Gen Z’s AI Confidence Cracks: Young Americans Now Fear Job Loss as Much as Older Workers Young Americans are becoming increasingly worried that artificial intelligence will eliminate jobs. According to a new Pew Research Center survey, 73% of U.S. adults aged 18–29 believe AI will lead to fewer jobs over the next 20 years, up sharply from 61% in 2024. That puts young adults almost level with people aged 30–49 (74%) and 50–64 (72%). Overall, 71% of Americans expect AI to reduce job availability, suggesting that concerns about automation are no longer primarily an older-generation issue—they are becoming a major concern for the workforce’s youngest entrants as well

AI Hardware Demand Growth and Representative US-Listed Companies June 2026 Executive Summary Nvidia’s transition to the Vera Rubin (VR200) platform marks a significant escalation in AI infrastructure complexity and cost. Our BOM teardown of the next-generation Rubin rack reveals a ~2x increase in total rack cost to approximately $7.8 million (vs. ~$4 million for GB300), driven not solely by the GPU/CPU but by sharp revaluations across the supply chain. Key highlights from downstream components include: • PCB content value +233% YoY, the largest increase. • MLCC +182%, reflecting higher density and count (e.g., ~600k MLCCs per VR200 NVL72 server, +30%+ vs. GB300). • ABF substrates +82%, power solutions +32%, and liquid cooling +12%. These upgrades align with broader AI scaling: 800G/1.6T optical transceivers ramping aggressively, glass-based technologies advancing for packaging and interconnects, and hyperscalers prioritizing performance, power efficiency, and thermal management. We expect sustained multi-year tailwinds for the AI hardware ecosystem into 2027+, with Rubin-driven demand accelerating in H2 2026. Investment Thesis: While Nvidia (NVDA) remains the core beneficiary, the supply chain offers diversified exposure. We favor companies with direct exposure to high-growth areas like advanced PCBs, high-speed optics, and glass substrates/optical interconnects. Risks include execution on new capacity, potential margin pressure from rapid scaling, and geopolitical supply chain factors. 1. PCB: Sharpest Value Uplift in Rubin BOM Morgan Stanley’s detailed analysis shows PCB content in the Rubin rack surging +233% versus GB300. This reflects needs for higher layer counts, advanced materials, better signal integrity, and larger formats to support increased power and interconnect density in AI servers. US Representative: TTM Technologies (TTMI) – Leading US PCB manufacturer with strong positioning in high-complexity boards for data center/AI applications. TTM has invested in capacity expansions (e.g., new facilities) to capture AI-driven demand for advanced HDI and high-layer PCBs. 2. MLCC: Density-Driven Surge Nvidia’s VR200 NVL72 platform requires ~600,000 MLCCs per server, over 30% more than GB300. Combined with the +182% value increase in the BOM, this underscores tightening supply for high-capacitance, high-reliability MLCCs in power delivery and decoupling for AI accelerators. Exposure Note: The MLCC market is dominated by Asian players (e.g., Murata, Samsung Electro-Mechanics, Yageo). US-listed indirect exposure may come through broader electronics or power solution providers, but direct pure-play opportunities are limited. Watch for capacity utilization tightness benefiting the ecosystem. 3. Optical Communication: 800G/1.6T Ramp Accelerating Chinese leader Zhongji Innolight reported Q1 2026 net profit +262% YoY, driven by strong 800G/1.6T shipments, with expectations of significant full-year growth. This mirrors industry-wide momentum as AI clusters shift toward higher-speed optics for reduced latency and power in scale-out/scale-up networking. Nvidia’s investments in photonics and CPO further validate the trend. US Representatives: • Coherent (COHR) and Lumentum (LITE): Key players in optical components and transceivers; Nvidia has made substantial equity investments to secure capacity. • Corning (GLW): Major beneficiary via optical fiber, connectivity, and glass technologies (detailed below). 4. Micro-LED/Glass Substrates & Optical Interconnects: Strategic Partnerships Accelerating On May 20, 2026, BOE announced a cooperation MOU with Corning covering glass-based encapsulation carriers, foldable glass, perovskite substrates, and optical interconnect applications. This aligns with industry shifts toward glass cores for superior flatness, thermal stability, and integration in advanced packaging and photonics—critical for next-gen AI as organic substrates hit limits. US Representative: Corning (GLW) – Central to Nvidia’s optical strategy with multi-billion partnerships, new US optical factories, and expansion in fiber/photonics for AI data centers. Recent deals position GLW for 10x+ capacity growth in key areas. AI Hardware Demand Growth & US-Listed Representative Companies Table Component Demand Growth (vs. GB300) Key Drivers US-Listed Reps Investment Rationale