OrangeGroveCapital

If you dont own $SIVE 3105 322310 6315 6451 6707 6830 $SOI going into 2027 you will be left in the permanent underclass.

I just interviewed a CEO who said three things that blew my mind: 1. We replaced our $600K Salesforce contract with a vibe-coded CRM, built within 3 weeks. 2. We will get rid of 80% of the SaaS we use internally. 3. If Anthropic doubled pricing, we would not change usage in any way.

It is so easy and obvious that people keeps ignoring it. $mu

$TSLA Sold 1/6 of my $TSLA @ $445. Think it has run too hot too fast. I don’t think Robotaxis scale significantly any time soon. Optimus delayed. Sure it might run up more, but I think I get a chance to buy back below $411.

Orange Grove Capital is of the opinion that $LMND is significantly undervalued today: $LMND should at worst be trading 4.5x higher or at best 8.5x higher. This prevents wonderful opportunity to the upside. Orange Grove Capital has built a starter position of 400 $LMND and will be looking to increase this significantly over the coming weeks and months. Full thesis as follows: (A) In-Force-Premiums - $10b by 2034 $LMND currently has $1b of In-Force-Premiums ("IFP") today. During their Investor Day Presentation in 2024 ("2024 IDP"), Maya Prosor, $LMND's Chief Business Officer, shared the $LMND aims to grow at a CAGR of 30% per year in the future, meaning it will take $LMND 9 years (2034) to reach $10b in IFP. It is important to note that $LMND management has not missed guidance over the past 17 quarters (source: 2024 IDP), and so achieving $10b IFP by 2034 is more likely than not. It bears mentioning that Management mentioned in the 2024 IDP that this growth from $1b to $10b will depend heavily on the car sector, so that will be one to watch closely. (B) Revenue Projections - $10b by 2034 Revenue is made up of (i) net earned premium; (ii) ceding commission income; (iii) net investment income; (iv) commission and other income. (i) Net earned premium will now be roughly 80% of Gross Earned Premium (as opposed to 45%, as $LMND previously reinsured 55% of its book, but has now dropped this to only 20%). Let's take to be $8b (i.e 80% of $10b). (ii) Ceding commission income has historically been approximately 25-30% of net earned premium when 55% was reinsured. Now that reinsurance has dropped to 20%, lets say this is 10%. 10% of $8b is $800m. (iii) Net investment income and (iv) commission and other income have historically been 17/18% of net earned premium, so lets say its $1.4b. This means that revenue is likely to be approximately $10b in 2034. (C) Net Profit Now to get to net profit figures, we must consider (i) Operating Expenses ("OpEx"); and (ii) Loss and loss adjustment expense + other insurance expense and subtract that from total revenue. (i) OpEx - $600m in 2034: OpEx minus Customer Acquisition Cost: $LMND shared in their 2024 IDP that their quarterly OpEx ex Customer Acquisition Cost ("CAC") has been holding steady at approx $85m. It sounds like $LMND fully intends for OpEx ex CAC to remain at this level even as the business expands - that was the general sentiment of the 2024 IDP. Customer Acquisition Cost: Looking at their quarterly CAC, it has been hovering around $40m (Q1 2025: $38.1m; Q4 2024: $40M; Q3 2024: $39.9m). Total OpEx: Assuming this holds constant, then the yearly OpEx for $LMND should be ($85m + $40m) * 4, which is $500m. Let's assume this increases by 20% in 2034. So, $600m OpEx in 2034. (ii) Loss and loss adjustment expense + other insurance expense - $6.8b in 2034: Loss and loss adjustment expense: This generally tracks the Gross Loss Ratio ("GLR"). Considering $LMND's GLR has been decreasing consistently and significantly, with the improvements in AI and their data sets I don't see why this can't hit 70% very conservatively. Let's say loss and loss adjustment expense is 70% of net earned premium, which is $5.6b (70% of $8b). Other insurance expense: I think the best way to think about this is that it is currently approx. 20% of NEP. Since this is likely tied to NEP (but has some items which will not scale linearly, such as headcount costs), let's say $LMND becomes slightly more efficient and gets this to 15% in 2034. I.e., other insurance expense will be $1.2b in 2034. (D) Fair value at 2034 Now assuming revenue is $10b at 2034, and OpEx is $600m; loss and loss adjustment expense is $5.6b and other insurance expense is $1.2b. That leaves us with $2.6b in profit. We'll then have to factor in tax of approx. 21% to get us to net profit of $2.08b. Apply a 30 P/E to this business, and it will be valued at $62.4 billion. Considering $LMND is valued at $3b today, this is a 21x opportunity in 9 years). Alternatively, $LMND is worth $26.46 billion today assuming 10% compounded cost of capital (i.e present value valuation) and no stock dilution. This is a 8.5x upside from here. Obviously these are all "target" numbers and assumes Management executes flawlessly. Key will be to observe their loss ratios and OpEx spend going forward, as well as how $LMND continues to scale into the car segment. (E) Fair value at 2030 All that being said, 2034 is admittedly very far out to look into, and most investors will likely look at a shorter time frame to value $LMND. Let's therefore look at 2030. In 2030: - IFP: 30% CAGR from $1b today = $3.7b - Net earned premium: 80% of $3.7b = £$3b - Ceding commission income: 10% of $3.7b = $370m. - Net Investment income + commission and other income: 17% of $3.7b = 629 million - Total Revenue: $4b - OpEx: assume $550m. - Loss and loss adjustment expense: 70% of NEP = $2.1b - Other insurance expense: 15% of NEP = $450m. - Profit: $900m. - Post-tax profit: approximately $720m. Value at 2030 applying a 30x P/E multiple = $21.6b. $LMND fair value today assuming 10% cost of capital = $13.41b, representing approx. 4.5 times from here. +++ If you found this research helpful, I'd be grateful if you can share this post to other people interested in $LMND.

