Fabric.AI / StableX Technologies (Nasdaq: $SBLX) and Kopin Corporation (Nasdaq: $KOPN): the MicroLED Interconnect Trade

Executive Summary

This is the kind of story that can light up the small-cap tape quickly because it combines several ingredients that retail traders actively chase: artificial intelligence infrastructure, data-center bottlenecks, optical interconnects, MicroLED, a low-float style microcap setup, a corporate name change, a ticker change, a financing, and a visible connection to a company with real operating history in advanced displays and optical systems. StableX Technologies, trading as $SBLX at the time of the announcement, is now presenting itself as Fabric.AI and intends to trade as $FABC. The new pitch is straightforward: AI factories need more than GPUs; they need faster, lower-power ways to move data between compute nodes. Fabric.AI wants to start with MicroLED-based optical interconnects, and Kopin is the enabling MicroLED partner.

The headline is powerful, but the story must be handled with discipline. Fabric.AI is not Nvidia, Broadcom, Marvell, Arista, Coherent, Lumentum, Astera Labs, Credo, or a proven AI-networking supplier. It is a newly repositioned public microcap using a prior listed shell-like corporate history, with minimal legacy revenue, a going-concern warning in its latest annual report before the new financing, a major strategic pivot, and a new Series K preferred-stock and warrant structure that could become highly dilutive if the common equity trades well and shareholder approvals are obtained. In other words, the upside narrative is huge because the addressable market is huge, but the verification gap is also huge.

Kopin is the more mature leg of the trade. Kopin has decades of optical and microdisplay experience, a real product portfolio, defense and industrial exposure, U.S. manufacturing relevance, and a 2025 revenue base that, while down year over year, is not theoretical. Kopin reported $39.3 million of total revenue for fiscal 2025, with defense representing the majority of revenue. It also strengthened its balance sheet through capital raises in 2025 and ended fiscal 2025 with $36.4 million of cash and cash equivalents excluding restricted cash, according to its annual report. Kopin is still a risky small-cap hardware company with customer concentration, defense-program timing exposure, ongoing operating cash burn, preferred-stock dilution, and the BlueRadios litigation overhang. But compared with Fabric.AI, Kopin has real manufacturing and technical depth that gives the story more credibility.

The connection between the two companies is the center of the deep dive. Fabric.AI says its Neural I/o™ chip is being developed in collaboration with Kopin and uses Kopin’s proprietary MicroLED technology and bi-directional NeuralDisplay™ architecture. Kopin’s own updated AI-infrastructure and optical-interconnect pages describe the Neural I/o™ offering as co-developed with Fabric.AI and designed to replace traditional copper and laser systems for ultra-fast, energy-efficient data movement in modern AI infrastructure. The key phrase is “designed to.” This is not yet broad commercial adoption. It is a development-stage, demonstration-chip and validation story.

The most important difference between a tradable catalyst and an investable platform is proof. For the trade, the announcement itself can matter. For the platform, the market will need evidence: technical benchmarks, prototype milestones, third-party validation, customer engagement beyond preliminary discussions, development timelines, manufacturing economics, intellectual-property clarity, margin model, delivery terms, and the exact commercial mechanics between Fabric.AI and Kopin. Until those arrive, the story should be treated as high-volatility, high-narrative, early-stage hardware speculation — potentially explosive, but not yet de-risked.

Why This Matters Now

The announcement lands at a moment when the market is no longer satisfied with generic AI software stories. The AI trade has moved down the stack. Investors now care about power, memory bandwidth, networking, optical links, advanced packaging, switches, retimers, accelerator clusters, and the infrastructure that allows GPUs to behave like one large compute fabric. In that context, “interconnect” is not a boring engineering term. It is one of the key bottlenecks in scaling artificial intelligence.

Modern AI clusters need to move enormous amounts of data between chips, boards, racks, and data-center zones. GPUs can become underutilized if they cannot exchange information quickly enough. As model sizes, inference workloads, and training clusters expand, the limiting factor often shifts from raw compute to data movement and power. Copper interconnects face distance, heat, signal-integrity, and power problems at scale. Laser-based optical solutions can solve parts of the bandwidth-distance problem, but they also introduce cost, integration, packaging, thermal, and reliability questions. That is the opening Fabric.AI and Kopin are trying to frame: use MicroLED pixels not merely as display emitters, but as ultra-fast optical transmitters.

