Altair
Skip to content

Trading Analysis Report: LITE

Analysis date: 2026-10-21 · Model: claude-opus-5-5 · AI-generated, not financial advice.

  • Analysis date: 2026-10-21
  • Rating: Underweight
  • Generated: 2026-10-21 11:18:22
  • TradingAgents 0.6.0: anthropic, deep claude-opus-5-5, quick claude-sonnet-5-5
  • Analysts: market, sentiment, news, fundamentals; research debate rounds 5, risk debate rounds 5
  • Data vendors: core_stock_apis yfinance, technical_indicators yfinance, fundamental_data sec_edgar,yfinance, news_data yfinance, macro_data fred, prediction_markets polymarket

V. Portfolio Manager Decision

Rating: Underweight

Executive Summary: Underweight LITE: reduce to the lowest of two sizes. The first is the size where a ~45% one-day earnings gap costs no more than 1% of the portfolio (about a 2.2% position for LITE alone; if you hold AI/optics peers, include them at a 10-15% read-through move). The second is 50% of target weight, not drifted weight. If the filing checklist is still open at the backstop, default to 25%; if you are already at or below the retained size, sell nothing, and if flat, stay flat. Sell about half the reduction over the next 2-3 sessions (plan on fills near 858), place the rest as limits near 902 and 931 as bonus fills, and sell anything unfilled at market on a price-blind backstop at the Oct 27 close, or earlier if confirmed Meta or LITE report dates require it. Buy nothing before the print and never average down; a weak guide or a weekly close below 636 cuts the position to 0-25%, 1016 is a post-print reassessment marker only, the horizon is 1-3 months through the Sep-quarter print, and the target is 787.

Current Price: 857.96

Price Target: 787.0

Confidence: Medium

Time Horizon: 1-3 months

IV. Risk Management Team Decision

Aggressive Analyst

Aggressive Analyst: I'll open the debate, since nobody else has spoken, and answer the arguments I expect from both of you.

The first charge will be that trimming to 40-50% of a standard position is timid. I think it's the aggressive call, because it acts on the numbers instead of hoping the story carries us through the print. Aggression means taking risk where you have an edge. The plan's own math says there is roughly zero edge at full exposure: 36-48% required odds against a 40-50% judgment. So the bold move is to size to what the evidence supports and keep the optionality.

LITE is up about 5.6x from its 52-week low. It trades around 26x sales and roughly 96x EBITDA, and the 23.9 forward P/E only works if you believe $35.92 of forward EPS against $1.50 of latest quarterly GAAP EPS, about $6 annualized. That is a stock paying for a six-fold step-up. We've also just seen how it trades when conviction wobbles. It fell from 1133 to 822 on a closing basis, about 27%, in roughly two weeks with no earnings report. If a no-news wobble does that, a print against this bar can do worse, which is why the plan sizes for a one-day gap of about 45% costing no more than 1% of the portfolio.

To the neutral analyst, who I expect to say hold: yes, the weekly and monthly SuperTrends are up, operating income went 7 to 64 to 174 million, and the PEG is 0.63. I agree, and that's why this is a trim, not an exit. But a trend that's intact on the weekly chart is a reason not to sell everything, not a reason to carry full size. The weekly stop is 635.87, about 26% below here, so that argument lets the position fall another quarter before it says anything. Meanwhile the daily SuperTrend is down, MACD has crossed bearish, price is under the 10 EMA and 50 SMA, and the 50 SMA has gone flat.

On "cheap," a PEG of 0.63 rests on a consensus number that hasn't been delivered. On reported earnings it isn't cheap.

On the 200-day bounce story, one StockTwits post says the last touch ran 56% in nine days. That is unverified and anecdotal. The 200 SMA is 784.49, almost exactly the 785 intraday low on 10-20, and a print can gap straight through it, so the plan refuses to size against it.

On "it's already oversold," it isn't. RSI is 41.89, the daily z-score is -1.12 and MFI is about 39. A 24% drop with no flush means no capitulation has cleared the decks. The 4.3% bounce came on about 2.5M shares against 3-7.6M on recent days, and the bar may be incomplete.

On the demand narrative, "sold out through 2029" is a headline I can't verify. Even if it's true, it cuts both ways. A capacity-constrained company's revenue is capped by what it can build, capex is $284M over nine months against $177M a year earlier, and one headline already flags production bottlenecks. The demand upside is largely priced in, while the execution downside isn't capped.

