Durable Moat

“The best business to own is one that over an extended period can employ large amounts of incremental capital at very high rates of return.” - Warren Buffett

Coda Octopus ( $CODA ) controls a unique asset - the only true real-time 3D sonar platform - with demand from the U.S. Navy and allied defense customers, supported by multi-year contracts and 66% gross margins. The business is debt-free, cash generative, and has only just recovered to pre-COVID revenue levels, suggesting an inflection point. A new high-growth product (DAVD) is scaling rapidly (200%+ YoY), while the broader NATO-driven expansion in unmanned underwater vehicles provides a long runway for demand. Despite this, the stock trades at a meaningful discount to defense peers on EV/Sales, creating an attractive entry for a multi-year compounding opportunity.

Bloodbath in AI stocks today yet $SIVE up 8% 😅

A lesson from $POET: Don’t invest in a company whose CEO only owns 0.0083% of the float There are people on X who own more than that

Top 4 highest-conviction asymmetric bets in the global drone/C-UAS/underwater drone space right now (in the post-Iran budget surge). 1. Red Cat Holdings ( $RCAT ) RCAT stands out as the top asymmetric bet because it delivers explosive, multi-domain growth (small tactical UAS + USVs) with proven U.S. Army Program of Record momentum and NATO wins. 2025 revenue surged 161% to $40.7M (Q4 alone +1,985% YoY), production capacity expanded 520%, and it holds a strong $168M cash position heading into 2026 with analysts eyeing 250%+ revenue growth. Black Widow orders, Blue Ops USV ramp (500–1,000 units/year potential), and recent Apium swarm AI acquisition position it perfectly for mass attritable offense and autonomy — the exact doctrinal shift driving the biggest Pentagon and ally spends. High-beta disruptor with real 2026 inflection. 2. DroneShield ( ASX:DRO / OTC: $DRSHF) DroneShield is the purest listed C-UAS play and a standout asymmetric winner in the $20B+ global counter-drone wave. FY2025 revenue exploded 276% to $216.5M with strong profitability, and Q1 2026 already delivered $74M revenue (+121% YoY) pushing committed FY2026 revenue to ~$155M (with $59M added since January). Battlefield-proven soft-kill systems (DroneGun, Sentry) are seeing repeat Western military, event (FIFA World Cup), and infrastructure orders, plus Australian policy tailwinds. As the only pure-play public name in a surging defensive category, it offers the cleanest torque to any new swarm headline or layered defense spend. 3.Ondas Holdings ( $ONDS ) ONDS combines autonomous UAS architecture with integrated C-UAS (Iron Drone, Sentrycs) and has delivered outsized contract momentum that justifies its high-beta profile. It raised 2026 revenue guidance to at least $375M (massive step-up), backed by multi-million World Cup security wins, Israeli border demining programs ($10M+ initial with $50M+ follow-on potential), and heavy engineering orders. Layered cyber-RF + kinetic solutions plus Palantir ties give it broad offense-to-defense exposure across military, border, and critical infrastructure — exactly where post-Iran demand is accelerating fastest. 4. Kraken Robotics ( $KRKNF / TSX:KRK) Kraken is the highest-conviction “picks-and-shovels” asymmetric play in the exploding underwater drone domain. It announced ~$24M in new defense orders (March 2026) plus another ~$28M in SeaPower batteries and Kraken SAS systems (April), hitting customers across five countries with growing backlog. As the critical enabler for long-endurance UUV/AUV batteries and advanced sonar in naval/Indo-Pacific programs, it rides the undersea autonomy surge with far less competition than aerial names. 2026 guidance of $165–175M revenue makes it a leveraged bet on the next frontier of unmanned warfare. These four give concentrated exposure across offense, pure defense, integrated systems, and maritime - with the highest torque to continued 2026 contract flow. DYOD!

The Quantum Narrative. Quick reality check : $IONQ: Best of the public bunch on paper—revenue growing, big cash pile from acquisitions, #AQ64 claims, high gate fidelity records. Quantinuum (private) has consistently shown stronger, more credible trapped-ion demos (logical qubits, error-corrected circuits, fidelity benchmarks) in peer-reviewed work. IONQ’s marketing has been… aggressive (handwriting recognition on quantum computers? testifying to Congress that quantum “doesn’t hallucinate” because it’s deterministic?). Recent public spats with Quantinuum over benchmarks show the narrative war is alive. Rigetti ( $RGTI ): Superconducting qubits. Respectable tech, but they’re a distant third/fourth behind IBM/Google in that modality. Cloud access is nice, but no one serious thinks they’re leading hardware. D-Wave ( $QBTS ): Annealing pioneer. They have real customers and hybrid solver usage is up, but annealing is not universal gate-model QC. Their “advantage” claims have been debated for 15+ years (classical solvers often match or beat them on practical problems). Now they’re pivoting to gate-model too—fine, but it underscores they weren’t there yet. Quantum Computing Inc.( $QUBT ): The clearest example of what experts like Aaronson is warning about. Photonics/room-temp sounds sexy on a slide deck. Reality: microscopic revenue, enormous valuation multiples, dilution, and multiple securities lawsuits alleging misleading statements.