My top 3 plays for the CPO buildout: 1. $SIVE - Lasers 2. FOCI (3363) - Fiber Array Units 3. Shunsin (6451) - Packaging

Update Longs 1/6/2026: Key Highlights: Sold $SNDK $207 -> $318 average with multiple trims and adds that accentuated this return. Was largest position by 2x. Sold $LITE from $240 to $390's average sale price. Portfolio Longs: $PDD ($119.3) $FLY ($25.3) $TE ($6.9) $ZETA (20.75) $FLNC ($20.8) $ALMU ($17.05) $ELVA ($7.72) $AMBA ($76.77) $SYM (69.0) Top of my focus list: $AEVA, $IMSR, $NVTS, $PONY, $VICR, $FTAI, $PATH, $ACMR, $ONDS, $OUST, $MELI, $POET, $MBLY, $MCHP, $ON

"But I’m very bullish on $SIVE since markets haven’t pieced together the connection to $POET and $MRVL yet." $SIVE 100mW 파워 부족론은 아키텍처를 모르고 하는 말인 것 같습니다. SIVE 100mW가 부족하다는 비판을 종종 봅니다. LITE, COHR, AAOI가 400mW ELSFP를 들고 나오는데 SIVE는 왜 100mW에 머무냐는 겁니다. 결론부터 말하면, 비교 자체가 틀렸습니다. LITE/COHR/AAOI의 400mW는 고출력 레이저 1개를 스플리터로 나눠 여러 SiPh 엔진에 분배하는 ELSFP 구조입니다. 1:8 분기 시 9dB 손실에 커플링/변조 손실까지 더하면 출발 파워가 400mW 이상이어야 링크 버짓이 성립합니다. LITE가 400mW UHP에서 800mW SHP, 나아가 1W로 bar를 올리는 것도 같은 논리입니다. (OFC 2026) SIVE의 경로는 두 가지입니다. 첫 번째는 Ayar Labs SuperNova입니다. 16파장 DFB 어레이가 TeraPHY 광 I/O 칩셋의 각 레인에 스플리터 없이 1:1로 들어가는 구조입니다. 채널당 100mW면 충분합니다. ECOC 2024에서 이미 16 Tbps 양방향 라이브 데모가 완료됐습니다. 경쟁력은 파워가 아니라 파장 수(8→16λ)와 채널 균일성입니다. 두 번째는 POET Optical Interposer ELS입니다. SIVE DFB 레이저와 POET 인터포저를 결합한 칩스케일 통합 ELS 모듈로, H1 2026 프로토타입 고객 데모를 목표로 하고 있습니다. POET는 단순한 부품 회사가 아닙니다. Marvell의 Celestial AI 인수 이후, Northland는 POET를 포함한 ELS 공급업체들의 장기 포텐셜이 구조적으로 개선됐다고 분석했습니다. 체인을 그리면 이렇습니다: SIVE DFB 레이저 --> POET Optical Interposer ELS --> Marvell/Celestial AI scale-up 광 인터커넥트 생태계 시장은 SIVE를 소형 InP 레이저 회사로만 보고 있습니다만, Ayar Labs의 NVIDIA 전략적 협업을 통한 미래 노출, POET를 통해 Marvell CPO 생태계에 동시에 노출된 구조가 아닐까요?