As micro watcher pointed out, look at what the market happily pays for hyped small-caps. $ALMU trades around 80x sales. $AMPX many times higher than this name. Meanwhile $AMPG sits at ~6x trailing and ~4x forward. So let me tell you the complete story of what you're actually getting at that price. 📡 THE RADIO. AMPG is the only American company to design and commercialize a 64T64R Massive MIMO O-RAN radio, the only one of its configuration certified by the DoD-backed Open6G OTIC, and the only 64T64R vendor at the latest O-RAN PlugFest. 🇨🇦 TELUS. ALREADY PRINTING. Per Telus's own VP: every Open RAN site runs AmpliTech antennas, 2 of the 5 radios per sector, live, alongside Samsung, on a Tier-1 carrier. Only ~15% of sites are Open RAN today. Starting line, not ceiling. 🤖 NVIDIA. AMPG's radio ran hand-in-hand with NVIDIA's Aerial in the world's first open-source AI-RAN demo, inside the DoD-funded Open6G hub. NVIDIA is now on its customer wall. (A collaboration/demo, not a signed supply deal, but the direction is loud.) 🛰️ AMAZON / KUIPER. Dec 2024: AMPG shipped space-qualified amplifiers to a "Fortune 50 satellite provider," LEO constellation, tens of thousands of units expected. The only Fortune 50 building its own LEO network is Amazon. Now on the customer wall too. (Still a deduction, still a prototype, but the dots keep landing.) ⚛️ QUANTUM. The only US maker of the cryogenic amplifiers superconducting qubits need, Google and IBM named as proof-of-concept recipients. (Optionality, not revenue yet, but real.) 🏛️ THE CUSTOMER WALL. This isn't a no-name. Straight from amplitechgroup.com, the logos AMPG puts under "Customers": NVIDIA, IBM, Amazon, Boeing, Lockheed Martin, Northrop Grumman, L3Harris, CPI, Fujitsu, HTC, Globecomm, C2Tech, Greins, Northeastern, Georgia Tech, University of Edinburgh, Digital Catapult, Paramount, DiscoveryPlus, Disney Channel, and more. (Honest framing: it's the company's own wall, spanning all divisions and years, including its distribution arm, so it shows who they've worked with, not the size of each deal. But you don't put Lockheed, NVIDIA and Amazon up on a whim as a public company.) 💰 THE FUNDAMENTALS. Debt-free. Gross margins 33% → 48%. $118M in LOIs across multiple carriers. A founder who started with $2 in his pocket, still owns ~10% , and hasn't sold a share. 🇺🇸 NOW THE BIG ONE. WHY THIS IS NATIONAL SECURITY? Look at who builds the world's wireless radios: Nokia (Finnish). Ericsson (Swedish). Huawei and ZTE (Chinese — banned in the US). Read that again. The backbone of America's communications, the same networks its military, its first responders, its critical infrastructure run on, is supplied almost entirely by foreign companies. That's not a convenience problem. It's a strategic vulnerability. Whoever builds the radios sits inside the nation's nervous system, which is exactly why the US ripped Huawei out of its networks over espionage and backdoor fears. You cannot have a trusted network on untrusted, foreign-controlled hardware. So the US government's answer is Open RAN, an open, multi-vendor architecture that lets American companies supply trusted, individual pieces (a radio here, software there) instead of depending on a single foreign empire. And the tip of that spear is Open6G: the DoD-funded (OUSD R&E, via an Army Research Laboratory agreement) research hub at Northeastern, built to create a domestic, secure foundation. That's the whole point. Washington is funding an entire architecture to get foreign hardware out of the nation's networks, and at the radio layer, where it matters most, there is exactly one American company at that spec. Already defense-qualified (Lockheed, Northrop, L3Harris, Boeing). Already inside the DoD-backed hub. Already certified. Already deployed at a Tier-1 carrier. When the government decides trusted, domestic, open infrastructure is a national priority, and there's only one US company that makes the radio, that company stops being "a micro-cap" and starts being strategic. So step all the way back and look at the asymmetry: The market pays 40–80x sales for hyped names with tiny revenue and no story like this. AMPG, real revenue, real customers, a debt-free balance sheet, NVIDIA and Amazon on its wall, designed into Telus, and sitting at the center of a national-security strategy, trades at single-digit sales. That's the story. That's the gap. And the market is only starting to close it. Not financial advice. I'm long $AMPG. DYOR. 📡

We need to think harder before selling. Spend as much time thinking about an exit as we would with buying. Don't sell just because something is painful TODAY.

$FCEL will be off to the races.