Most investors are always chasing the last wave. The AI infrastructure cycle is playing out in layers — and each one is bigger than the last. Here’s the full roadmap: Wave 1: Semiconductors ✅ (Priced In) $NVDA. $AMD. $AVGO. The GPU arms race. Everyone knows this trade. Raw compute was the foundation. Most of the upside? Already captured. Wave 2: Memory & Storage ✅ (Confirming) More data. More inference. More throughput. Flash, DRAM, HBM all followed. $SNDK ’s recent move is the latest confirmation. The storage bottleneck is real — and the market is finally catching up. Wave 3: Photonics & Optical Networking 👀 (In Motion NOW) This is what most retail investors are still sleeping on. You cannot move AI workloads at scale with copper wire. Data centers need ultra-fast, low-latency interconnects — and photonics is the only answer. Silicon photonics is actively replacing copper-based communication inside and between hyperscaler campuses. The plays institutions are quietly accumulating: $AAOI — pure-play optical components, deep data center exposure $COHR — vertically integrated photonics at scale $LITE — transceivers powering the hyperscaler build-out $POET — integrated optical engines, early-stage, asymmetric upside $CIEN — backbone optical networking infrastructure $FN — precision optical manufacturing, best-in-class margins This wave is not coming. It is here. Wave 4: Power & Energy Infrastructure (Early Stage) AI doesn’t run on ambition. It runs on electricity. Data centers are projected to consume 8–10% of U.S. power by 2030. That is not a tailwind. That is a structural demand shock the grid is not prepared for. Nuclear is the only baseload solution that can sit next to a hyperscaler campus and deliver consistent, carbon-light power at scale. The plays: $CEG & $VST — utilities with nuclear-powered AI offtake contracts $CCJ — uranium supply, the fuel behind the revival $OKLO & $SMR — next-gen small modular reactors $VRT — liquid cooling and power management inside data centers The capex commitments from $MSFT, $AMZN , and $GOOG are already locked in. The power infrastructure just needs to catch up. This wave is early. That is the opportunity. Wave 5: Robotics 🤖 (Loading…) Once AI is trained, powered, and connected — it needs a body. Physical deployment into warehouses, factories, logistics, and healthcare. This is the final and potentially the largest layer of the entire cycle. Hardware-software integration is still 12–24 months from full ignition. But the smart money is already laying the groundwork quietly. The pattern is simple: Each layer enables the one after it. → Semis built the brain → Memory gave it recall → Photonics gave it nerves → Power gives it fuel → Robotics gives it hands The investors winning this decade are not chasing yesterday’s leader. They are identifying tomorrow’s bottleneck — and positioning before the crowd arrives. The question is not which wave already ran. The question is: which wave are YOU in front of? $AAOI $COHR $LITE $POET $CIEN $FN $CEG $VST $OKLO $SMR $CCJ $VRT DOYR , bookmark this and retweet for others.

Super bullish on $MRLN. But there’s a 20-day window before Sep 16, 2026 that decides how much dilution shareholders get. Stock price during that window sets conversion: • ≥ $12 → lower dilution • $7.5–12 → moderate • < $7.5 → maximum dilution This period matters a lot.

Anders Storm’s repeated, specific praise is a strong and credible counter to blanket “pumper” accusations against Serenity (@aleabitoreddit). @stormdirac Storm is not a random retail account: He was CEO of Sivers for 8–10 years (built the photonics side). He knows the tech, partnerships (Ayar Labs, Jabil, etc.), and execution realities better than almost anyone outside the current management. He’s not hyping blindly - he engages with Serenity’s threads on supply-chain mapping and calls them “great deep research.”

Thanks, nice analysis. A big chunk of the current revenue run-rate is tied to a conflict the U.S. (across administrations) has signaled it wants to wind down. If aid scales back or the war de-escalates meaningfully, the revenue cliff is real. They’re not “entirely” dependent (U.S. Gov direct was ~78% of 2025 revenue overall, with other customers and international pursuits), but the Ukraine piece is outsized enough to matter. In summary: Ukraine-concentrated with PE overhang and margin pressure.