PCB +233% value Higher layers, HDI, signal integrity TTM Technologies (TTMI) Direct AI server/backplane exposure; US capacity expansion MLCC +182% value; +30%+ count Power density in servers Limited direct (ecosystem via power suppliers) Supply tightness supports pricing/volume Optical Comm (800G/1.6T) Strong ramp (e.g., +262% profit ex.) Scale-out networking, CPO transition Coherent (COHR), Lumentum (LITE), Corning (GLW) Nvidia investments; transceiver/fiber boom Glass Substrates/Interconnects Emerging (MOU-driven) Packaging, photonics, thermal/optical Corning (GLW) Nvidia factory deals; US manufacturing tailwinds Power & Liquid Cooling +32% / +12% Higher TDP (e.g., 2300W GPUs) Indirect (ecosystem) Secondary but critical for rack deployment Source: Morgan Stanley BOM analysis, company reports, industry data. Growth metrics approximate from Rubin teardown. Outlook & Risks We project robust 2026-2027 growth in AI capex, with Rubin shipments catalyzing another leg-up in component demand. Optical and advanced substrate shifts could extend the cycle beyond traditional GPU focus. Hyperscalers’ vertical integration and US onshoring (e.g., Corning/Nvidia factories) add resilience. Key Risks: Cyclical capex pauses, yield/execution challenges on new tech (glass/CPO), commodity volatility in passives, and intense competition in Asia-heavy segments. Valuation multiples in the space have expanded; selectivity is key. Recommendation: Overweight select supply chain names with strong Nvidia alignment (e.g., TTMI for PCBs, COHR/LITE/GLW for optics/glass). Monitor Q2 2026 earnings for confirmation of Rubin ramp momentum.

Read-through from $GLW (Corning) Q3 2026 earnings, for the sector. 1. $NVDA: AI fiber is growing faster than GPUs ship. Gen AI product sales more than doubled YoY and outgrew GPU units, so the fiber attached to each GPU is rising. All of that is scale-out, linking GPUs across a cluster, at roughly 16 fibers per GPU. Corning signed a multiyear manufacturing partnership with NVIDIA in May and says it sits inside the scale-up ecosystem. Weeks is "quite convicted that this accelerates in 2028." Read: neutral to bullish. The networking around NVIDIA systems keeps getting richer, but this is not a GPU volume signal. (inferred: NVIDIA not named on this call) 2. $META: hyperscaler capex is being prepaid, not just planned. In January Meta signed a deal worth up to $6B through 2030 for Corning fiber, cable and connectivity, anchoring a cable plant in Hickory, NC. Corning didn't name Meta on this call. It did say hyperscalers now put cash up front to fund capacity, which Corning returns as they buy the committed volume. Net customer deposits reached $639M in the first nine months. Customers prepaying for fiber they'll use in 2027 and 2028 is a harder durability signal than any capex guide. Read: neutral for $META's stock (it's spend), bullish for the networking suppliers selling into that build. (inferred: Meta not named on this call, deal confirmed separately) 3. $VZ $T: carriers are now buying fiber for data centers. Corning signed new long-term agreements with Verizon and AT&T this quarter (plus Zayo, which is private). They center on DCI, data center interconnect, the long-haul fiber linking one data center to another. Corning now says it will beat its old target of a $1B DCI business by 2030 on both timing and size. It also expects fiber-to-the-home builds by the big carriers to keep growing. For the carriers, it means hyperscaler wholesale demand is real enough to lock in multi-year supply. Read: mildly bullish for their wholesale fiber revenue, neutral on the stocks until that capex earns a return. (direct: management named them) 4. $LUMN: the DCI demand behind its fiber build is bigger than Corning planned for. Lumen came up on the call as an existing carrier partner. In 2024 it reserved 10% of Corning's global fiber capacity for two years to more than double its intercity network for cloud data centers. Corning raising its own DCI expectations is a demand check on exactly that bet. The catch is that supply is tight. Corning hands its newest high-density cable first to long-standing customers and those who prepay for capacity, so Lumen's next tranche likely needs a bigger commitment. Read: bullish on demand, neutral on supply terms. (direct on the relationship, inferred on the allocation) 5. $AAPL: cover glass fully onshored, device market soft. By the end of 2026, all cover glass for iPhone (including iPhone Duo) and Apple Watch sold worldwide will come from Corning's Kentucky plant, which makes a tariff-exposed input fully U.S.