The narrative is clever because it takes a technology Kopin already understands — MicroLED and display architectures — and attempts to repurpose it for a higher-growth infrastructure market. For Kopin, the strategic value is obvious: the company has historically been treated by the market as a defense and optical-systems small cap, not as a core AI-infrastructure name. If the Fabric.AI partnership gives Kopin a credible foothold in AI data centers, the valuation conversation can change. For Fabric.AI, the value is also obvious: it can enter the AI-infrastructure conversation without having to build a MicroLED manufacturing foundation from zero.

That is why the two-stock angle matters. $SBLX / Fabric.AI is the high-beta narrative vehicle. $KOPN is the technology-enablement and supplier/partner vehicle. Traders may chase Fabric.AI because of the tiny market cap optics, the ticker change, and the radical pivot. Traders may chase Kopin because it has a clearer operating base and a direct financial/ownership angle in the Fabric.AI story. The smartest editorial framing is not to say one replaces the other, but to say each carries a different risk/reward profile inside the same catalyst.

The Actual News: What Was Announced

On April 28, 2026, StableX Technologies announced the official launch of Fabric.AI and presented a MicroLED-based optical interconnect technology as the first step in building what it calls next-generation “AI factory” infrastructure. The company says it has exited its prior digital-asset treasury strategy and is now singularly focused on AI infrastructure. It also disclosed that it is changing its corporate name to Fabric.AI and intends to change its Nasdaq symbol to $FABC beginning with the market open on April 29, 2026.

The first product concept is Neural I/o™, a chip being developed with Kopin. According to the company, the chip replaces traditional electrical interconnects with MicroLED-based optical links to enable ultra-high-bandwidth, low-latency communication between compute nodes while improving energy efficiency. Fabric.AI says the technology uses Kopin’s proprietary MicroLED and patented bi-directional NeuralDisplay™ architecture, repurposing programmable MicroLED pixels as optical transceivers.

The press release disclosed that Kopin will own 19.9% of Fabric.AI as part of the development deal. The subsequently filed 8-K gives the relationship more detail: Fabric.AI entered into a Joint Development and License Agreement with Kopin on April 27, 2026, under which Fabric.AI agreed to pay Kopin up to $15 million for development of the Project Technology through at least one successful prototype demonstration. The filing also states that Fabric.AI agreed to issue an initial $5 million purchase order within ten business days after the JDA date, and that the parties expect to negotiate a production-deployment plan after a successful demo, with an additional expected payment of approximately $15 million to $25 million. That is more concrete than a vague memorandum of understanding, but it is still a development-stage milestone, not mass production, hyperscaler qualification, or commercial deployment.

Fabric.AI also announced a $21.5 million financing led mostly by existing investors through the sale of 21,500 shares of newly designated Series K convertible preferred stock with a stated value of $1,000 per share, convertible into common stock at an initial conversion price of $2.51, plus accompanying warrants exercisable at $2.51. Full conversion and full warrant exercise are subject to shareholder approval. This is a critical part of the story because it can fund the pivot, but it can also create a future dilution overhang.

Fabric.AI / StableX Technologies: From Broken Legacy Story to AI Infrastructure Pivot

Fabric.AI is not starting from a clean operating base. That is one of the most important points in the entire analysis. The company’s latest annual report still reflects the older business history. The issuer had been AYRO, then StableX Technologies, and now Fabric.AI. The legacy business had been tied to electric vehicles, specifically the Vanish platform, before the company paused manufacturing and shifted strategy. In the 2025 annual report, the company reported zero revenue for the year, compared with only $63,777 in 2024. It reported cost of goods sold despite no revenue, general and administrative expense above $8 million, and a net loss of approximately $21.1 million for 2025.

The annual report also included a substantial doubt / going-concern note. As of December 31, 2025, the company had approximately $5.0 million in cash and cash equivalents, restricted cash of about $110,000, and marketable securities of approximately $3.17 million. It reported a net loss of approximately $21.1 million and negative operating cash flow of approximately $7.70 million for the year. Management concluded that the existing cash at the time would not be sufficient to fund operations for at least twelve months from issuance of the financial statements. That was before the newly announced $21.5 million financing, but it remains crucial context.

