Lattice’s miss is not the signal; the guide says the inventory trough is ending with margins intact

The thesis from this print is deliberately narrower than “recovery”: Lattice has likely moved from an inventory-clearance debate to an operating-leverage debate, and the Street’s reaction should hinge less on the Q2 revenue miss than on whether Q3 validates the midpoint guide. What was priced in was a company still struggling to escape the post-2023 revenue reset, with revenue expected at $131.7 million and EPS expected at $0.24. What actually surprised was the split: EPS landed exactly at $0.24, while revenue came in at $124.0 million for a -5.9% miss. That combination matters because the miss did not flow through to EPS on the Street-comparison basis, implying the investment question is not whether the model is broken, but whether the top-line shortfall was the last quarter of digestion before sequential growth resumes.

The reason the miss should not be dismissed, however, is that Lattice’s revenue base has not yet earned the benefit of the doubt. Revenue is still far below the 2023 peak run-rate, and Q2 FY2025 at $124.0 million was essentially flat with the year-ago period at -0.1% YoY. Gross margin is the counterweight: at 68.4%, it is back near the company’s pre-reset range rather than confirming the Q4 FY2024 trough at 61.1% as structural. The variant perception is that investors may be over-anchoring on the missed top line while underweighting the fact that margin has already normalized before revenue has recovered, which creates more torque if Q3’s revenue midpoint is real.

That financial trajectory makes the Q3 guide the center of the call, not the Q2 result. Lorenzo A. Flores gave the commitment plainly: “For Q3 2025, we expect revenue to grow into the range of $128 million to $138 million, with gross margin expected to be 69.5%, plus or minus 1% on a non-GAAP basis.” The wording matters because management is not merely calling for stabilization; the midpoint is above the Q2 actual and pairs growth with a non-GAAP gross margin level that would sit above the Q2 GAAP gross margin of 68.4%. In the company’s own reported framing, Fouad G. Tamer sharpened the point by saying, “We now expect $133 million in revenue at the midpoint or 7.2% sequential growth, the largest we've achieved in 3 years.” That is the sentence the market has to underwrite: if the guide is hit, the Q2 miss becomes stale; if it is not, the bull case loses its cleanest evidence.

The bridge between revenue recovery and earnings power is expense discipline, and the company gave enough data to make the case quantifiable rather than aspirational. Flores said non-GAAP operating expense was $51.8 million, with roughly 1% sequential growth, while non-GAAP operating margin expanded 150 basis points to 27.5%. That is the operating leverage PMs should care about: revenue in the company’s own account increased 3% to $124 million, yet operating income grew faster according to management’s comment that revenue was up 3.2% and operating income was up 8%. The message is not that the cost base is solved forever; it is that Lattice can convert modest sequential revenue growth into margin expansion before returning to its old revenue scale.

Cash flow supports that interpretation more cleanly than GAAP EPS, which was only $0.02 in the quarterly history for Q2 FY2025. Operating cash flow improved to $38.5 million from $31.9 million in Q1, and free cash flow reached $31.3 million with a 25.2% free cash flow margin. The company then used that cash aggressively, repurchasing approximately $46 million of common stock during the quarter. The buyback is not the thesis by itself, but it is a useful signal: through the first half of 2025, repurchases were $71 million, which management said equated to 100% of operating cash flow. A company still seeing severe demand deterioration would be less likely to return that much of the first-half operating cash flow while guiding sequential revenue up.

The product-cycle evidence is the second plank of the thesis, because revenue recovery without new-product mix would look more like channel refill than durable growth. Tamer said Lattice is “on track to have about a 70% growth year-on-year between '24 to '25 in new products revenue,” which is the most important non-quarterly number on the call. The AI disclosure adds a useful mix lens: Edge AI is expected to contribute high teens of revenue in ’25 and mid-20% of revenue in ’26. By end segment, management framed about 60% of AI-related revenue in Compute and Comms and about 40% in industrial, automotive and consumer markets. That split reduces the risk that the AI story is only a data-center attachment narrative, but it also makes the 2026 target more exposed to execution across fragmented edge markets.

The second-order read-through is clearest for wafer suppliers, even though Lattice did not name customers in the data pack. For GlobalFoundries, TSMC, and UMC, the Q3 revenue range of $128 million to $138 million implies better FPGA wafer demand than the Q2 miss alone would suggest. The magnitude is not enormous against those suppliers, but the sequencing matters: Lattice is guiding to revenue growth while gross margin is expected at 69.5%, plus or minus 1%, which argues against a recovery driven by heavy discounting or low-quality inventory flush. The supplier read-through is therefore incremental wafer pull for FPGA programs rather than a broad foundry-cycle call, with UMC tied to 28nm/40nm FPGA wafers and TSMC tied to advanced FPGA wafers.