To the conservative analyst, who I expect to say sell it all: the daily TD-9 is at +8, one bar from completing, so dumping at market near 858 may sell the low of a bounce. That's why the plan uses limit orders in 2-3 tranches toward 902 and 931, making the market come to us. A full exit also abandons a company that is executing. Operating income and FCF have risen for three straight quarters, cash is about $2.04B, and the convertible overhang looks largely cleared. If the print is a blowout, the retained core participates. A close above 1016, the daily SuperTrend stop, tells us the trim thesis is wrong, and we reassess instead of arguing with the tape.

On the open items, I won't pretend the -$6.94B TTM loss is a smoking gun. It conflicts with three profitable quarters and equity rising to $4.64B, so it's probably a data error or a non-cash item. But unresolved is unresolved, and I wouldn't carry full size into a print with trailing earnings unverified, along with the dilution question and the cash drop of about $574M in the June quarter. The plan rounds that outflow higher, so that number needs checking too. Insider selling of about $62M in six weeks with no buys isn't a bullish tell either. Some of it may be plan-driven, so I'm not leaning on it.

The aggressive part comes after the trim. It converts exposure into dry powder. If the print is strong, we still hold half. If it breaks, we have cash and levels to redeploy at, like 785 and 636, instead of averaging down from 900. Rates give no help either, since the market prices 97% odds of no Fed cuts in 2026. Any support has to come from earnings, and earnings are a binary event. I'll concede that an 8% recession probability is a real tailwind for demand. That is another reason this is a trim and not a full exit.

I don't know your actual holdings, so apply the sizing to your own book. If you're flat, stay flat, with no starter position. If you're long, take the 40-50% allocation, sell into 902-931, never average down, and cut to 0-25% on a weekly close below 636 or a weak guide. Don't take full size when the edge is zero. Aggressive Analyst: I predicted the neutral would say hold and the conservative would say sell everything. Neither did. The conservative backs the sell with 25% or less retained, and the neutral keeps the trim and adds a schedule. So we're really arguing about three things: how much of LITE to keep, how to get down to that size, and what we do after a gap.

First, where I was wrong. The neutral is right that my six-fold step-up was anchored to the March quarter. If June gross profit is around $554M and operating expenses sit near $175-183M, June operating income is roughly $370M. That's inference, since the June income statement isn't in the feed. Against that, $35.92 of forward EPS is a two-and-a-half to three-times hurdle, not six. But that still means the average quarter in the forward year has to earn about two and a half times what June probably did. We also don't know which fiscal year that number covers. The company has to do that while capex is up 60% and a headline already flags bottlenecks. A lower hurdle is still a hurdle, which is why I keep half and not all.

I also leaned on equity rising to dismiss the -$6.94B loss, and the neutral's reading is better. An induced-conversion charge would be booked as a loss and credited back to equity, so a huge non-cash charge can coexist with rising equity. That fits the roughly 75M weighted shares against 88-90M now. It's still inference until the filing confirms it. It also undercuts my claim that the convert overhang looks largely cleared. Current liabilities are still $2.48B against $393M a year earlier, and debt-to-equity implies about $1.67B of debt. I can't tell what's in there, so I withdraw "largely cleared." The plan's "roughly $700M" cash outflow is $574M in the data, so that line needs fixing too. And the conservative is right about the 785 and 636 buy levels. If a print can gap through them, I shouldn't pre-commit buy orders there. Same standard for me on the headlines: I won't lean on the bottleneck piece any more than the others lean on the 56% bounce post.

On retained size, I side with the neutral over the conservative's 25%. The conservative's arithmetic is right: a 45% gap takes the stock to about $470, and a 1% loss at that gap means roughly a 2% position. But the plan already sizes to that gap budget, so the smaller number already wins by construction. If your standard allocation is above about 4.4% of the book, the budget binds and you end up under half. If not, half already fits. The preference for 25% isn't derived from anything. The plan's own numbers are a 40-50% judgment against 36-48% required odds. At the midpoint that's 45 against 42, a small positive edge, and half size is what that supports. Quarter size is what you'd hold if you believed the pessimistic end of the range. The evidence doesn't put us there: operating income went 7, 64, 174, June looks stronger, cash is $2.04B, and the weekly and monthly trends are up.

The conservative is also right that a weak guide arrives as a gap and nobody can sell it. But no one claimed the 0-25% rule protects against the first move. It decides what we do with the remainder the next morning, and I'd rather have a rule than a hope. Sizing is the real control, and the weekly 636 level is a process rule for the primary trend.