This is why the new financing is central. Without new capital, Fabric.AI’s pivot would be more fragile. With new capital, the company has a path to fund early development, but the cost is a preferred-stock and warrant structure. At the initial $2.51 conversion/exercise price, the preferred stock alone represents a large potential common-share issuance if converted, and the accompanying warrants can add another layer if exercised. Because full conversion and full warrant exercise are subject to shareholder approval, traders must watch the follow-up proxy and 8-K details closely. The capital structure can change dramatically.

The positive case is that Fabric.AI is doing what public microcaps often need to do to survive: abandon a broken or exhausted legacy strategy and reposition into a stronger secular market. The negative case is equally obvious: the company is early, promotional risk is high, the prior operating base is weak, and the market may price the AI pivot faster than the company can prove the technology. That gap between narrative and proof is where volatility lives.

What Fabric.AI is trying to become

Fabric.AI describes itself as an infrastructure company building fabless semiconductor technologies for AI factories. “Fabless” means the company is not trying to own the entire manufacturing footprint. The strategic model is closer to system architecture, chip design, integration, IP development, and partnership-based execution. The first technology area is MicroLED optical interconnects. Over time, the company says it wants to expand into other system-critical semiconductor solutions for AI workloads.

That ambition sounds large, but the market should not confuse ambition with de-risked execution. Building chips and interconnect technologies for data centers is difficult. It requires technical proof, packaging integration, reliability testing, thermal validation, supply-chain discipline, customer qualification, and often long design-in cycles. Hyperscalers and AI-infrastructure customers do not deploy new interconnect technologies simply because the story is attractive. They need cost, reliability, performance, power, and compatibility evidence.

This is the key editorial point: Fabric.AI has announced a direction, not completed a platform. The direction is interesting enough to deserve a deep dive. The platform still needs to be proven.

Kopin Corporation: Why $KOPN Is the Credibility Layer

Kopin is the reason this story deserves more attention than a typical microcap AI rebrand. Kopin is not a newly created AI shell. It is a long-operating optical-systems and microdisplay company. Its portfolio includes active-matrix LCD, FLCoS, MicroLED, OLED, optics, modules, eyepiece assemblies, image-projection modules, and low-power ASICs. The company sells into defense, enterprise, professional, industrial, consumer, and medical applications, with defense currently the largest revenue contributor.

For fiscal 2025, Kopin reported total revenue of $39.324 million, down from $50.335 million in fiscal 2024. Its 2025 revenue mix remained heavily defense-oriented: defense represented approximately $29.361 million of revenue, while industrial revenue was approximately $3.025 million, medical was approximately $594,000, and R&D revenue was approximately $4.590 million. The company disclosed that product sales to defense customers were 74% of total revenue in 2025, compared with 82% in 2024. Customer concentration remains high, with DRS Network & Imaging Systems representing a major portion of revenue.

The 2025 story at Kopin was messy but not empty. Fourth-quarter revenue was hurt by government shutdown and procurement delays, according to management. Kopin also had a significant legal overhang from the BlueRadios litigation, with a $19.7 million judgment accrued and an appeal filed. To post the appeal bond, Kopin deposited approximately $24.2 million, classified as restricted cash. That matters because headline cash balances can look stronger than freely available cash if investors do not separate restricted and unrestricted amounts.

At the same time, Kopin raised substantial capital in 2025. The company reported approximately $51.2 million raised during fiscal 2025 through common stock, pre-funded warrants, and preferred stock. It ended fiscal 2025 with $36.4 million in cash and cash equivalents excluding restricted cash, and management stated this was sufficient to support operations and obligations for at least twelve months from issuance of the financial statements. The company also reported net income of $2.6 million for fiscal 2025, but that result was influenced by non-operating and accounting items; operating cash flow was still negative by approximately $15.5 million.