Concerned for the many retail investors who bought $AVEX up nearly 100% in a matter of days without fully understanding the risks. Reasons I would not own this stock: 1. The vast majority of revenue is tied to programs supporting Ukraine. When this conflict ends or the U.S. scales back support, revenue faces significant and immediate risk. 2. High private equity ownership. With a lockup expiring in October, the historical PE playbook is to sell systematically into retail enthusiasm. We’ve seen this happen many times. 3. Gross margins were 21.8% in 2025, down from 28.1% the year prior. For a business scaling rapidly in a single conflict, this number is moving in the wrong direction. 4. IPO proceeds were used primarily to restructure debt, not fund growth. $AVEX issued no formal forward guidance in their prospectus. For a stock trading at nearly double its IPO price, that is a significant lack of visibility. 5. The entire product portfolio is concentrated on offensive strike missions. No counter-UAS solutions, no ground robotics, no missile defense, no stratospheric ISR, no $PLTR integration. The FY2027 defense budget allocates $20.6B to counter-UAS alone, up 424% YoY. $AVEX has no meaningful presence in any of the fastest-growing segments. $AVEX is a real business with a strong backlog, proven deliveries, and impressive IP, but the risk/reward is not there for me personally. $AVEX will likely benefit to some degree from the 2027 defense budget, but right now, they are a wartime contractor with high PE ownership, declining margins, negative FCF, and extreme revenue/customer concentration tied to a single conflict that the U.S. wants to end. I wish everyone well with their investments!

$AVEX A big chunk of the current revenue run-rate is tied to a conflict the U.S. (across administrations) has signaled it wants to wind down. If aid scales back or the war de-escalates meaningfully, the revenue cliff is real. They’re not “entirely” dependent (U.S. Gov direct was ~78% of 2025 revenue overall, with other customers and international pursuits), but the Ukraine piece is outsized enough to matter. In summary: Ukraine-concentrated with PE overhang and margin pressure.

This is insane. The $1.5T U.S. defense budget lays out $75B for drones. The combined market cap of U.S. pure play drone primes like $AVEX $AVAV $KRATOS is $27B.

The AI boom started as a massive GPU boom—and it still is for the heaviest parts—but it's evolving. Most headlines scream "Nvidia GPUs everywhere" because training frontier models and high-throughput inference on big LLMs are insanely GPU-dependent. But the full picture (especially in 2026) shows CPUs handling a growing share, particularly as workloads shift toward inference, agentic AI, RAG, and orchestration. This doesn't kill the GPU story—it just means the overall compute stack needs way more balanced CPU:GPU ratios. Picks? ARM has the most torque (smallest base + direct silicon upside) but also highest execution risk on the new chip. TSM is "safer" asymmetric — proven moat, but bigger market cap caps the % upside vs ARM. AMD is the hedge play if Arm doesn't take 100% share.

$NOW The share price has been obliterated, but before calling it a screaming buy, the core issue might be a potential structural shift in how customers use the platform. ServiceNow's traditional business is heavily seat-based (or user-based licensing). Revenue grows when companies add more employees who need access to IT, HR, security, or customer service workflows. AI tools like Now Assist (and future AI agents) let companies automate more tasks. Customers can handle the same (or growing) workload with fewer people — what management itself calls "headcount-neutral" growth. CEO and executives have openly highlighted examples: They’ve automated ~90% of certain customer service use cases, allowing clients to grow revenue and free cash flow without adding headcount. This is great for customer efficiency and ServiceNow’s margins (they’re guiding for strong operating and FCF margins in 2026), but it can slow the top-line seat expansion that has driven past hyper-growth.On top of that:UBS downgraded to Neutral and slashed its price target, citing risks that autonomous AI agents could disrupt traditional workflow automation (including parts of ServiceNow’s customer service business, ~10% of revenue). BTIG cut its target to $185, explicitly citing "limited organic upside" in the FY2026 guidance. They see the current ~19.5–20% subscription revenue growth outlook as relying more on acquisitions (e.g., Moveworks contribution) and AI upsell than pure base business momentum. They’re also skeptical that growth stays that strong into 2027–2028. Broader sector rotation: Many software names got hit on fears that “AI will eat software” — with concerns that generative/AI agents reduce the need for complex enterprise suites.