On execution, the conservative and neutral are both right that limit-only leaves the first tranche exposed. The neutral's estimate is about 70% odds of touching 902 and a bit over half for 931 in the sessions before an early-November report. That rests on assumed volatility, but a 7.9% ATR supports it. So I accept the thirds, the end-of-October deadline if the date can't be confirmed, and a market backstop about three sessions before the report. I'd disagree on one thing. The conservative treats 902 and 931 as obstacles, but the 10 EMA and 50 SMA are where the daily trend has been rejected, so they're where I want to be the seller. Selling the first third into the next session or two is fine as long as it's only a third and not a dump on a bar that may be incomplete.

The collar or put spread is the one idea I find interesting. It keeps the upside and caps the gap, which is the aggressive way to hold risk. But I have no options data, and with ATR near 8% heading into a print, I'd expect protection to be expensive. Selling a call to fund it caps the blowout the retained piece exists to capture. Price it first. If the premium is steep, selling shares is cheaper insurance.

On 1016, the neutral is right. A close there is 18% above here and almost certainly means the print worked. That's a reassessment with half in hand, not an order. The cost of the trim in that scenario is the upside on the half we sold, and I'll pay that premium.

The aggressive part is what the cash does afterward. A clean print that still gets sold because the bar is high, with the TTM loss reconciled, the share count known and capacity commentary clear, is where I want to buy hard. That's why there's no buying before the print.

If you're flat, stay flat. If you're long, get to the gap budget on the schedule, never average down, and rebuild only after the print. Aggressive Analyst: We've converged more than any of us expected, so I'll say what I'm giving up and then spend my time on the one slice that's still open.

I concede four things. The neutral is right that I can't read a three-point edge out of two ranges that are each ten to twelve points wide. I'll stop calling it a small positive edge. If the edge straddles zero, it can't choose the size, and the risk budget has to. You're both right that the 10 EMA and 50 SMA aren't validated rejection points. They're free limit prices as long as a backstop exists. The neutral is also right that I shouldn't wait for the print to learn whether I'm allowed to buy afterward. The TTM loss, the share count, the $574M cash outflow (the plan says $700M, so fix that) and the roughly $1.67B of debt all sit in a filing for a closed quarter, so I'd read it this week. And the conservative is right about the basket. LITE has a beta of 1.54 and trades with AAOI, COHR and CRDO, and Meta's capex commentary can land before LITE reports. If your book holds those names, the 1% gap budget covers the group, and the LITE slice may end up well under half of standard. That's the budget working.

On execution I'll take the front-loaded version. Sell about half the reduction over the next two or three sessions, spread across days. Today's bar may be incomplete, and the daily TD-9 could complete on the close, so I don't want one dump on a possible low. Leave two limits near 902 and 931 for the rest. The conservative's driftless-walk point is a good one: 785 is about as likely to be touched first as 902, so the backstop does the real work. I'd put it at about the October 29 close if the date isn't confirmed. If the date is confirmed for early November, count back four to five sessions and use the earlier one.

What's left is how much we keep. The conservative wants a 25% cap until the filings are clean and 40% at most after. The neutral wants the budget to set the number, with a 50% ceiling and a move toward 25% if the filings disappoint. I side with the neutral, for three reasons.

First, the cap is tied to homework that takes one sitting, and the first half of the reduction takes two or three sessions. By the time we decide the last slice, the check is done, so the cap rarely binds. If it somehow isn't done by the backstop date, I'll accept a default of 25%. That's a fair price for not finishing the work.

Second, look at how small the disagreement is. Take a 4% standard position. The ten points between 40% and 50% is 0.4% of the book, and at the plan's 45% stress gap that's about 0.18% of the portfolio. And if standard is above about 4.4%, the budget binds before either number. We're arguing over less than a fifth of one percent of the portfolio in the stress case, while the stake is in a name where operating income went 7, 64, 174 and, on inference, something like 370 million.

Third, the argument that we'd cut the retained half to a quarter after bad news, so we should be there now, proves too much. The post-gap cut is conditional on the thesis breaking. A weak guide means the capacity story cracked, the stock has gapped, and the trend is under pressure. Being at 25% unconditionally means being at 25% in the branch the stock is priced for, with a 23.9 forward P/E that only works if the number is delivered. By the same logic we'd already be at zero, because we'd also cut to zero below 636 on a weekly close. A conditional rule isn't a reason to pre-pay the whole cut.