For the Fabric.AI story, Kopin’s importance is technical. Kopin says the Neural I/o™ offering uses MicroLED and bi-directional NeuralDisplay™ architecture, repurposing MicroLED pixels as high-speed transmitters. Kopin’s site now includes AI Infrastructure and Optical Interconnects pages, which is notable because it shows the company is publicly positioning the partnership as a new market application, not merely a one-off press-release mention. Kopin also states that it is the only U.S.-based company with full MicroLED production capability. That claim is strategically valuable if customers care about domestic supply chains and secure manufacturing.

The Technology: Why MicroLED Optical Interconnects Are Interesting

The core problem is data movement. AI training and inference are increasingly distributed across large clusters. Every accelerator needs fast access to memory, other accelerators, networking fabric, storage, and sometimes rack-scale or data-center-scale communication. A cluster is only as strong as the interconnect fabric that lets its processors exchange information. If the interconnect is too slow or too power-hungry, expensive compute capacity is wasted.

Traditional electrical interconnects use copper. Copper is mature, cheap, and deeply integrated into existing systems, but it becomes problematic as bandwidth, distance, and density increase. Signal degradation, heat generation, electromagnetic interference, and power requirements can become harder to manage. Optical interconnects move data using light rather than electrical signaling. In theory, this can improve bandwidth-distance performance and reduce power per bit. In practice, optical systems introduce their own tradeoffs, including coupling efficiency, packaging complexity, driver electronics, temperature management, cost, yield, and long-term reliability.

MicroLED is usually discussed in the context of displays: tiny light-emitting diodes that can produce bright, efficient, high-resolution images. The Fabric.AI / Kopin concept is different. The idea is to repurpose programmable MicroLED pixels as optical transmitters. Instead of showing an image to a human eye, the MicroLED array sends data optically at high speed. Kopin’s NeuralDisplay™ architecture is described as bi-directional, which suggests the architecture is designed around communication rather than one-way display output.

The attractive theoretical features are easy to understand. MicroLED emitters can be compact, potentially efficient, and arrayed at high density. If they can be modulated fast enough, packaged effectively, and integrated with receivers, drivers, and system-level interfaces, they could become a novel path for short-reach optical links inside AI systems. The company language says the solution is expected to consume less power per bit than existing copper or laser-based solutions. That is the commercial promise.

The hard questions are equally important. What is the demonstrated bandwidth per lane or per array? What is the bit-error rate? What distances are targeted: chip-to-chip, board-to-board, rack-to-rack, or optical backplane? What temperature range can it handle? What is the cost per gigabit? What packaging does it require? How does it compare with silicon photonics, VCSEL-based short-reach links, co-packaged optics, electrical retimers, active copper cables, and established optical module approaches? How much of the system can be built with Kopin’s existing process, and how much requires new manufacturing development? These are not small details; they decide whether the technology becomes a product.

That is why the demonstration chipset matters. A demo chip can move the conversation from concept to measured performance. But the market should demand numbers, not adjectives. “Ultra-high-speed,” “low-latency,” and “energy-efficient” are not enough. The next de-risking moment would be a technical update with measured data, customer testing, or a named design partner.

Financial Snapshot and Capital Structure

Fabric.AI / $SBLX

Fabric.AI’s financial snapshot has to be read in two layers: the pre-pivot annual report and the post-announcement financing. The pre-pivot annual report is weak. The company reported no revenue in 2025, a net loss of approximately $21.1 million, negative operating cash flow of approximately $7.70 million, cash and cash equivalents of roughly $4.98 million, and marketable securities of about $3.17 million. It also disclosed substantial doubt about its ability to continue as a going concern for one year from issuance of the financial statements.

The announced $21.5 million financing changes the short-term liquidity picture but introduces dilution. The Series K preferred stock has a $1,000 stated value and an initial conversion price of $2.51. The accompanying warrants also have an initial exercise price of $2.51 and expire five years from issuance. Full conversion and full warrant exercise are subject to shareholder approval. If common shareholders approve the structure and the common stock remains active, the share count can expand materially. This is not a small footnote. It is one of the main risk factors for common shareholders.

The financing is not automatically bad. Development-stage semiconductor stories need capital. The question is whether the money buys enough progress to justify the dilution. If the company can produce credible technical proof and attract real commercial interest, dilution may be absorbed by a larger valuation narrative. If the company fails to validate the product, the financing becomes a bridge to more dilution rather than a bridge to value creation.

