유리: AI의 마지막 병목 현상인가?

THE DAILY H.E.A.T. Torching Wall Street’s Obsolete Playbook. Delivered straight to your inbox 5 days a week. Subscribe Home Posts Glass: The Last AI Bottleneck? Glass: The Last AI Bottleneck? Apr 21, 2026 I’ve been a trader and investor for 44 years. I left Wall Street long ago—-once I understood that their obsolete advice is designed to profit them, not you. Today, my firm manages around $4 billion in ETFs, and I don’t answer to anybody. I tell the truth because trying to fool investors doesn’t help them, or me. In Daily H.E.A.T. , I show you how to Hedge against disaster, find your Edge, exploit Asymmetric opportunities, and ride major Themes before Wall Street catches on. Table of Contents H.E.A.T. News vs. Noise: What’s Moving Markets Today ETF News In Case You Missed It H.E.A.T. The Glass Gatekeeper SCHMID Group ($SHMD) and the Last Physical Bottleneck Standing Between Today’s AI and Tomorrow’s Infrastructure   TL;DR ▪  Thesis: Glass substrates are the next hard bottleneck in 1.6T-era AI interconnects — the layer the industry has been ignoring. ▪  Gatekeeper: Glass at commercial scale requires vertical, touchless wet processing. SCHMID is already shipping it into production facilities. ▪  Why now: Intel’s 2026–2030 glass roadmap plus first confirmed U.S. deliveries pull the thesis from “theoretical” to “operational.” ▪  Risks: Dilution, customer concentration, adoption timing, microcap volatility. Read the bear case before sizing a position. The copper is running out of road. For decades, the semiconductor industry answered every demand surge the same way: denser traces, faster clocks, more copper. It worked — until it didn’t. The problem stops being compute and starts being what compute sits on: the substrate. Compute is scaling. Packaging isn’t. And the gap between those two trajectories is where the next investment bottleneck lives. The “1.6T specification” — shorthand for the next generation of terabit-class AI server interconnect standards being pursued by Intel, Nvidia, and the hyperscalers — demands substrate performance that copper-filled organic materials simply cannot deliver. The physics are brutal and non-negotiable: thermal expansion mismatches, signal loss at micron-scale trace geometries, chemical contamination during fabrication. Glass clears the physical bar. Organic materials increasingly don’t. But glass only works if you can fabricate it at scale without contaminating the surface layers, without inducing the “puddling effect” from horizontal chemistry, and without physically contacting — and cracking — the panel during transport. That’s the hidden bottleneck. And it’s where SCHMID shows up. I’ve been mapping this transition since last year. The thesis is structural: each time a bottleneck migrates in a technology buildout, a narrow window opens where a handful of infrastructure suppliers become irreplaceable. We’ve already called the compute and memory legs of that trade. The substrate layer is the next migration — and one German engineering company nobody was talking about six months ago just proved it’s not a theory anymore. On a single trading day, SCHMID Group ($SHMD) surged over 30%. The catalyst: confirmed delivery of its InfinityLine H+ panel-level packaging platform into U.S. defense and aerospace manufacturing. That’s not a lab demo. That’s revenue, with verified deployment, loading onto a truck headed for a facility where reliability isn’t a preference — it’s a legal requirement. SECTION I The 1.6T Era: Why Glass Changes Everything The 1.6T specification represents the next generational step-change in AI server architecture — systems requiring terabit-class signal throughput where every micron of substrate geometry matters. The benchmark isn’t academic. Intel, Nvidia, and the major hyperscalers are building toward it now, and their packaging roadmaps are converging on one material transition: glass. TRADITIONAL PCB Organic resin expands unevenly under heat. Fine-line trace integrity degrades. Horizontal chemical processing leaves residue that ruins signal layers. THE PHYSICAL WALL At 1.6T trace widths, a single contamination event destroys panel yield. Any mechanical contact risks fracture. Horizontal chemistry creates the “puddling effect” — chemical pooling that kills fine-line uniformity. GLASS SUBSTRATES Dimensional stability across temperature swings. Lower signal loss. Higher-density interconnects. The material science is solved — the constraint is now manufacturing execution.   