The customer-side implication has to be expressed indirectly because the data pack lists no named Lattice customers, but management’s AI mix still creates named competitive context. Tamer said about 55% of AI applications are companion chips to “AI accelerators and GPUs and switches,” while about 45% are on the data path or running edge AI on the chip. That is not a claim that Lattice competes with NVDA accelerators; it is a claim that Lattice can monetize board-level complexity around them. The peer table reinforces the distinction: NVDA’s latest reported quarter showed revenue YoY of +85.2% and gross margin of 74.9%, while Lattice’s Q2 FY2025 revenue YoY was -0.1% with gross margin of 68.4%. Lattice is not participating in the same growth curve yet, but if even the companion-chip portion scales, the earnings beta can be meaningful because the gross-margin gap to NVDA is far smaller than the revenue-growth gap.

That comparative point also keeps the valuation debate honest. Lattice’s gross margin profile is closer to high-quality semiconductor and platform economics than its current revenue trajectory implies, but the top line has to prove it can grow again. Against mega-cap platform peers in the table, Microsoft reported 67.6% gross margin and +18.3% revenue YoY, while Google reported 62.4% gross margin and +21.8% revenue YoY. Lattice has the margin quality to be viewed as a scarce programmable-logic asset, but it does not yet have the growth proof. The print’s investable edge is that the next quarter offers a fast falsification point: Q3 midpoint revenue of $133 million would start to close that growth proof gap, while a miss would reframe Q2 as demand weakness rather than timing.

The delivery of the call strengthens the case that management wanted investors to shift attention from Q2 actuals to Q3 acceleration, but the tone data also says not to treat the script as risk-free. In the tone history, Q2 FY2025 sentiment rose to 0.40 from 0.24 in Q1 FY2025, and guidance_tone improved to 0.44 from 0.35. Prepared_sentiment jumped to 0.74, but qa_sentiment was only 0.28, showing the scripted message was much more positive than the analyst exchange. That split is exactly what one would expect in a transition quarter: management has a guide it wants the market to price, while investors are still interrogating the durability of demand after the revenue reset.

The most important nuance in the tone data is not the positive script; it is the lower evasiveness in Q2 FY2025. The qa_evasiveness index fell to 6.6 from 43.7 in Q1 FY2025, while uncertainty was 45.2 compared with 43.6. Those two numbers conflict mildly: management sounded less evasive in Q&A, but uncertainty did not fall. The interpretation is that executives were more direct about the recovery path, yet the business still carries enough demand uncertainty to keep the language from becoming cleanly low-risk. That is consistent with the call excerpts, where analysts pressed on whether 2026 could return to a 15% to 20% revenue growth target and whether OpEx can reach the long-term range of 30% of revenue. The company has not yet earned those longer-term numbers in reported revenue, but Q3 guidance is the first real attempt to make the bridge.

The debate also turns on inventory, because that is the cleanest explanation for why Q2 revenue missed while Q3 can still accelerate. Management said it expects “a continued improvement in channel inventory,” and tied that to a 69.5% midpoint gross margin and $0.28 EPS at the midpoint. The phrase matters because it frames the recovery as channel normalization rather than a one-quarter shipment push, and the EPS midpoint would be above the $0.24 delivered in Q2 on the company’s own non-GAAP basis. If channel inventory is truly improving, investors should see sequential revenue growth without gross-margin damage. If instead Q3 revenue requires concessions, the gross-margin guide should be the first place the thesis breaks.

The bear case is straightforward and cannot be ignored: the actual Q2 revenue miss was large relative to expectations, and the company’s historical quarterly revenue remains well below 2023 levels. The stock should not get full credit for a recovery until actual revenue exits the $120 million area and the guide converts into reported results. But that is also why the setup is interesting: the Street got the EPS number right at $0.24 despite missing revenue by -5.9%, and management guided Q3 revenue to $128 million to $138 million with non-GAAP EPS of $0.26 to $0.30. In other words, the company is asking investors to underwrite not a vague second-half rebound, but a near-term revenue step with EPS leverage and gross margin protection.

What to watch next is precise. For Q3 FY2025, the thesis is confirmed if revenue lands inside the $128 million to $138 million guide, with the midpoint of $133 million validating Tamer’s 7.2% sequential-growth claim, and if non-GAAP gross margin holds near 69.5%, plus or minus 1%. It is strengthened if non-GAAP EPS reaches the $0.26 to $0.30 range while non-GAAP operating expenses stay between $52 million and $54 million. It breaks if Q3 revenue misses the low end of $128 million, if gross margin falls below the guided band, or if management backs away from the new-products trajectory that it described as about 70% growth year-on-year between ’24 to ’25. The next quarter does not need to prove the entire 2026 growth story; it needs to prove that Q2’s revenue miss was the last echo of the trough rather than the first sign that the trough is lower than investors thought.