Kopin / $KOPN

Kopin’s financial picture is more complex but more substantial. Fiscal 2025 revenue fell to $39.3 million from $50.3 million in 2024. Defense remained the central revenue base. The company reported net income of $2.6 million, but operating cash flow was negative by $15.5 million. It raised approximately $51.2 million during the year and ended with $36.4 million in cash and cash equivalents excluding restricted cash. Restricted cash is important because of the BlueRadios litigation bond: Kopin deposited $24.2 million related to the appeal bond, and that cash is not the same as free operating cash.

Kopin’s balance sheet was improved by financing, but financing also changed the equity structure. The company issued common stock, pre-funded warrants, and preferred stock. Its preferred stock carries conversion features and dividends. This means Kopin is not a clean, no-dilution hardware compounder. It is a small-cap hardware turnaround with capital raises, litigation, government-program timing, customer concentration, and a new AI-infrastructure optionality layer.

Why the Market Can Get Excited

The bullish setup is not hard to understand. AI infrastructure remains one of the strongest themes in the market. Data movement is a recognized bottleneck. Optical interconnects are a real strategic category. Kopin is a real optical technology company. Fabric.AI has a public vehicle, a fresh name, a fresh ticker, a funding event, a partner, a product name, and a simple story that retail traders can understand quickly. That combination is exactly what creates momentum.

There is also a scarcity angle. Many AI infrastructure leaders already have huge valuations. Traders looking for smaller, more explosive names often move toward second- and third-derivative plays. A microcap connected to MicroLED optical interconnects can become the “cheap AI infrastructure” story of the day, even if the commercial path is long. Kopin can benefit from the same scarcity effect because it is not a mega-cap, yet it can point to real MicroLED capability and actual defense/optical revenue.

For $SBLX, the ticker-change element adds heat. Name and symbol changes often attract attention because screeners, chat rooms, and social feeds start rediscovering the company. The planned move from $SBLX to $FABC is also thematically clean: Fabric.AI sounds like AI compute fabric, not a leftover treasury or EV story. Traders love clean ticker narratives. That does not create fundamental value by itself, but it can create short-term attention.

For $KOPN, the bullish market reaction would likely focus on three questions. First, can the $15 million purchase order contribute meaningful development revenue relative to Kopin’s revenue base? Second, can Kopin’s expected 19.9% Fabric.AI ownership become valuable if Fabric.AI’s market capitalization expands? Third, can Kopin use this partnership to reframe itself from a defense microdisplay company into a broader AI-infrastructure optical technology company? If the answer to even one of these becomes credible, the valuation discussion becomes more interesting.

Why the Story Can Fail

The bearish case is just as important. Many microcap pivots into hot sectors fail. A company can change its name, ticker, website, and strategy faster than it can build a commercial semiconductor product. Fabric.AI’s prior financials show a company with no 2025 revenue, major losses, and a going-concern warning before the new financing. That does not mean the new strategy is impossible, but it does mean the burden of proof is high.

Technology risk is the biggest non-financial risk. The idea may be compelling, but performance claims need validation. Competing solutions are not standing still. Copper interconnects are improving. Optical modules are improving. Co-packaged optics, silicon photonics, VCSEL arrays, advanced SerDes, retimers, active electrical cables, and network-switch architectures are all evolving. Major AI-infrastructure customers are conservative when reliability matters. A new approach must beat existing alternatives not only in a lab, but in deployable systems.

Commercial-cycle risk is also serious. Even if the technology works, customer adoption can take years. Data-center operators need qualification, integration support, software and hardware compatibility, supply assurance, and cost modeling. If Fabric.AI is early in preliminary discussions with hyperscalers, that is interesting but not enough. Preliminary discussions are not purchase commitments. The market may ask for named customers, development agreements, pilot deployments, or measurable design wins.

Capital risk cannot be ignored. Fabric.AI’s financing helps the company move forward, but the preferred/warrant package may pressure common shareholders. Kopin’s financing history also reminds traders that hardware development is capital-intensive. If the demo-chip stage leads to more spending before commercial revenue appears, both companies may need additional funding later.