$BRUN has real revenues backed by a massive and growing backlog Boost Run Investor update here: boostrun.com/investors/

Started a challenge account. Full ported $XFAB Can we full port our way to 1m from 30k? 2x incoming 🚀

$XFAB - 10 REASONS why INVEST 1. High-Growth Megatrends Focus on Key Markets No Consumer Risks Stable Long-Term Growth 2. Strong Customer Loyalty Unique Manufacturing Processes Sole-Source Supplier High Switching Costs 3. Electric Vehicle (EV) Boom More Chips Needed Automotive Certified Critical Component Supplier 4. Next-Gen Power Materials Silicon Carbide Pioneer Gallium Nitride Expertise Green Energy Leader 5. Secure Financial Agreements Long-Term Contracts Take-or-Pay Clauses Pre-Funded Expansion 6. Advanced Medical Technology Medical Chip Demand Lab-on-a-Chip Solutions Life-Saving Electronics 7. Safe Global Footprint Six Global Factories No Geopolitical Trap Local Supply Chains 8. Long Product Lifecycles Decades of Revenue Low Upgrade Costs Reliable Cash Flows 9. Leadership in MEMS Sensors Micro-System Pioneers Heterogeneous Integration Smart Device Futures 10. Direct Research Connections Lab-in-Fab Concept Fast Market Entry Cost-Efficient Innovation Not financial advice.

Time in the market is teaching us an important lesson. Don't let price action change your thesis.

The stock market is such a gift. There are multiple companies that we can full port and go to sleep. Wake up richer on the other side.

$HPE just CRUSHED IT. Biggest EPS beat since 2018. Stock surging 30% after hours. Q2 FY2026 results: - Revenue: $10.68B vs $9.79B est - Adj EPS: $0.79 vs $0.53 est - Server Revenue: $5.45B vs $4.66B est - Free Cash Flow: $915M (record Q2) - AI Systems Orders: $1.8B (record backlog)

We at Frontrun Capital are not attached to a stock. We are attached to fundamentals. $SIVE doesn't have the fundamentals to command that price. $AAOI and $IQE do. Long here. Exit there ☝🏽

Listened to the $SIVE earnings call today. I have a 7% position in $SIVE When the CEO talks about the 'opportunity pipeline', I can read between the lines. I was also a CEO/founder. It's usually a sum of the 'potential value' of 'customer engagements' from the CRM app the company uses. Note that these aren't LOIs or contracts, and sales people usually exaggerate these numbers to 'look good' infront of management. Venture building has taught me a lot, and knowing how organizations are built tells you what might be happening under the hood. $SIVE is still very early. If you have to be bullish, then be so with knowledge and a dose of skepticism.

The more I pick stocks/companies, the more I learn that I am price sensitive. (1) I don't like overpriced companies no matter how rosy the future is (2) I love a re-rate story. Being able to front-run the price discovery ahead of everyone else = satisfaction (3) Commercial inflection points backed by patents/IP = treasure hunting Know thyself.

Probable $DELL $IQE connection? $IQE makes InP epiwafers ▼ MACOM uses epiwafers to make laser drivers, TIAs, EMLs ▼ Transceiver OEMs $LITE $AAOI use MACOM components to build 800G/1.6T modules ▼ Transceiver modules installed by $TSSI in server racks ▼ $DELL buys those specific transceivers which are made from $IQE InP epiwafers ▼ Dell PowerEdge AI server ships to customer like $CRWV

Rosenblatt industry update on Indium Phosphide Supply and Demand. They estimate 12x supply increse 2025-2030. NVIDIA have asked for 20x supply increse over the same period. Supply constrained #AAOI $NVDA $LITE

Wild idea I was discussing with @Ren_aramb The idea is to create a sub-portfolio of 30-50 microcaps and then double down on winners.. could turn 50k into 1m... Margin is a bad idea on this one but otherwise not too different from how VCs do pre-seed investments 😅

Fantastic write up on $IQE vs $SOI by @AlmaCap114204 I had already trimmed our position in $SOI as we feel the price reflects the business of tomorrow I don't have clarity on TAM, so unless the conviction develops - I will sit it out. I do like the monopoly but not at this valuation. I continued to hold $IQE through the ER today. It's a much better company today than it was 3 months ago.

$IQE FY2025: headline revenue -17.6% hides the real story. Photonics 15% to £57.1M which is now the majority segment MACOM just bought 11.5% of the company and locked in long-term supply agreements. A customer doesn't invest £45M in their supplier unless they see accelerating demand ahead. "Multiple Tier 1 InP design wins for data centre" volume production. FY2026 guided 20% growth. Bank debt eliminated. Balance sheet clean. FY2026: exceeding 20% growth guided The smartphone drag is fading. The AI photonics inflection is here. Every optical transceiver connecting GPUs needs InP wafers. IQE makes them.

Why HLIT's TAM is the 500M global homes, not 85M US: Every cable modem whether in Ohio, Germany, Brazil, or Japan needs DOCSIS technology to function. When cable operators upgrade to DOCSIS 4.0, the headend software (Harmonic cOS) must be upgraded regardless of geography. HLIT's 41M modems on cOS today is only 8% of the 500M global installed base. The other 92% is the multi-year runway. $HLIT

$HLIT is the lifeline for cable operators in the US - powering 85m homes.