Yes, AI infrastructure stocks are booming, but a GPU glut is bound to happen in the future. Too much supply of high-end AI GPUs relative to demand, which would push down rental prices, reduce utilization, and hurt neocloud profits (like Nebius and CoreWeave). In the short term (next 1–3 years), the risk of this happening is very low. The market is still in strong shortage mode: GPU rental prices are rising (H100s up ~20–40%, Blackwell up ~22%), systems are sold out, and demand is “off the charts.” NVIDIA itself is guiding to roughly $1 trillion in cumulative demand for Blackwell and Rubin platforms through 2027. Meanwhile, real bottlenecks are not chips but power, data centers, and memory. Both Nebius and CoreWeave report high utilization and massive backlogs (e.g., CoreWeave ~$67B), supported by hyperscalers like Meta and Microsoft continuing aggressive AI spending. Bottom line: this period strongly supports growth, high pricing, and near-full capacity. The first real risk window appears in the 2028–2032 period, where a glut becomes moderately possible. By then, supply will have scaled significantly (new GPU generations and large data center buildouts), so the key variable becomes AI monetization (ROI). If companies generate strong returns from the hundreds of billions being invested, demand can absorb supply. But if ROI disappoints or spending slows, the industry could enter a classic overbuild cycle, leading to excess capacity, especially in older GPUs. History suggests at least one correction is likely in this phase. For Nebius and CoreWeave, execution matters: long-term contracts, prepayments, and differentiation (like Nebius’ software stack) provide some protection, but CoreWeave’s higher debt load could make it more sensitive in a downturn. Over the 10+ year horizon (2035+), outcomes depend on whether AI delivers sustained economic value at scale. If AI continues to expand like the internet — driving productivity, new industries, and real revenue — demand could stay strong enough to absorb supply for a decade or more. However, most infrastructure booms eventually face cycles of overbuilding and declining returns, so periodic gluts are likely even in a bullish long-term scenario. Overall takeaway for investors: short term is clearly favorable, medium term (late 2020s) is the key danger zone, and long term will likely be cyclical. These are high-risk, high-reward businesses — watch GPU pricing, utilization rates, hyperscaler capex, and any signs of idle (“dark”) capacity closely.

"For the past 100 years, aviation also has been built around pilots. Its next 100 years will be built around autonomy." - Matt George, Founder and CEO Merlin, Inc. (MRLN) Merlin’s thesis is that it’s building the “operating system for aircraft” — not just autopilot, but a full AI pilot that can run missions end-to-end. The key insight is that aviation autonomy is far easier than self-driving because the environment is controlled, structured, and predictable. So while the market thinks “autonomy = impossible like Tesla Full Self Driving,” Merlin is solving a much more tractable problem and could become the dominant platform for autonomous flight. The business model is powerful because it retrofits existing aircraft instead of building new ones. Each plane generates about $3M upfront and $2M recurring annually, creating a high-margin, software-like revenue stream layered on top of a massive installed base. With 800+ aircraft already under contract and a total market of 13,000+, this can scale into billions of recurring revenue. The combination of cost savings, pilot shortages, and defense demand makes adoption economically compelling and potentially sticky. The opportunity exists because the market is mispricing the company due to SPAC stigma and complexity. While public investors ignore it, insiders and sophisticated investors have doubled down, and ownership is highly concentrated. The thesis is that this is not a speculative SPAC but a misunderstood, early-stage platform with asymmetric upside — where even modest execution could lead to large returns, and full success could be massive. $MRLN

Michael Burry just called the bottom on software stocks He opened a new 3.5% position in PayPal $PYPL at $49.38 yesterday And is adding Salesforce $CRM and MSCI $MSCI this morning, per his Substack His recent purchases: • PayPal $PYPL • Fiserv $FISV • Adobe $ADBE • Autodesk $ADSK • Veeva $VEEV

EOS.AX (Electro Optic Systems) looks like one of the better long-term buys in the exploding counter-drone space right now. With cheap drone swarms becoming a nightmare in Ukraine, the Middle East, and pretty much every modern conflict, militaries are scrambling for affordable ways to knock them down instead of burning through million-dollar missiles on $5,000 drones. That’s exactly where EOS shines. Their lineup mixes solid kinetic options like the Slinger remote weapon system with high-energy lasers (HEL) that can take out threats at almost zero cost per shot. The recent €71 million deal for a 100kW laser to the Netherlands was a big one — basically the first major export of its kind — and they’ve stacked up solid US contracts plus a huge Middle East Slinger order. The backlog is sitting around A$459 million, which gives them real visibility heading into 2026 and 2027. What I like most is that they’re not just another hype story. They’ve got actual in-house optics and laser expertise, they’re building production capability in Singapore for Asia, and they’re staying flexible on exports (ITAR-free helps a lot with allies in Europe and the Middle East). As drone warfare keeps evolving, the shift toward directed-energy “digital steel” defenses feels inevitable, and EOS is one of the few companies with proven hardware already winning real contracts. It’s not without risks — defense deals can slip and execution always matters — but if you’re looking for real exposure to the anti-drone future, EOS.AX feels like one of the more grounded plays in the sector.

If AI capex ROI disappoints, or hyperscaler spending slows, multiples can compress quickly (20–40% drops are common in tech). Geopolitics, energy constraints, or competition add layers. Valuations assume continued 30–60%+ growth for leaders like NVDA—any slowdown could trigger a reset. • History lesson: The internet wasn’t a bubble (it transformed the world), but valuations were. AI infrastructure demand is real and secular, but we’re in a maturation phase where multiples matter more.