-made. Separately, Corning sized the device market for Q3: notebooks and handhelds down about 20%, TVs down 4%, with memory prices blamed. That's an industry figure, not an Apple one. Read: neutral, a supply-chain positive against a weak unit backdrop. (direct: management named them) 6. $CRDO: copper holds scale-up through 2027, and 2028 is the clock. Scale-up, the short links between GPUs inside one system, is 100% copper today per Corning. That is where Credo's AECs live (active electrical cables, copper with signal-boosting chips). Corning expects optics to start taking scale-up in 2028, and fully optical scale-up means about 160 fibers per GPU, 10x today. Credo is hedging with its own optical lines (ZeroFlap optics, plus active LED cables slated for fiscal 2028), but the copper franchise now has a stated window. Read: neutral near term, bearish on long-dated copper share. (inferred) 7. $LITE: Corning's scale-up date lines up with Lumentum's. Lumentum has guided to its first scale-up CPO shipments by late calendar 2027 (co-packaged optics, lasers and optics mounted right next to the chip) and calls optical scale-up its biggest growth driver. Corning sits at the fiber layer and independently puts the acceleration in 2028. Two suppliers at different layers pointing at the same window makes the timeline more credible. It is still management conviction, with no adoption rate attached. Read: bullish on timing, not yet on size. (inferred) 8. $APH: Corning's sold-out high-density cable is Amphenol's opening. Amphenol closed its $10.5B purchase of CommScope's connectivity business in January, which makes it a direct competitor in data center fiber connectivity at about $4.1B of 2026 sales. Corning says demand for its newest high-density cable is outrunning its ramp, so it rations by relationship length and willingness to prepay. Buyers who miss Corning's allocation have to source somewhere else. Read: bullish for $APH share gains, until Corning's planned 10x U.S. connectivity capacity build closes the gap. (inferred) Bottom line: The strong read is fiber demand. Enterprise sales +84%, Gen AI sales more than doubled, and the DCI target beaten early are direct evidence that hyperscalers and carriers are still building, and the deposit model shows that spend is committed into 2027 and 2028. That supports the carriers ($VZ $T $LUMN) and the connectivity competition ($APH). The speculative read is scale-up timing. The $NVDA, $CRDO and $LITE chain rests on Corning's conviction and Lumentum's matching timeline, and nobody has given an adoption rate. $GLW also traded down about 4% on a beat, so the market isn't paying for "demand is strong" alone anywhere in this chain.

# Practical and Useful Patterns with ADK 🎭 Fixed-prompt tests can't replicate the unpredictable behavior of real users. ADK's User Simulation feature enables persona-driven conversational testing — "impatient customer," "vague requester" — to automatically verify agent robustness. 📌 Title: User Simulation 🔗 URL: 🧩 Overview ADK's User Simulation dynamically evaluates conversational agents using AI-generated user prompts. Instead of fixed test cases, a simulated user adapts to agent responses, making it ideal for testing unpredictable conversation flows. A ConversationScenario consists of three elements: `starting_prompt` (initial message), `conversation_plan` (high-level goals the user must achieve), and `user_persona` (user traits and interaction style). 🛠 Usage Basic user simulator configuration: ```json { "user_simulator_config": { "model": "gemini-flash-latest", "model_configuration": { "thinking_config": { "include_thoughts": true, "thinking_budget": 10240 } }, "max_allowed_invocations": 20 } } ``` ADK provides three pre-built personas: | Behavior | EXPERT | NOVICE | EVALUATOR | |----------|--------|--------|-----------| | Advance | Detail-oriented | Goal-oriented | Detail-oriented | | Answer | Relevant questions | All questions | Relevant questions | | Correct inaccuracies | Yes | No | No | | Troubleshoot errors | Once | Never | Never | | Tone | Professional | Conversational | Conversational | Define custom personas directly in ConversationScenario JSON: ```json { "user_persona": { "id": "IMPATIENT_USER", "description": "User in a rush, easily frustrated", "behaviors": [ { "name": "Short responses", "description": "Provides brief, incomplete replies", "behavior_instructions": [ "Keep responses under 10 words", "Omit polite phrases" ], "violation_rubrics": [ "Response exceeds 10 words", "Response is overly polite" ] } ] } } ``` Auto-generate evaluation scenarios: ```bash adk eval_set generate_eval_cases \ --user_simulation_config_file= ``` 🏗 Practical Patterns **Robustness Testing with "Impatient Customer" Persona**: Use an IMPATIENT_USER persona that gives sub-10-word responses and skips pleasantries. This tests whether your agent can gather sufficient information through proper follow-up questions despite receiving minimal input. Especially valuable for customer support agents. **"Vague Requester" Persona**: Simulate users who give abstract, unspecific requirements. Test whether your agent asks clarifying questions to pin down needs, rather than making assumptions or going off-track with ambiguous input. **Leveraging Pre-Built Personas**: EXPERT validates technical accuracy, NOVICE verifies that explanations are beginner-friendly, and EVALUATOR provides neutral performance assessment. Each reveals different aspects of agent quality. **Auto-Generated Scenarios for Coverage**: Use `generate_eval_cases` to automatically create diverse conversation scenarios, discovering edge cases you wouldn't think of manually. Specify `generation_instruction` for scenario direction and `environment_context` for backend state to produce realistic scenarios. 