Finally, narrative-risk is high. AI infrastructure is a crowded storytelling arena. The more crowded the theme, the more aggressively traders must separate verified facts from promotional language. The correct stance is not cynicism; it is discipline. The story is worth tracking, but it is not yet proven.

Retail Sentiment: What Traders Are Likely to Focus On

Retail sentiment around a story like this usually forms quickly around a few simple hooks. The first is “AI infrastructure microcap.” The second is “Kopin owns 19.9%.” The third is “$15 million purchase order.” The fourth is “ticker changing to $FABC.” The fifth is “MicroLED optical interconnects can replace copper and lasers.” These hooks are powerful because they are easy to repeat. That makes them useful for momentum, but also dangerous if traders do not understand what remains unproven.

On Reddit, Stocktwits, and X/Twitter, the optimistic version of the discussion will probably focus on scarcity and optionality. Traders may argue that Fabric.AI is extremely early but has asymmetric upside if it becomes associated with AI data-center infrastructure. They may also argue that Kopin is underappreciated because the market has historically valued it as a defense display company rather than an AI infrastructure component supplier. The retail bull case will likely be emotional and fast: “real AI hardware,” “domestic MicroLED,” “GPU-to-GPU bottleneck,” “new ticker,” “low market cap,” “Kopin stake.”

The skeptical retail discussion will likely focus on dilution, reverse-split history, old AYRO baggage, zero revenue at Fabric.AI, financing terms, and the absence of named hyperscaler customers. Skeptics will also point out that “designed to” language is not proof and that a demonstration chipset is not mass production. For Kopin, skeptics will point to revenue decline, litigation, customer concentration, and dilution from prior capital raises.

The balanced view is that retail attention itself can become a catalyst, but it is not fundamental validation. Social sentiment can move a small-cap stock in the short term. It cannot manufacture product-market fit. For Merlintrader readers, the correct approach is to monitor sentiment as a trading variable while keeping the factual checklist separate.

Bull, Base, and Bear Scenarios

Trader Checklist: What to Watch Next

• Fabric.AI 8-K: exact financing documents, preferred terms, warrant terms, ownership mechanics, development agreement terms, and any risk-factor updates.
• Ticker change: trading under $FABC from April 29, 2026 if completed as announced, with possible scanner and retail-flow impact.
• Demo-chip timeline: any milestone around design, tape-out, prototype, testing, or customer sampling.
• Technical benchmarks: bandwidth, latency, power per bit, distance, reliability, packaging, and comparison with existing copper/laser/silicon-photonics alternatives.
• Customer validation: named hyperscaler, AI-infrastructure partner, data-center integrator, or strategic investor would be much stronger than preliminary discussions.
• Kopin revenue recognition: timing and margin profile of the up-to-$15 million development funding, the initial $5 million purchase order, and any later $15-25 million production plan.
• Kopin litigation: BlueRadios appeal updates and any change in restricted-cash or liability treatment.
• Dilution: shareholder votes, conversion approvals, warrant exercises, and future registration statements.

Bottom Line

Fabric.AI and Kopin have produced one of the more interesting small-cap AI-infrastructure stories of the day. The concept is genuinely relevant: data movement and power efficiency are real bottlenecks in AI clusters, and optical interconnects are a serious area of industry focus. Kopin’s presence gives the story technical credibility that many microcap AI pivots lack. The up-to-$15 million development-funding structure, the initial $5 million purchase order, the possible $15-25 million production-plan payment after a successful demo, and the expected 19.9% Kopin ownership of Fabric.AI give traders tangible hooks beyond generic AI language.

But the story is still early. Fabric.AI’s legacy financial base was weak, the financing is potentially dilutive, the product needs proof, and the market needs to see whether the partnership can move from press release to measurable engineering progress and customer validation. Kopin is the more established company, but it has its own risks: lower 2025 revenue, operating cash burn, litigation, customer concentration, and capital-structure complexity.

The cleanest conclusion is this: $SBLX / Fabric.AI is the explosive narrative leg; $KOPN is the more tangible technology and operating-company leg. Both are connected to the same catalyst, but they are not the same trade. For traders, the opportunity is in the volatility and the possible repricing of a new AI-infrastructure theme. For investors, the burden of proof remains high. The next filings and technical milestones matter more than the first headline.