I'll also grant the conservative's hurdle arithmetic as a fair illustration. Averaging $800 million a quarter from about $370 million means roughly 30-35% sequential growth, four quarters running. But the last three sequential steps were roughly 9x, 2.7x and 2.1x. The capacity limit is the real constraint, and capex up 60% is the supply response. We also don't know whether consensus is GAAP or adjusted, or which fiscal year it covers. That uncertainty cuts in both directions, which is why we keep half and don't add. It's also why I'm not leaning on "sold out through 2029" or the bottleneck headline.

On options, I'd price a collar before using one, and the premium is probably rich with ATR near 8% going into a print. The default is to sell shares.

Here's where I land. Confirm the earnings date and do the filing check this week. Sell about half the reduction over the next two or three sessions, leave limits at 902 and 931, and keep the market backstop about four to five sessions before the report. Retained size is the lower of the gap budget, covering the basket if you hold peers, and 50% of standard. It drops toward 25% if the filings show dilution still ahead, if the loss isn't the non-cash conversion charge we suspect, or if the check isn't finished by the backstop. Treat 1016 and the weekly 636 as reassessment markers, not orders. Buy nothing before the print, never average down, and keep the cash for the post-print rebuild, when the unknowns are already resolved.

That's still the aggressive play. We stop paying for exposure the numbers don't support, we keep a real stake in the strongest operating turn in the data, and we have both cash and clean information if the market overreacts either way. So the Sell stands as a trim, not an exit. If you're flat, stay flat. Aggressive Analyst: We've converged, so I'll say what I'm changing, where I still think you're reaching, and then give the version I'd hand the trader.

First, the $700 million. I said twice that it should be corrected to $574 million, and the neutral's arithmetic is better than mine. TTM free cash flow of about $225 million minus the $104 million through March implies roughly $121 million of positive FCF in June. If cash still fell $574 million, then about $695 million left through investing or financing, which is almost exactly the plan's number. That's inference from the same data, not a reported figure, but I was treating a different question as a typo. The right move is to ask what it was. Cash-settling converts is balance-sheet cleanup and arguably good news. Buying securities or an acquisition says something else.

The conservative is also right that my "buy hard" was a reflex. A clean November print retires one step of four, not the risk. I'll take the neutral's three-step rebuild: the first third within a couple of sessions of the reaction if the checklist was clean and the guide holds, the second on trend repair, and the third on a retest that holds. I prefer the neutral's version to waiting for the daily trend to stop making lower highs. The print is an information event, and if the guide clears the bar and the stock sells anyway, waiting for the chart means buying after the move.

On the schedule, the conservative identified the most likely way this plan fails. I said a close above 1016 means reassess. Before the print, that's wrong. The reason for the trim is the gap budget, and a higher price doesn't shrink the gap. A move toward 1016 raises the bar while the unknowns stay open, so we finish the schedule. 1016 is a post-print marker only.

On the backstop, I side with the neutral: one date, written down today. That's October 29, or the session before Meta if Meta lands earlier, or four sessions before LITE reports, whichever comes first. Neither Meta's date nor LITE's is confirmed in our data. There's one consequence nobody has said out loud. With the backstop on October 29, the limit window is about six sessions, not ten. Using the neutral's own assumption of 4-5% daily moves and no drift, I get roughly 60% odds of touching 902 and mid-40s for 931. So the backstop will do real work. The plan's 902.56 headline is no longer the price for the whole reduction. It's the upside case for the last slice, and the expected exit for the first half is near today's 858. The trader should update that.

I agree with the neutral that tranche precision isn't where the risk is. An unfilled slice sold 15% below a limit costs under 4% of a standard position, while a 45% gap on a half-size position costs about 22%. The risk is in what we hold through the print.

I like the pass-fail checklist, with one addition. It cuts both ways. If the filing shows a one-time non-cash conversion charge, a fully diluted count within a few percent of 90 million, debt covered by cash or maturing beyond a year, and an explained outflow, then the two biggest bear items, a mystery loss and dilution, are gone. The retained piece should go to the cap then, not sit at 25% out of habit. The 25% default should only mean we didn't finish the homework by the backstop. It shouldn't be a view on the stock. The check is one sitting, so do it today. I'd also add the one the conservative flagged: is $35.92 GAAP or adjusted, and for which year? Adjusted EPS usually runs above GAAP, so if it's adjusted, the hurdle against GAAP operating income is lower than the neutral's dollars suggest. If it's GAAP, it's tougher.

Here is why this is still the aggressive play. The neutral put the hurdle plainly: do what June did, about $170 to $230 million of added operating income, four quarters in a row. That's demanding but not heroic, and the operating turn of 7, 64, 174 and roughly 370 million is the strongest thing in the data. That's why we keep half. We sell the other half because this stock fell 24% in two weeks without a print, and the print is a binary event. The cash isn't for sitting out. It's for acting on verified information, with a rebuild rule, while others are guessing.

