headline received. Q4 non-GAAP operating expense was $56.4 million, up roughly 5% sequentially and 7% YoY, so management did not manufacture the quarter by starving the P&L. Yet Q4 non-GAAP operating margin expanded 170 basis points to 30.7%, and EBITDA margin increased 90 basis points to 36.5%. The Q1 FY2026 guide carries non-GAAP operating expense of $59 million to $61 million, which means the company is allowing investment to rise while still guiding non-GAAP EPS of $0.34 to $0.38. Tamer’s Q1 EPS language was unusually explicit about the incremental economics: “Our Q1 EPS guidance of $0.36 at the midpoint represents nearly 65% year-over-year growth as we expect to continue to deliver earnings growth that is faster than revenue growth.” That is the variant perception in one sentence: the model is being priced as a cyclical rebound with margin risk, while management is guiding to a revenue recovery that should carry EPS faster than revenue if the Q1 gross-margin band holds.

Cash conversion strengthens that interpretation because the quarter was not merely a shipment rebound. Full-year 2025 GAAP net cash flow from operating activities increased to $175.1 million from $140.9 million in 2024, with GAAP operating cash flow margin improving to 33.5% from 27.7%. Free cash flow was $133 million with a 25.3% free cash flow margin, up from $120 million and 23.5% in 2024. The quarter-level numbers were even cleaner: Q4 GAAP net cash flow from operating activities increased to $57.6 million from $47 million in Q3, with GAAP operating cash flow margin of 39.5%, up from 35.3% in Q3, and Q4 free cash flow was $44 million with a 30.2% free cash flow margin, up from $34 million and 25.2% in Q3. The buyback adds a capital-allocation floor to the thesis without being the thesis: the company repurchased approximately 1,800,000 shares or $100 million in full-year 2025, completed that program in Q4, and received authorization for an additional $250 million. That authorization matters because it arrives after free cash flow expanded, not in place of it.

The second-order implication for the wafer suppliers is more direct than for end customers, because the data pack names no customers of Lattice but does name suppliers. For GlobalFoundries, which supplies FPGA wafers, TSMC, which supplies advanced FPGA wafers, and UMC, which supplies FPGA wafers at 28nm/40nm, the read-through is a demand normalization signal rather than a material revenue call on any one foundry. The relevant magnitude is Lattice’s revenue path: $117.4 million in Q4 FY2024, $120.2 million in Q1 FY2025, $124.0 million in Q2 FY2025, $133.3 million in Q3 FY2025, and $145.8 million in Q4 FY2025, followed by Q1 FY2026 revenue guidance of $158 million to $172 million. That sequence implies wafer demand moving with a product recovery that has now produced +24.2% YoY revenue growth in Q4 FY2025 and a guide described as over 37% year-over-year growth at the midpoint. Because no customer names are provided, there is no defensible named-customer implication to draw; the customer-side point is instead that the attach and ASP commentary points to servers without giving a named account. Tamer said, “So, you multiply this by the number of server units, our attach rate growing, and our ASP is growing as well from the $3, call it, to above 4.” The actionable read-through is to the foundry suppliers listed above, not to a named system customer the data does not provide.

The peer context also argues against treating Lattice as just another late-cycle semis rebound. In the subsector table, NVDA posted $81,615.0 million in revenue, 74.9% gross margin, and +85.2% revenue YoY, which is a different scale and growth regime. Lattice’s latest reported quarter was $145.8 million of revenue, 62.4% gross margin, and +24.2% revenue YoY, so the company is not competing with the AI accelerator growth curve on magnitude. The more relevant comparative point is the margin and recovery shape: Lattice’s Q1 FY2026 non-GAAP gross-margin guide of 69.5% plus or minus 1% would place its forward margin framework above MSFT’s 67.6% and below NVDA’s 74.9% and META’s 81.9% in the peer table, while its Q4 FY2025 +24.2% YoY revenue growth sits above MSFT’s +18.3% and GOOGL’s +21.8% but below META’s +33.1% and NVDA’s +85.2%. That comparison does not make Lattice an AI mega-cap proxy; it shows the market may be underpricing a smaller-cycle recovery whose forward margin structure remains closer to high-quality silicon economics than the Q4 FY2025 gross-margin trough suggests.

The call delivery supports the same conclusion, but with one caveat that PMs should track. The tone history shows Q1 FY2026 call sentiment at 0.53, up from 0.45 in Q4 FY2025, and guidance_tone at 0.44, up from 0.41. More important than the direction is the quality of delivery: tone_confidence rose to 0.44 from 0.23, while uncertainty fell to 36.3 from 52.2. That combination usually matters more than sentiment alone because it indicates management is using firmer language with fewer uncertainty markers. The caveat is that prepared_sentiment fell to 0.00 from 0.63 while qa_sentiment rose to 0.54 from 0.30, and qa_evasiveness increased to 36.5 from 5.2. The conflict says the scripted remarks were less promotional, while Q&A was more positive but also more evasive by the model. I would read that as a net positive for the bull thesis, because the hard guide gave the revenue and EPS bands, but it argues against paying for a full-year 2026 extrapolation until management narrows the bridge beyond Q1.

That delivery nuance fits the unusual breadth of the Q1 guide. Gary Mobley flagged on the call that “the revenue bracket in the guide this quarter is $14 million when it's normally $10 million,” and that is the cleanest way to state the remaining risk. Management is confident enough to guide Q1 revenue to $158 million to $172 million, but the band itself admits more dispersion than usual. That dispersion matters because a Q1 FY2026 result near the low end would still be up from Q4 FY2025 revenue of $145.8 million but would not prove the same acceleration embedded in the $165 million midpoint commentary. Likewise, non-GAAP operating expense guidance of $59 million to $61 million leaves less room for disappointment if gross margin does not return toward 69.5% plus or minus 1%. The market should not ignore that the Q4 FY2025 diluted EPS line in the quarterly history was -$0.06, while the street-comparison EPS actual was $0.32; those are different bases, and the investment case relies on the non-GAAP earnings path management guided, not on pretending GAAP and street EPS are interchangeable.

The key debate into next quarter is therefore not whether Q4 beat consensus enough, because +1.8% revenue surprise and 0.0% EPS surprise are too small to carry the stock by themselves. The debate is whether Q1 FY2026 confirms that Q4 was the handoff from recovery to acceleration. Confirmation would be revenue inside or above the $158 million to $172 million range, gross margin near the 69.5% plus or minus 1% non-GAAP guide, non-GAAP operating expense inside $59 million to $61 million, and non-GAAP EPS inside or above $0.34 to $0.38. A break in the thesis would be revenue near the bottom of the $158 million to $172 million range paired with gross margin failing to approach the 69.5% plus or minus 1% band, because that would imply the +9.3% QoQ Q4 revenue move was not carrying the margin recovery investors need. I would also watch whether tone_confidence stays near 0.44 and uncertainty stays below 36.3 on the next call; a relapse toward Q4 FY2025 uncertainty of 52.2 would make the wider guide look like a warning rather than prudence. If the company delivers the Q1 midpoint framework of $165 million revenue and $0.36 non-GAAP EPS, the market will have to stop treating this as an inline Q4 print and start underwriting a recovery with earnings growing faster than revenue.