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 $5 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.
The pluggable era isn’t ending — it’s being segmented. Where it matters most, CPO moves the profit pool upstream: packaging, test, optical coupling, laser supply, and platform control. Here’s who collects the toll.
1.6 Tb/s Architecture decision point: CPO vs. XPO/LPO/NPO | 70% Optical interconnect power improvement — Broadcom Bailly (disclosed) | $5.5B+ CPO/NPO module opportunity by 2027 in some forecasts; definitions vary | XPO / LPO / NPO Extend pluggables where serviceability matters; CPO wins highest-density segments |
On a Tuesday morning in June 2026, a room full of optical engineers sat down to reason through why Co-Packaged Optics is hard. Ninety minutes later, the consensus was three words: CPO is really not easy.
That is not a reason to ignore it. That is the entire investment thesis.
Every technology transition that created generational wealth had the same property: the underlying engineering was genuinely difficult, and the window between ‘this is solvable’ and ‘this has been solved’ was short. The people who made money were not the ones who waited for certainty. They were the ones who correctly mapped which problems remained open — and which companies owned the tools to close them.
The pluggable optics market is not disappearing. XPO, LPO, LRO, and NPO architectures can extend the front-panel model further than the simple ‘copper is dead’ narrative suggests. Arista’s XPO, announced in March 2026, is explicitly designed to push density and bandwidth well beyond current pluggable ceilings using liquid-cooled, high-density pluggable modules. That is a real counterforce.
But at 1.6T per port and beyond, the economics change. Power, signal integrity, cooling, and density start fighting each other. In the highest-density AI fabrics, the optic can no longer be designed as a detachable afterthought. It becomes part of the package, the test flow, and the platform architecture. That is where the profit pool migrates: packaging, wafer-level test, photonic foundry capacity, external laser supply, and ASIC platform control.
CPO is a bet that the highest-value part of optical networking moves from detachable modules into package-level integration. Here is our map of who wins that transition.
WHAT IS CPO — AND WHY DOES IT MATTER NOW Co-Packaged Optics (CPO) integrates optical transceivers directly onto the switch ASIC package, eliminating the electrical interconnect loss that makes pluggable optics unsustainable at 1.6Tb/s+ per-port bandwidth. Broadcom’s Bailly platform, the first confirmed production CPO Ethernet switch, delivers a 70% improvement in optical interconnect power efficiency versus a comparable pluggable configuration. The physics advantage compounds at scale. |
Part I: The Segmentation, Not the Extinction
The datacenter optical market has operated for two decades on a simple premise: build the switch, build the optic, connect them at the faceplate. Pluggable transceivers were the right answer for the bandwidth regimes of 10G, 40G, and 100G. The electrical trace from ASIC to faceplate was short enough that signal integrity remained manageable and power overhead was acceptable.
At 1.6Tb/s, that premise does not break all at once — but it breaks where it matters most. The electrical loss across the SerDes interface at those speeds consumes a disproportionate share of the link power budget before the signal even reaches the optic. The re-timers and retransmitters required to compensate add latency, complexity, and yet more power.
The market’s response is not a single answer. It is a portfolio of architectures:
XPO (Arista’s External Pluggable Optics): Liquid-cooled, high-density pluggable optics that preserve front-panel serviceability while dramatically increasing bandwidth density. Arista’s white paper describes a 12.8Tbps module and 400W+ liquid-cooled capability. XPO is the clearest near-term competitor to CPO timelines and the most credible delay mechanism.
LPO / NPO (Linear-Drive / Near-Package Optics): Reduce the re-timer complexity by simplifying the electrical interface between switch and optic. Near-Package Optics move the module closer to the ASIC without full co-packaging. A meaningful middle ground that extends the pluggable window.
CPO (Co-Packaged Optics): Full integration of optical engines onto or within the switch ASIC package. Eliminates the faceplate interconnect entirely. Maximum power and density efficiency. First confirmed production platform: Broadcom Bailly, a 51.2T CPO Ethernet switch.