So the Sell stands as a trim, not an exit. Do the filing check today and confirm both dates. Sell about half the reduction across the next two or three sessions, not in one block. Leave limits near 902 and 931 as bonus fills, and run a price-blind market backstop on one date. Retained size is the lower of the gap budget, covering the whole AI basket if you hold peers, and 50% of standard, defaulting to 25% if the checklist is open at the backstop. Never average down and buy nothing before the print. Treat 1016 and a weekly close below 636 as reassessment markers, not orders. Rebuild in thirds afterward. If you're flat, stay flat. Aggressive Analyst: I'll take the new points in order, say what I'm changing, and then price the one thing I think we've been underweighting.

Conservative, your debt point is the best thing said in this round, and it knocks out one of my own pillars. I kept citing $2.04 billion of cash as proof the balance sheet is fine. If roughly $1.67 billion of the debt is in-the-money convertible paper, cash against face value tells us very little. The neutral's refinement helps, and I'd add that net settlement means you pay principal in cash and the excess in shares. If principal goes out in cash, $2.04 billion less $1.67 billion leaves about $370 million, which looks thin next to a capex ramp already at $284 million over nine months. If it settles in shares, you get the 6-18% dilution the neutral sketched. Either way it costs something, and the filing tells us which. The $1.67 billion is itself inferred from a debt-to-equity ratio.

One more clue, with the same caveat. If the roughly $7 billion charge is a conversion loss, it's about what was handed over minus the carrying value of what was retired. Liabilities fell about $1.4 billion, so the consideration was worth several times book. That would mean those notes converted far below market. If the remainder was issued on similar terms, face value understates it by a wide margin, and the neutral's 6 to 17 million shares at $858 is $5 to $15 billion of conversion value. That is why the checklist needs teeth, and why I accept 25% as the default if it isn't finished by the backstop.

On the backstop, I'll stop hedging. Meta usually reports the last Wednesday of October, which would be the 28th. The session before is the 27th, four sessions from today, and the 29th is six. I haven't confirmed either date, and nobody here has. Pick one date today and write it down. Plan on the first half going out near 858, with the limits at 902 and 931 as bonus fills.

You're right about "standard." It's a target weight, and after a 5.6x run the position may have drifted well above it. Half of target could mean selling most of what you actually hold, which makes the tax point real. I'd sell highest-basis lots first and stop at the budget. The trim is also a ceiling, not a quota. If you're already at or below the retained size, you sell nothing. For the rebuild I'll take the neutral's version: the lower of target weight and the gap budget for the window you're in, rerun before each print. Rebuilding to a drifted weight would be a mistake.

The thing I want to price is the cost of being wrong. If the print works and the stock gaps toward 1,016, the half we sold gives up about 18%. On an illustrative 4% position, that's about a third of a percent of the book. The stress case it protects against is a 45% gap on that same sold half, about 0.9% of the book. I'd pay that premium. But the conservative is right that the most likely way this plan fails is the group watching a bounce and deciding the thesis was wrong. If the stock runs to 930 or 1,000 before the print, the cost of the trim is already priced above, and the schedule doesn't change. The failure on the other side is the backstop selling into a decline. I've made peace with that, because an unfilled slice sold well below its limit costs a few percent of a standard position, while what we hold through the print can cost many times that.

I also agree the clean checklist removes unknowns, not the bar. The bar is still roughly "do what June did, four quarters running" in a capacity-limited build. A clean check lifts the cap to 50% and does nothing more. The first rebuild third also has to wait for room in the post-print budget.

So the Sell stands as a trim, and the trader's write-up needs these changes: - The 902.56 entry is the upside case for the last slice, not the price for the reduction. - The 1016 stop is a post-print marker only. Before the print it isn't a trigger, and a rally doesn't pause the schedule. - The cash line should read "about $700 million left through investing or financing, unidentified." - The checklist gets the guide, the if-converted share count and settlement terms, and whether $35.92 is GAAP or adjusted and for which year. - Retained size is the lowest of the gap budget (LITE at the stress gap, peers at a read-through gap) and 50% of target weight, defaulting to 25% if the checklist is open at the backstop.

What makes this the aggressive play is that we decide the size before the event instead of discovering it during the gap. We keep a real stake in the strongest operating turn in the data, and the cash comes with a rule. Never average down, buy nothing before the print, and rebuild in thirds once the loss, the share count and the guide are all known. If you're flat, stay flat.