💡 Use Cases 🎭 Verifying customer support agent adaptability across diverse user types 🧪 Discovering edge-case conversation paths beyond fixed tests 🤖 Domain-specific robustness testing with custom personas 📊 Comprehensive quality evaluation through mass scenario generation ⚠️ Caveats - Without proper `max_allowed_invocations` settings, simulations may run indefinitely. Adjust from the default of 20 based on your domain. - Auto-generating scenarios requires the Vertex Gen AI Evaluation Service API. - Vague `violation_rubrics` in custom personas can lead to inconsistent persona behavior. Define specific violation conditions. - Enabling thinking in the simulator model produces more natural dialogue but increases costs. ✨ User Simulation breaks through the limitations of fixed tests. Use personas to simulate "how real users would behave" and verify agent robustness from multiple angles! #ADK# #AIAgent#

# Practical and Useful Patterns with ADK 🧪 How do you verify that a prompt change hasn't degraded your agent's quality? ADK's Evaluation framework provides two approaches — test file method and dataset method — to systematically validate agent performance. Integrate it into CI to automate regression testing on every prompt change. 📌 Title: Evaluation Overview 🔗 URL: 🧩 Overview ADK provides two primary evaluation methods. The **test file method** uses `.test.json` files containing single sessions (with multiple turns) for unit-test-like fast iteration. The **evalset method** supports multiple lengthy sessions across many turns, suitable for integration testing. Some criteria (such as response quality, safety, and multi-turn quality) require the Vertex Gen AI Evaluation Service API. Built-in criteria include `tool_trajectory_avg_score`, `response_match_score`, `hallucinations_v1`, `safety_v1`, and more. 🛠 Usage Test files describe user queries, expected tool use trajectories, and expected responses. Three execution methods are available: **Web UI** for interactive evaluation: ```bash adk web path/to/agents_dir ``` **Pytest integration**: Call `await AgentEvaluator.evaluate()` with `agent_module` set to the target agent module name (e.g., `"home_automation_agent"`) and `eval_dataset_file_path_or_dir` pointing to the test file path (e.g., `"tests/integration/fixture/test.test.json"`). **CLI execution**: ```bash adk eval --config_file_path= ``` Configure evaluation thresholds in `test_config.json`: ```json { "criteria": { "tool_trajectory_avg_score": 1.0, "response_match_score": 0.8 } } ``` By default, tool trajectory requires 100% match while response matching accepts 80% similarity. 🏗 Practical Patterns **Regression Testing for Agent Updates**: Run existing test files after every prompt or tool definition change. `tool_trajectory_avg_score` quantitatively verifies tool selection accuracy, while `response_match_score` checks response quality. Early detection of quality degradation lets you iterate on prompts with confidence. **Fast Iteration with Test File Method**: During development, the `.test.json` approach works like unit testing. Define expected outcomes for single sessions, run them via pytest, and get quality feedback in seconds. Gradually expand test coverage by adding test files for each edge case. **Batch Evaluation with Dataset Method**: For complex, production-like scenarios, use the evalset method to evaluate many sessions at once. Multi-turn conversation flows and diverse user patterns can be comprehensively tested as integration tests. **CI Pipeline Integration**: Embed `adk eval` in your CI pipeline to automatically run quality checks on every pull request that modifies prompts. Fail the CI when scores drop below thresholds, preventing quality-degraded prompts from being merged. 💡 Use Cases 🔄 Automated regression testing on prompt changes ⚡ Fast development iteration with the test file method 📊 Comprehensive batch evaluation with the dataset method 🔗 Quality gate integration in CI/CD pipelines ⚠️ Caveats - Some criteria (such as response quality, safety, and multi-turn quality) require the Vertex Gen AI Evaluation Service API, which is a paid Google service. - Setting thresholds too high increases false negatives; too low risks missing quality degradation. Adjust incrementally from defaults. - LLM-based criteria (e.g., `hallucinations_v1`) can vary between judge model runs. Use `num_samples` for multiple sampling to improve reliability. ✨ ADK's Evaluation framework transforms agent quality management from intuition to quantitative benchmarks. Start with the test file method and progress to CI integration! #ADK# #AIAgent#

$GLW KEY