"The faceplate pluggable architecture is no longer the only answer. At 1.6T and beyond, the market splits: XPO and NPO extend pluggables where serviceability matters; CPO wins where power, density, and signal integrity overwhelm the front-panel model." |
CPO is not a bet that every pluggable dies. It is a bet that the highest-value architectural decisions in optical networking — the ones that determine platform lock-in, test infrastructure, laser supply contracts, and photonic foundry relationships — are being made right now, before volume production. The companies establishing those positions are not buying a module. They are buying a decade of toll-collection.
Part II: The Three Bottlenecks — Where the Investment Opportunity Lives
Bottleneck 1: Packaging — EIC/PIC Stacking and the Thermal Trap
The central engineering challenge of CPO is co-packaging: placing the photonic integrated circuit (PIC), which handles optical functions, alongside or beneath the electronic integrated circuit (EIC), which handles the ASIC logic. The two components run at wildly different thermal profiles and have incompatible process technologies. Getting them into the same package without destroying yield or thermal management is the hard problem.
Hybrid bonding — the technique of bonding chips at micron-level precision without solder bumps — is where the industry is converging. It allows tighter integration and better thermal pathways than traditional flip-chip approaches. But it requires different equipment, different inspection methodologies, and significantly more complex yield management than anything the optical supply chain has done before.
TSMC’s COUPE roadmap is the most visible foundry-level validation of this direction. TrendForce reports TSMC is targeting 2026 COUPE production, integrating silicon photonics with CoWoS-based advanced packaging. The foundry that owns the yield recipe for EIC/PIC co-packaging owns the platform. This is not a component business. It is an architecture business.
HYBRID BONDING — THE TECHNIQUE THAT MAKES OR BREAKS CPO Hybrid bonding eliminates solder bumps entirely, connecting chips via copper-to-copper direct bonding at pitches below 10 microns. Better electrical performance, lower power, dramatically tighter integration — but it requires sub-angstrom surface preparation and alignment precision that pushes the limits of current metrology tools. Yield at volume is the unresolved question. |
Bottleneck 2: Test Complexity — Burn-In Is the New Moat
Traditional transceiver testing is hard enough. CPO testing is a categorically different problem. When you co-package optical and electronic components, you cannot test them independently at final assembly — the package is the product. Any fault found at the packaged level is a full-assembly loss.
This drives the economics of CPO test in a direction that favors companies with deep process control capability: the industry will be forced toward comprehensive wafer-level and die-level test before packaging, because the cost of a packaged failure is prohibitive. Burn-in testing — thermal and electrical stress screening that accelerates early-life failure — becomes essential infrastructure, not optional insurance.
TrendForce identifies CPO testing as a critical technology bottleneck that must be mastered before TSMC’s COUPE platform moves into mass production. The test flow involves wafer-level optical and electrical test, followed by known-good-die screening, followed by final package and system-level verification. That is a three-stage test architecture that does not exist in the pluggable world.
Every CPO package must pass through a much more complex test stack before it ships. The companies that build and own that test architecture are not incidentally positioned. They are at the highest-value chokepoint in the supply chain. Tollbooth economics at its most literal.
Bottleneck 3: Lasers — Cooling, Coupling, and the External Source Question
CPO does not eliminate the laser problem. It relocates it.
Some CPO architectures integrate continuous-wave (CW) laser sources close to or within the package — generating significant heat in an already thermally constrained environment. The engineering challenge is twofold: cooling a high-power laser co-located with a hot ASIC, and maintaining optical coupling efficiency at the fiber-chip interface.
But the industry is also moving toward external laser sources precisely to improve serviceability and thermal isolation. The OIF’s ELSFP (External Laser Small Form-Factor Pluggable) implementation agreement defines a front-panel pluggable laser source for CPO systems, placing the laser in the cooler section of the rack and allowing field replacement. That is a different supply chain than an embedded source.
Either way, high-power CW laser reliability, coupling loss, eye safety, and thermal management become first-order design constraints. Grating couplers — the most common fiber-to-chip interface — introduce insertion loss that erodes the power budget. Improving coupler efficiency, or transitioning to edge-coupling architectures, is an active engineering race. The practitioners in our seminar were emphatic: optical loss at the PIC level is not a secondary consideration. It is a primary yield driver.