Intel has publicly committed to glass substrate adoption across its advanced packaging roadmap through 2026–2030. TSMC and Samsung are not far behind. But a material transition at this scale requires one thing almost nobody is pricing: the wet process equipment that actually fabricates the substrate. That’s the chokepoint. And SCHMID holds it. “Compute is scaling. Packaging isn’t. The market is about to pay for that gap.” SECTION II SCHMID’s Mechanical Moat: Three Advantages That Take Years to Replicate Competitive moats in semiconductor equipment don’t come from patents alone. They come from years of process refinement that lives in the institutional knowledge of a few hundred engineers. SCHMID has three advantages that are not quickly replicable — and the defense qualification signal confirms they’ve already cleared the highest bar in the industry: InfinityLine H+: Panel Scale That Few Platforms Can Match SCHMID’s InfinityLine H+ handles 700x700mm panels in production deployment — among the largest panel-level packaging (PLP) platforms commercially available. Larger panels compress unit costs per AI cluster board dramatically, accelerating the economics of glass adoption for Tier-1 customers. Panel-level packaging (PLP) refers to processing multiple chips simultaneously on a large-format substrate rather than individual package-by-package assembly — the approach required to make glass economics work at AI infrastructure volumes. Vertical Etching: Solving the “Puddling” Problem Horizontal processing traps chemistry on the substrate surface. SCHMID’s vertical architecture ensures uniform chemical distribution across the entire panel. For the micron-level trace widths required by SAP (Semi-Additive Process) and mSAP (modified SAP) technologies — the fabrication methods used to build the ultra-fine copper traces in advanced AI server boards — this isn’t a preference. It’s the only method that delivers acceptable yield at production volumes. Touchless Transport: Eliminating the Breakage Failure Mode Glass is brittle. In a high-volume fab, any mechanical contact with the panel surface is a yield event. SCHMID’s non-contact transport system eliminates this failure mode entirely. At Intel- or TSMC-grade production volumes, the yield improvement translates directly into margin at the OEM level — which is why customers qualify and rarely switch. The defense delivery matters for a reason that goes beyond the revenue line. Defense and aerospace customers don’t evaluate equipment — they qualify it. That process is measured in years, not quarters, and it involves reliability standards that commercial fabs never require. SCHMID has already cleared it. A competitor starting that qualification process today is 2–3 years behind on the procurement calendar. SECTION III The Financial Turnaround: From Debt Anchor to Asymmetric Setup Six months ago, the bear case on SCHMID was clean: great technology, broken balance sheet. Debt from the legacy business transformation was weighing on the equity, and the stock was priced as a distressed restructuring rather than a growth infrastructure play. That calculus has changed. €100M+ 2026 Revenue Guidance €95M 2025 Order Backlog >12% Target Adj. EBITDA Margin $27M Debt Converted to Equity   The January 2026 debt-to-equity swap — converting approximately $27 million of obligations held by XJ Harbour into shares — was a dilutive event, but the right one. A clean balance sheet re-rates the equity. With $30 million in convertible notes secured (first $15M deployed in January; second $15M tied to operational milestones), the company had capital to execute. Today’s U.S. delivery is the milestone. That second tranche is unlocking. The market cap at the time of writing hovers around $400M–$450M. The 2025 order backlog of approximately €95M (~$100M) represents a substantial portion of that figure — real contracted demand already on the books, before the glass substrate ramp has properly begun. A company delivering 30–40% revenue growth YoY, on a path to >12% EBITDA margins, in a substrate market that Intel is racing to scale — that is the asymmetric setup. Revenue recognition in complex equipment contracts follows delivery and installation milestones rather than order booking — so the €95M backlog doesn’t show up in one quarter. But it does provide a level of forward revenue visibility that is unusual for a company this size. Once a platform like the InfinityLine H+ is qualified and installed, switching costs are high: re-qualification timelines at defense or Tier-1 semiconductor fabs are measured in years, not months. The backlog isn’t just demand. It’s durability.  