READ-THROUGHS FROM CORNING’S Q3 2026 EARNINGS CALL Corning’s results point to an increasingly uneven technology investment cycle: AI connectivity demand is expanding through higher content, greater network density, and premium products, while rising component costs are suppressing consumer-device volumes. The strongest positive implications concern optical interconnect suppliers, active optical components, and operators monetizing existing fiber infrastructure. The strongest negative implications concern PC manufacturers and display producers facing simultaneous demand and procurement pressure. The critical distinction is between growth in infrastructure content and growth in underlying equipment units. Corning’s performance supports higher networking expenditure per accelerator, not an equivalent acceleration in GPU shipments or a universal earnings upgrade across technology. The following read-throughs have high conviction on the direction of the economic mechanism; magnitude assessments describe fundamental exposure, not predicted share-price moves. OPTICAL CONNECTIVITY AND AI NETWORKING HIGH-DENSITY CONNECTIVITY CAPTURES MORE VALUE PER DEPLOYMENT, FAVORING SPECIALISTS OVER COMMODITY FIBER EXPOSURE (READ-THROUGH 1) Affected companies: Amphenol (APH: United States); Fujikura (5803: Japan). Directional impact and magnitude: Positive. Material for the relevant optical-connectivity businesses and potentially meaningful at the consolidated level, but not a basis for transferring Corning’s growth or margins directly to either company. Call evidence: Optical Communications revenue increased 44%, Enterprise Networks revenue increased 84%, and Gen AI product sales increased more than 100%. Management attributed the performance to faster and broader adoption of its innovations, explicitly noting that growth exceeded GPU unit growth. Management also stated, “Our overall price is going up,” and explained that new products generate “more revenue per element of capacity.” The transmission mechanism is higher connectivity content and greater engineering value per installation. More densely connected AI clusters require additional fiber assemblies, connectors, termination systems, and increasingly demanding cable-management solutions. Products that reduce installation space, labor, and deployment time can capture value beyond their raw-material content. This supports revenue growth through both physical demand and product mix, rather than requiring accelerator shipments to grow at the same rate. Amphenol is a particularly relevant listed exposure because the acquired CommScope Connectivity and Cable Solutions assets now sit within its earnings base. Its July earnings call identified IT datacom as 43% of group sales and described advanced optical interconnect as a major growth driver within the acquired business. Fujikura has directly applicable high-density technology, including its commercialized 4,000-fiber SWR/WTC cable for Japanese data centers. These are verified product exposures, not inferred awards from Corning’s customer agreements. The incremental insight is that capacity measured in fiber volume can understate revenue opportunity when customers migrate toward higher-value connectivity systems. Conversely, owning commodity fiber capacity alone is insufficient. Corning emphasized integration, manufacturing capabilities, customer co-development, and long-term relationships as competitive advantages. That supports differentiation within the cable sector rather than a blanket positive view on all fiber producers. Near-term catalyst: Amphenol’s forthcoming Q3 results and the November Japanese earnings cycle provide opportunities to test whether order conversion, optical mix, and capacity availability support stronger forward expectations. Amphenol had already guided for strong sequential IT datacom growth, so Corning’s results corroborate an established trend rather than establish an entirely new thesis. Longer-duration implication: Through 2027 and beyond, increasing connectivity complexity can sustain growth above underlying equipment-unit growth. However, Corning’s customer commitments belong to Corning. They cannot be assigned to competitors, and fiber shortages can constrain downstream assembly suppliers as well as benefit upstream producers. ACTIVE OPTICS HAVE A CURRENT SCALE-OUT GROWTH ENGINE, WITH SCALE-UP PROVIDING A LATER OPPORTUNITY (READ-THROUGH 2) Affected companies: Lumentum Holdings (LITE: United States); Coherent (COHR: United States). Directional impact and magnitude: Positive. Potentially substantial for datacenter optical components and systems; significant consolidated relevance for both companies. The immediate implication is stronger demand durability, not a mechanically calculable revenue beat. Call evidence: Management emphasized that current optical growth is driven by scale-out. It contrasted approximately 16 fibers per GPU in current architectures with approximately 160 in an illustrative fully optical scale-up