Part III: Scale-Up vs. Scale-Out — Two CPO Markets, Not One
The mainstream CPO narrative is a datacenter interconnect story: CPO switch ASICs replacing pluggable transceivers at the top-of-rack and spine layers. That is real. But it is half the market.
The second CPO market is scale-up optical I/O — accelerator-to-accelerator, XPU-to-XPU, and in-package optical interconnect for AI systems. Nvidia and Marvell’s partnership explicitly includes silicon photonics collaboration for NVLink Fusion scale-up networking. This is not the same supply chain as switch CPO.
In scale-out switching, the platform owner is the switch ASIC vendor (Broadcom, Marvell) and the system OEM. In scale-up optical I/O, the platform owner may be the GPU/accelerator vendor itself, with optics integrated into the chiplet architecture. These two markets develop on different timelines, involve different foundry relationships, and reward different supplier positions.
The common thread is not the package label. It is the migration of value from detachable modules to co-designed electrical/photonic integration. Wherever that migration occurs, the same economics apply: packaging capability, test infrastructure, and platform co-development relationships determine who collects the toll.
Where the Economics Move
LAYER | COMPANIES | VERDICT | WHY IT MATTERS | RISK |
ASIC / Platform Control | AVGO, MRVL | WINNER | ASIC vendor sets the integration spec and controls the platform. Broadcom’s Bailly is the first confirmed production CPO switch. Marvell is tied to Nvidia NVLink Fusion silicon photonics collaboration. | Execution risk on first-gen CPO platforms; yield problems delay flagship programs |
Lasers / Optical Engines | LITE, COHR, AAOI | WINNER | Nvidia has committed strategic partnerships with both Lumentum and Coherent. Coherent has demonstrated 6.4T CPO and 400G-per-lane InP technology. External laser (ELSFP) architecture creates a recurring serviceable component market. | ELSFP vs. integrated laser architecture choice not yet settled; wrong bet means supply chain exclusion |
Foundry / Silicon Photonics / Advanced Packaging | TSM, AMAT, BESI, AMKR | WINNER | TSMC COUPE targets 2026 silicon photonics + CoWoS production. Foundry with the yield recipe for EIC/PIC co-packaging owns the platform. Advanced packaging equipment vendors (bonding, deposition, metrology) are mandatory infrastructure. | Samsung targeting 2029 CPO turnkey — lagging TSMC creates concentration risk in foundry supply |
Test / Burn-In / Metrology | TER, ATEYY, KEYS, AEHR, ONTO, CAMT | NEAR-TERM WINNER | CPO creates a three-stage test architecture (wafer-level, known-good-die, system-level) that does not exist in the pluggable world. Aehr has shipped wafer-level burn-in for silicon photonics transceivers. TrendForce explicitly identifies this layer as a production bottleneck. | Test methodology not yet standardized — equipment purchasing stalled until architecture is settled |
Thermal / Data Center Infrastructure | VRT, ETN, MOD, NVT | NEAR-TERM WINNER | CPO concentrates thermal density around ASIC + optics + laser power. Liquid cooling is mandatory at scale. Infrastructure vendors benefit from both CPO ramp and parallel XPO liquid-cooling buildout. | Timeline risk — delayed CPO ramp pushes liquid cooling upgrade cycle right |
EDA / Photonic Design Tools | SNPS, CDNS, SIEGY | NEAR-TERM WINNER | Co-simulation of optical and electrical domains simultaneously is a genuine capability gap. New toolchain investment is mandatory for PIC design at CPO scale. | Large EDA incumbents could acquire capability and compress the pure-play margin window |
Front-Panel-Only Pluggable Assemblers (without XPO/LPO/NPO/CPO roadmap) | — | PRESSURE | The pluggable market does not disappear. But growth and margin pool migrate toward architectures that solve power, density, cooling, and serviceability at 1.6T+. Module assembly without a credible next-generation roadmap is a shrinking position. | XPO/LPO success could extend the transition window; faster hyperscaler CPO adoption compresses it |
Pressure Points
PRESSURE POINT | WHAT TO WATCH | TIME HORIZON |
First CPO Switch Production Volume | Broadcom Bailly entering hyperscaler production signals the transition start. Watch for confirmed deployment announcements from cloud providers and system OEM design-win disclosures. | H2 2026–2027 |
Hybrid Bonding Yield Data | Early yield reports on EIC/PIC co-packaging — from foundries, test equipment companies, or IPO filings — will determine whether the CPO roadmap holds. Any public disclosure is a repricing event for packaging and test vendors. | 2026–2027 |
Hyperscaler CPO Specification Release | When a major cloud provider publishes a CPO interface specification (as they did with OCP for networking), it collapses technology optionality and forces supplier convergence. Enormous for equipment and component vendors. | 2026–2027 |
