To The Moon

Here's how I model $Applied Optoelectronics(AAOI)$ post the $600M ATM offering AH on a Friday...: 1. Revenue: Still fair to assume $5.6B in FY28 revenue (probably slightly conservative excluding ELSFP contribution). I've been over these numbers a lot. 2. Margins: 22% non-GAAP net margins is a solid base case for 2028. At $5.6B revenue scale, we should see operating leverage expansion. We see $Lumentum(LITE)$ guiding towards 36%, Innolight ~31%, and Eoptolink ~37%. 22% is also currently in line with analyst estimates (who have so far been conservative on the $Applied Optoelectronics(AAOI)$ numbers but best to take a

$Tempus AI(TEM)$ is still wildly undervalued. Here's how I think the business should be valued over the next 12 months: DATA & APPLICATIONS -> Annualized revenue based on Q2 is currently ~$410M. Growth at a similar rate into FY27 would put us at $525M. -> This business has 73% gross margins so at 28% YoY growth and those margins it's putting up figures like a solid SaaS company. -> $525M * 10x sales = $5.3B -> $525M * 12x sales = $6.3B DIAGNOSTICS -> Generated $293.3M in Q2 alone which is $1.2B annualized. Growth at a similar rate gets us to $1.5B in FY27. -> And plus, $Tempus AI(TEM)$ recently received FDA approval on xT CDx test which brings tumor DNA

Humanoids are coming and there are three public options now. We have Unitree (688836) now public at $53B. -> 2025 revenue ~$250M (humanoids were half of that). -> 2025 humanoid shipments ~ 5,500. -> ASPs ~$24k -> Production already demonstrated in the multi-thousand unit annual output range. Nomura has already initiated a 25x (yes...25x) 2027 P/S multiple for Unitree. We have $Churchill Capital Corp XI(CCXI)$ (Agility) expected to go public at $2.5B pre-money. -> Trailing annual revenue ~$37M -> Cash burn ~$100M -> 150 units but 65,000 real world hours -> ~$125k BOM with a path to $30k long0term -> $300M multi-year contracted order. And we have Optimus /

$NVIDIA(NVDA)$ vs $Cerebras Systems(CBRS)$ $NVIDIA(NVDA)$ is far cheaper but is it a better bet today? For now, yes. But let's see where $Cerebras Systems(CBRS)$ can get to when it's through this early IPO period. Firstly, $Cerebras Systems(CBRS)$ valuation was insane. It came in with an IPO that was 20x oversubscribed and then hit ~90x FY26 revenue estimates on day 1. With the stock now trading ~$52B on $900M in FY26 numbers...the valuation is still insane at ~57x 2026 forecasts. But we're also trad

Just want to add to this $Amazon.com(AMZN)$ math: E-Commerce: ARR is currently $547B. It's likely worth anywhere between 1 -1.8x sales which puts the e-comm valuation ~$800B. Digital Ads: FY26 revs is heading towards $80B+ with operating margins likely at +50% which puts operating income +$40B. I think a conservative multiple on this is in the 20x EBIT range which values the ads business ~$800B as well. That's a total of $1.6T in e-comm and ads. Overall valuation is $2.8T which means ~$1.2T is available for: -> AWS: Likely worth +$3T by 2030 -> Amazon Leo -> Zoox -> One Medical -> Prime Video -> Equity stakes

Just some back of the napkin maths on $Amazon.com(AMZN)$from MS: $Amazon.com(AMZN)$ AWS: AWS Q2 2026 revenue came in at $42.2B. That's an annualized run rate of $169B. $Amazon.com(AMZN)$ management have said it's very possible AWS reaches a $1T annual revenue business. Constraint is entirely capacity driven with a huge $496B backlog. But let's model out the path to $1T in annual revs: Capacity ~14 GW in 2025 -> 120 GW in 2035. -> ~6 GW in 2026 and then ~8 GW per year thereafter. -> Monetization rises to ~$12 revenue per incremental watt. -> Which means at $12/W, each 8 GW of new capacity adds $96B in ann

Here's my 2030 $NEBIUS(NBIS)$ estimate: -> 5GW capacity is contracted by YE 26 with 1GW capacity to be deployed per year from FY 27. -> Mid-term ACV is ~$20-25M as per management (for reference it used to be ~$12M). So a BULL case model would look something like this: 5.2 GW connected capacity by 2030 on a blended ARR/MW of $14M (that's a realistic haircut on utilization, mix, and timing vs the $20-25M forecasted. 5,200 MW * $14M = $72.8B ARR With EBITDA margins forecasted to stabilize near 45%... we're then looking at ~$32.7B in EBITDA. Bull case multiple: 25x Base case multiple: 18x Bear case multiple: 13x 25x * $32.7B -> $817B 18x * $32.7B -> $588B 13x * $32.7B -> $425B Or go more conservative (which is fair) and

Some of the more exciting themes right now: 1. Silicon Photonics Safest bet: $Lumentum(LITE)$ $Coherent(COHR)$ $Marvell Technology(MRVL)$ $Credo Technology Group Holding Ltd(CRDO)$ Moderate risk: $Applied Optoelectronics(AAOI)$ Highest risk: $POET Technologies Inc(POET)$ $Aeva Technologies Inc.(AEVA)$ 2. Solid Oxide Fuel Cells Safest bet:

Here's the potential 3x opportunity for $Aeva Technologies Inc.(AEVA)$: 1. OPTICS SEGMENT Management are guiding for "millions and millions of units" likely giving them "multiple hundreds of millions" per annum in revs. Conservatively, I think $300M by 2029 and $400M by 2030 is the range we're talking given H2 27 is the initial ramp up with a big production ramp up in 2028. 2. PERCEPTION BUSINESS Still pre-scale but looking at ~$522M in revs as per analysts in FY29. 3. TOTAL REVS Given the risks, I think it's sensible to work ~60-80% risking factors here. 60% * $822M = $493M in revs 80% * $822M = $657 in revs 4. SHARE DILUTION ESTIMATES Current share count is at 69.7M. SBC is unavoidable at this stage. 3-4M shares per annum vesting

95% of X is focused on semis, photonics, and memory it seems? What about the AI drug discovery theme? -> $AbCellera Biologics(ABCL)$ +55% in a week. -> $Tempus AI(TEM)$ still wildly undervalued, especially post $Microsoft(MSFT)$ partnership. -> $Eli Lilly(LLY)$ as the $1.1T (future -> $4T drug discovery leader). What about the critical minerals/materials/metals theme everyone on X went crazy about last year and then forgot about? -> $Global X Copper Miners ETF(COPX)$ is likely no longer