configuration, while reiterating, “We continue to be quite convicted that this accelerates in 2028.” The transmission mechanism runs through optical endpoints, bandwidth, and component complexity. Expanding optical networks require the lasers, receivers, optical engines, and transceiver systems that transmit and recover signals. Lumentum and Coherent participate in these active optical layers, whereas Corning’s disclosed growth primarily validates the passive connectivity and physical-network expansion supporting them. The most important implication is that current earnings growth does not require immediate mass adoption of optical scale-up. Existing scale-out deployments are already producing exceptional connectivity demand. That supports the runway for active optics used in today’s networks, while scale-up creates an additional market rather than the sole justification for current investment. The 2028 timing should not be interpreted as the first possible revenue from in-rack optics. Lumentum’s August release already disclosed increasing demand for high-power co-packaged-optics lasers, an initial external-laser-source module order, and near-packaged-optics engagements. Corning’s statement concerns acceleration of adoption, not the absence of early commercial activity. This distinction prevents both premature revenue extrapolation and an erroneous conclusion that all emerging optical products remain commercially irrelevant until 2028. Near-term catalyst: Coherent’s November 4 results and the upcoming Lumentum fiscal Q1 reporting cycle should clarify capacity conversion, optical product mix, and forward demand. Their prior disclosures already described strong AI-related demand, so the value of Corning’s call is independent ecosystem confirmation and evidence that demand remained strong through the September quarter. Longer-duration implication: Optical penetration increases the addressable market, but it does not preserve every existing product architecture. Co-packaged optics can shift value between pluggable modules, optical engines, and external lasers. The high-conviction conclusion is increased demand for optical functionality, not permanent protection of every current module format. The 16-to-160 fiber comparison also cannot be converted into a 10x revenue forecast for either supplier. SCALE-OUT NETWORKING DEMAND REMAINS STRONG WITHOUT REQUIRING AN EARLY SCALE-UP INFLECTION (READ-THROUGH 3) Affected company: Arista Networks (ANET: United States). Directional impact and magnitude: Positive, with moderate incremental importance to near-term expectations and substantial relevance to the longer-term AI networking opportunity. Call evidence: Enterprise Networks revenue increased 84%, Gen AI product sales more than doubled, and management stated that current growth remains driven by scale-out. Customers continued to request “more of our new products, faster and broader.” The transmission mechanism is demand for the network connecting accelerators and compute clusters. More extensive optical fabrics require switching capacity, network operating software, telemetry, congestion management, and deployment engineering. Arista participates through its datacenter networking systems and software, including its October 7 announcement of open Ethernet rack-scale architectures addressing scale-up and scale-out applications. The actionable point is that the present networking investment cycle is not waiting for optical scale-up. Corning’s results support continued deployment of the larger fabrics that underpin Arista’s existing opportunity. The evidence is stronger for the persistence of networking expenditure than for a particular protocol or supplier’s share. That limitation is essential. Corning sells connectivity used across different architectures. Its fiber growth does not establish Ethernet share gains over InfiniBand, prove Arista wins against competing systems, or imply an identical increase in switch ports. Fiber count, link speed, topology, and switching capacity do not move in fixed proportions. Near-term catalyst: Arista reports Q3 results on November 3. Its August release already guided to approximately $3.3 billion of Q3 revenue following 37.7% Q2 revenue growth. Corning therefore supports the demand backdrop behind that outlook rather than independently establishing a beat. The more useful trading signal will be the subsequent-quarter outlook, delivery constraints, and evidence of customer acceptance and revenue conversion. Longer-duration implication: Increasing network complexity can make system-level performance and software execution more valuable. The call supports a durable networking opportunity, but not an indiscriminate upgrade to every AI hardware supplier. Portfolio exposure to Arista, optical components, and connectivity assemblies remains correlated through the same hyperscaler deployment cycle. TELECOMMUNICATIONS AND DATA CENTER INTERCONNECT EXISTING CONDUIT BECOMES MORE VALUABLE AS HIGH-DENSITY FIBER EXPANDS