ELSFP / Integrated Laser Architecture Convergence | Whether the market converges on external laser sources (ELSFP) or integrated CW sources determines which laser companies win the design-in and what the serviceability model looks like. | 2027–2028 |
XPO Hyperscaler Adoption Rate | If XPO captures meaningful hyperscaler deployment before CPO reaches volume, the CPO revenue curve shifts right. Monitor Arista’s hyperscaler XPO design wins as the leading indicator of CPO timeline risk. | 2026–2027 |
CPO Burn-In / Wafer-Level Test Standardization | The OIF, IEEE, or a hyperscaler consortium defining the CPO test methodology locks in a multi-year equipment procurement relationship. Resolution of this debate is a purchasing catalyst for test vendors. | 2027–2028 |
Credibility Firewall
COMPANY-DISCLOSED | MODEL-DERIVED | EDITORIAL VIEW |
Broadcom Bailly: industry’s first 51.2T CPO Ethernet switch platform, delivering 70% optical interconnect power improvement (Broadcom press release, confirmed) | CPO/NPO module market opportunity estimated at $5.5B by 2027 in CIR forecasts; definitions of CPO addressable market vary significantly across analyst sources — treat as directional | The 70% power number is the cleanest hard anchor in this issue. Build all valuation arguments around disclosed data, not TAM estimates. |
TSMC COUPE: targets 2026 silicon photonics + CoWoS-based CPO production (TrendForce, April 2026) | Foundry yield at volume for EIC/PIC co-packaging has not been publicly disclosed. Hybrid bonding yield data for CPO context is not available — directional inference only. | TSMC’s COUPE timeline is the most important undisclosed datapoint in the CPO supply chain. Any slip is a sector-wide repricing event. |
OIF ELSFP implementation agreement: defines front-panel pluggable external laser source for CPO systems (OIF press release, confirmed) | Whether the majority of CPO deployments will use integrated vs. external laser sources is an unsettled architectural question — not yet determined by hyperscaler specs | The laser architecture decision is a binary for Lumentum, Coherent, and AAOI. ELSFP confirmation creates a serviceable component market; integrated architecture concentrates laser revenue in foundry yield. |
CPO testing identified as a critical production bottleneck by TrendForce; Aehr has shipped wafer-level burn-in systems for silicon photonics transceivers (TrendForce Insights; Aehr public disclosures) | Bill-of-test economics for CPO at volume are not yet public. Test complexity and cost impact on CPO system economics are directional estimates based on practitioner input | The burn-in / wafer-level test opportunity is the most underpriced layer in this issue. No analyst model currently treats CPO test infrastructure as a standalone investment thesis. It should be. |
Arista XPO announced March 2026: 12.8Tbps module, 400W+ liquid-cooled pluggable capability, front-panel serviceability preserved (Arista press release, confirmed) | XPO hyperscaler qualification status and deployment volume projections are not public — Arista has not disclosed confirmed hyperscaler CPO or XPO design wins | Arista XPO is not evidence of CPO adoption. It is the most credible near-term delay mechanism. Both theses can be true: XPO extends pluggables in segments where serviceability matters, CPO wins where density and power overwhelm the faceplate model. |
Bear Case: Why CPO Could Disappoint
BEAR CASE SPOTLIGHT CPO has been 'three years away' for longer than anyone in the optical industry wants to admit. The cleanest delay mechanism is XPO: if Arista's architecture gives hyperscalers sufficient density and cooling while preserving pluggability, the CPO revenue curve shifts right and the first winners may be XPO suppliers, not full CPO suppliers. Hybrid bonding yield at volume has not been demonstrated in a CPO context. The hyperscalers have their own in-house silicon photonics programs (Google, Microsoft, Amazon) that could vertically integrate past external suppliers entirely. Test methodology is unsettled, which means equipment purchasing is stalled. And the CPO thermal problem may require liquid cooling infrastructure that datacenter operators are not yet ready to deploy at scale. The transition happens. The timeline is genuinely uncertain. 'Hard and late' is a buying opportunity. 'Hard and late because XPO works well enough for long enough' compresses the window significantly. |
Five Things To Do With This Information
1. Map packaging capability, not module volume. The CPO winners are companies with advanced hybrid bonding, precision substrate, and thermal interface materials capability. Revenue scale in pluggable transceivers is not a proxy for CPO positioning. The supply chain reorganization required is substantial, and it runs through a set of companies Wall Street currently prices as if nothing has changed.