SECTION IV Winners, Losers & Pressure Points WINNERS TICKER COMPANY THESIS $SHMD SCHMID Group The gatekeeper. Among the first commercial-scale PLP wet process platforms for glass substrates at 700x700mm panel size. Defense-qualified. Backlog-backed revenue inflection in 2026. $LPKF / LPKFF LPKF Laser & Electronics Laser drilling partner in the glass core production chain. SCHMID’s vertical wet process + LPKF’s laser structuring = the complete glass substrate fabrication stack. Trades OTC as LPKFF for U.S. investors. $INTC Intel Corporation Most aggressive glass substrate adopter among Tier-1 chip manufacturers. A validated SCHMID supply relationship is a direct enabler of Intel’s advanced packaging roadmap through 2030. $NVDA Nvidia Corporation Blackwell Ultra and Rubin architectures require substrate precision achievable at yield only with vertical PLP systems. Indirect beneficiary as glass lowers per-unit packaging cost for AI compute.   PRESSURE POINTS PRESSURE POINT RISK VECTOR Traditional PCB Suppliers Companies whose revenue base depends on organic substrate volume for high-density AI applications face structural displacement as glass adoption scales past the 2026–2028 inflection. Horizontal Wet Process Incumbents Legacy horizontal chemistry platforms cannot solve the puddling problem at 1.6T trace geometries. Customers migrating to glass will re-qualify equipment. Incumbents without vertical PLP platforms are in a development race against an already-deployed competitor. Late-to-Qualify Competitors The defense delivery signals SCHMID has cleared multi-year qualification processes with U.S. government-adjacent customers. Competitors starting that process today are 2–3 years behind on the procurement calendar — a lead that compounds with each new customer qualification.   THE BEAR CASE Dilution risk is real. The January debt-to-equity swap cleaned the balance sheet but reduced per-share value. The second $15M convertible note tranche introduces further potential dilution if converted at depressed prices. Customer concentration. SCHMID’s revenue is not broadly diversified. A delay or cancellation at a single Tier-1 account would materially impact 2026 guidance. Revenue recognition timing. Complex equipment contracts recognize revenue on delivery and installation milestones, not order booking. A slip in installation schedules pushes revenue quarters forward. Glass substrate adoption pace. Intel’s public roadmap spans 2026–2030 — a five-year window. If adoption slows or a competing substrate material gains traction, the demand curve compresses. Foreign listing and currency exposure. $SHMD trades on a U.S. exchange but is a German-domiciled company. Euro/dollar currency dynamics, EU regulatory changes, and disclosure standards differ materially from U.S. domestic peers.   Five Things to Take Away From This Issue 1.  The bottleneck is migrating up the AI stack. GPUs get the headlines. Substrates decide whether 1.6T-class architectures can be manufactured at yield. The market hasn’t priced that migration yet. 2.  Glass is the only material that clears the physical bar for next-generation AI server interconnects — but the material isn’t the constraint anymore. Manufacturing execution is. That’s where the investment lives. 3.  SCHMID isn’t “a bet on glass.” It’s a bet that vertical, touchless wet processing becomes the gating toolset for glass substrates at panel scale — and that the company already shipping it into production has a durable head start. 4.  The defense delivery is a qualification milestone, not just a sales event. It is a signal about reliability expectations and yield performance — and a proof point that SCHMID’s technology has cleared the hardest customer standard in existence. 5.  The setup is attractive if execution holds: a ~€95M backlog (~$100M) representing a meaningful portion of current market cap, growing revenues, and a path toward >12% EBITDA margins in a market Intel is actively scaling. This is still a microcap. Dilution, customer concentration, and adoption timing are real risks and must be monitored continuously.  The glass revolution isn’t coming. It’s loading onto a truck. We’ll be watching the second convertible tranche, the Q1 revenue confirmation, and any incremental announcements on the Intel production relationship. When the next catalyst hits, subscribers who did the work will already be positioned. Space Exposure Designed For Potential Weekly Payouts The space economy is no longer just a government program — it’s becoming part of a global industry that is gradually advancing. But why view long-term growth as the only potential path forward?  