ITS EARNING CAPACITY (READ-THROUGH 4) Affected company: Lumen Technologies (LUMN: United States). Directional impact and magnitude: Positive. Material to the strategic value and potential economics of Lumen’s intercity network; moderate near-term earnings significance because construction, activation, and revenue recognition take time. Call evidence: Management said carrier customers want newer products because they can fit approximately 2x to 4x more fiber into existing conduit. It also stated that DCI would “clearly exceed both the timing and the magnitude” of Corning’s previously communicated $1 billion opportunity by the end of the decade. The transmission mechanism is improved monetization of existing network infrastructure. Rights-of-way, conduit, permitting, and construction are costly and time-consuming to replicate. Installing more fiber within existing routes can expand sellable connectivity without proportionately increasing civil-engineering expenditure. The economic benefit therefore extends beyond lower cable cost to faster deployment and better utilization of embedded infrastructure. Lumen has a verified relationship with Corning involving next-generation fiber for its intercity network and Private Connectivity Fabric. Its previously disclosed architecture uses existing conduit alongside new network construction. The read-through is consequently more direct than a generic positive view on telecom spending. The non-obvious implication is that AI networking demand can increase the value of existing physical infrastructure even while conventional telecom services remain mature or declining. Carrier growth of only 4% at Corning also argues against interpreting the call as evidence of a broad telecom recovery. The attractive opportunity is concentrated in DCI and specific network assets. Near-term catalyst: Lumen reports Q3 results on November 4. The relevant measures are contracted-project conversion, customer activation, construction milestones, cash funding, and the relationship between incremental investment and recognized revenue. Corning’s commentary validates deployment economics, but it is not a new Lumen contract announcement. The larger DCI opportunity had also been discussed before this call. Longer-duration implication: Through 2027 to 2028, high-density deployment can improve the return on existing routes and support network-service growth. However, 2x to 4x more fiber per conduit does not mean 2x to 4x more revenue or profit. Utilization, pricing, customer concentration, and construction obligations remain decisive. Corning’s customer-funding discussion also reinforces that supplier cash receipts can correspond to earlier cash commitments elsewhere in the chain; financing risk is redistributed rather than eliminated. PERSONAL COMPUTERS AND MEMORY PC REVENUE CAN RISE WHILE THE UNDERLYING UNIT AND MARGIN ECONOMICS DETERIORATE (READ-THROUGH 5) Affected company: HP Inc. (HPQ: United States). Directional impact and magnitude: Negative. Material for Personal Systems volume and margin expectations, with meaningful consolidated earnings sensitivity. The call does not independently establish a miss against HP’s existing guidance. Call evidence: Corning reported that notebook and handheld units declined approximately 20%, with those markets “significantly impacted by memory prices.” Glass Innovations sales declined only 3%, illustrating that a supplier benefiting from higher content can appear resilient while underlying device demand weakens substantially. The transmission mechanism is a procurement-cost and demand-elasticity squeeze. Higher memory costs increase the bill of materials. Passing those costs through can raise device prices and reduce unit demand; absorbing them protects volumes at the expense of gross profit. Either outcome can weaken operating leverage in a relatively low-margin hardware business. HP’s already-reported fiscal Q3 results demonstrate why revenue alone is an inadequate measure of health. Personal Systems revenue increased 18%, but total units declined 16%, and the segment operating margin was 4.6%. Corning’s September-quarter commentary extends the evidence of weak device volumes rather than predicting that already-reported HP quarter. The incremental concern is that persistent component inflation limits the pace of a subsequent margin recovery. The distinction between consumer and enterprise spending also matters. Strong AI infrastructure expenditure does not imply a synchronized refresh cycle in end-user PCs. Memory suppliers and AI infrastructure vendors can capture a greater share of technology budgets while device affordability worsens. Near-term catalyst: HP’s fiscal year-end results and initial fiscal 2027 outlook should establish whether pricing, mix, and procurement actions can offset the unit pressure. The most important indicators are Personal Systems margins, units, channel inventory, and memory-cost assumptions rather than nominal revenue growth. Longer-duration