2. Watch burn-in and wafer-level test vendors specifically. The CPO test complexity identified by practitioners is not in analyst models. The company that defines the burn-in architecture for CPO packages a recurring, high-margin revenue stream that compounds as the installed base grows. TrendForce identifies this as the production bottleneck. The market has not priced it.
3. Treat the laser architecture decision as a binary. ELSFP vs. integrated CW sources determines which laser companies win the design-in and what the serviceability revenue model looks like. Nvidia's partnerships with Lumentum and Coherent are already placing bets. The decision will be made in design-in conversations happening now, not in volume announcements.
4. Monitor Arista XPO hyperscaler deployment as the leading indicator of CPO timeline risk. XPO is not a CPO confirmation. It is the most credible delay mechanism. If XPO captures meaningful hyperscaler volume in 2026-2027, the CPO revenue curve shifts right. Watch procurement and BOM disclosures, not press releases.
5. Position before the hyperscaler CPO specification. When a major cloud provider publishes a CPO interface spec — as they did with OCP for networking — it immediately prices the winning supply chain into the stocks of companies with confirmed design-ins. The co-development relationships that determine CPO winners are being established now. Waiting for disclosed design wins means paying for what is already priced.
The longest-running theft in American history?
Porter Stansberry's argument: when new money is printed, the dollars you already hold
quietly lose purchasing power, a "theft" you never see on any statement. Whether or not
you agree, the Porter & Company Porter Portfolio Index ETF (PCPP) was built for
investors who'd rather hold a diversified mix than sit in cash. It tracks a rules-based index
split roughly four ways: P&C insurers, capital-efficient equities, hard assets, and
short-term fixed income.
Investing involves risk including possible loss of principal.
Read the prospectus at porterandcofunds.com.
Consider the fund's investment objectives, risks, charges, and expenses carefully before
investing. This and other information are in the prospectus at PorterandCoFunds.com or (833)759-6110; read it carefully before investing. Investing involves risk, including possible loss of principal. PCPP is new and non-diversified. Distributed by Foreside Fund Services, LLC.
News vs. Noise: What’s Moving Markets Today
Yesterday was an attempt to find a bottom, which I think would have worked just fine if we didn’t have this headline….
Markets are only looking to give back some of yesterday’s gains so far. Oil stocks, which we have been constantly telling you to buy on dips, are up nicely. Gold also looks strong this morning. That’s important as war news has been bad for gold lately, so far this morning it’s not.
$GOOGL ( ▲ 0.65% ) has earnings tonight, which ought to be important.
In the whole scheme of things I’m not concerned about war based selloffs, they tend to turn pretty quickly after the next tweet about a cease fire. I’m more looking for signs that the leverage unwind has ended. Yesterday’s buying did not look orderly, so far this morning’s selling has.
ETF News
A Stock I’m Watching
Wrote about this earlier in the week as a beneficiary of Kimi.
In Case You Missed It
Great conversation on wide ranging topics with Kenny Polcari…
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.