SPCI seeks to offer approximately 100% concentrated exposure to the space industry while striving to generate weekly income through a disciplined put credit spread options strategy.  Gain exposure to the “final frontier” and a potential weekly distribution while you wait for the theme to play out. Discover SPCI. Distributor: Foreside Fund Services | Investing involves risk including possible loss of principle. News vs. Noise: What’s Moving Markets Today Not shockingly last week’s truce didn’t hold and oil spiked yesterday, but the market didn’t really budge. Most likely traders assuming we would get a TACO Tuesday from Trump. Futures are up, but the ceasefire ends on Wednesday and Trump has said he is unlikely to extend it. THE SIGNAL While everyone's watching oil and ceasefire deadlines, the real story is quietly getting stronger. Japanese semiconductor equipment purchases just hit their highest absolute level ever in March, up 11% Y/Y and accelerating — that's your capex supercycle confirmation in real time. MRVL is working with Google on two additional custom chips. Credo just acquired DustPhotonics and guided to >$500M in optical revenues in FY27 — roughly 2x ahead of what the Street had modeled. CRDO is targeting 75% revenue growth next year versus the Street's 53%. CLS remains the highest-conviction EMS name with Google TPU continuity, five sole-sourced 1.6T networking programs ramping in the back half, and an Anthropic rack-scale program entering production soon. This is the AI infrastructure buildout continuing to compound in broad daylight while investors trade Iran headlines. The dumb story of the day is the ceasefire tape bomb. The smart trade is still the same one it was last week. Software also continues to bounce…. I’m showing you an ETF here as an example, if I were to buy software for any more than a trade you have to be extremely careful as some companies will benefit from AI, others will get taken out. What Iran Tells Us About UFO Disclosure When governments confront unknown threats in their airspace, defense budgets surge and the same aerospace and surveillance companies move hardest. On March 2nd, Northrop jumped 6% and Lockheed 3.3% on the Iran news — and President Trump has since ordered the formal release of government UAP files, with the Pentagon confirming compliance. So if a conventional conflict can move these stocks this fast, what happens when the bigger story breaks? See the UFOD holdings: [thetruthisoutthereufod.com ETF News A Stock I’m Watching $AVEX ( ▲ 7.72% ) This defense stock just IPO’d and it’s up big. Also up 10% pre market. People are comparing them to Anduril so pretty rarified air. In Case You Missed It The H.E.A.T. (Hedge, Edge, Asymmetry and Theme) Formula is designed to empower investors to spot opportunities, think independently, make smarter (often contrarian) moves, and build real wealth. The views and opinions expressed herein are those of the Chief Executive Officer and Portfolio Manager for Tuttle Capital Management (TCM) and are subject to change without notice. The data and information provided is derived from sources deemed to be reliable but we cannot guarantee its accuracy. Investing in securities is subject to risk including the possible loss of principal. Trade notifications are for informational purposes only. TCM offers fully transparent ETFs and provides trade information for all actively managed ETFs. TCM's statements are not an endorsement of any company or a recommendation to buy, sell or hold any security. Trade notification files are not provided until full trade execution at the end of a trading day. The time stamp of the email is the time of file upload and not necessarily the exact time of the trades. TCM is not a commodity trading advisor and content provided regarding commodity interests is for informational purposes only and should not be construed as a recommendation. Investment recommendations for any securities or product may be made only after a comprehensive suitability review of the investor’s financial situation.© 2026 Tuttle Capital Management, LLC (TCM). TCM is a SEC-Registered Investment Adviser. All rights reserved. Keep Reading Apr 9, 2026 • 3 min read The Industrial Space Age: Bezos just filed for 51,600 satellites Apr 1, 2026 • 14 min read THE $7 TRILLION AI CONSTRUCTION BOOM Apr 10, 2026 • 17 min read If Your Heart Is Beating, We Will Find You View more