implication: During 2027, recovery requires either more affordable memory, greater customer willingness to pay, or enough product differentiation to protect both price and volume. Counterevidence must be retained: HP raised its full-year outlook after fiscal Q3. Corning’s call strengthens the operating-risk assessment but does not justify treating previously recognized cost pressure as a new, automatic guidance cut. MEMORY PRICING POWER IS BEING CONFIRMED DOWNSTREAM, BUT DEMAND DESTRUCTION IS A REAL LIMIT (READ-THROUGH 6) Affected company: Micron Technology (MU: United States). Directional impact and magnitude: Positive for near-term pricing durability, with moderate incremental significance to earnings expectations. Negative for the assumption that consumer memory demand can absorb indefinitely higher prices without adjustment. Call evidence: Corning explicitly linked approximately 20% declines in notebook and handheld units to memory prices, while simultaneously reporting more than 100% growth in Gen AI connectivity products. The transmission mechanism is a transfer of economic value toward a constrained input supplier. The fact that downstream manufacturers are reporting material demand effects from memory prices provides independent evidence that pricing pressure is commercially significant. For Micron, stronger realized prices can support revenue and gross profit even when some consumer-device categories contract. The important distinction is between device units, memory bits, and memory revenue. A decline in notebook shipments does not necessarily produce an equal decline in aggregate memory demand because capacity per device and datacenter consumption can change. Conversely, strong pricing does not establish healthy consumer bit demand. Corning provides evidence of downstream affordability pressure, not a complete DRAM or NAND supply-demand balance. Micron had already reported fiscal 2026 results before Corning’s call and expressed confidence in fiscal 2027, supported in part by strategic customer agreements. Corning therefore corroborates the pricing environment rather than creates a fresh forecast for a completed Micron quarter. It does not quantify an incremental Micron selling-price increase or distinguish the effects across DRAM, NAND, and high-bandwidth memory. Near-term catalyst: Late-2026 and early-2027 pricing negotiations, shipment commentary, and fiscal 2027 guidance updates will show whether pricing strength continues to outweigh weaker consumer volumes. Corning’s evidence is more supportive of existing pricing assumptions than of raising unit-demand forecasts. Longer-duration implication: The memory cycle can remain favorable while downstream hardware economics deteriorate, creating meaningful cross-portfolio divergence. However, persistent affordability pressure eventually encourages lower specifications, deferred replacements, and inventory adjustment. The high-conviction positioning implication is to distinguish upstream pricing exposure from downstream device-volume exposure, not to treat all technology demand as moving together. DISPLAY PANELS AND CONSUMER ELECTRONICS GLASS PRICE IMPLEMENTATION ADDS A 2027 COST HEADWIND TO WEAK DISPLAY END MARKETS (READ-THROUGH 7) Affected company: BOE Technology Group (000725: China). Directional impact and magnitude: Negative for display manufacturing economics before mitigation. Moderate potential earnings significance, but no defensible company-specific EPS reduction can be calculated from the disclosed information. Call evidence: Corning reported TV units down 4% and notebook and handheld units down approximately 20%. Management said display-glass price increases would begin taking effect in Q4 but were “really geared around 2027 and beyond,” addressing a weaker yen and rising input costs. The transmission mechanism is pressure on a key manufacturing input at a time when downstream demand limits pricing flexibility. BOE has a verified, long-standing display relationship with Corning, reinforced by the companies’ May 2026 strategic-collaboration announcement. Higher glass procurement costs can reduce panel margins unless offset by panel-price increases, supplier negotiations, currency movements, manufacturing efficiencies, or favorable product mix. Corning’s September 11 announcement provides useful scope beyond the call: the adjustment was 15% or more for yen-denominated display-glass products, varying by product and generation size. That is not a 15% increase in BOE’s total manufacturing cost, nor necessarily a 15% increase in its renminbi procurement cost. BOE’s supplier mix, contract currency, hedging, and effective purchase-price changes were not disclosed. The more actionable insight is timing. A relatively limited Q4 impact should not be interpreted as disappearance of the issue. Corning explicitly placed the intended economic effect in 2027. Forecasts assuming a clean display-margin recovery should therefore incorporate the possibility that substrate costs rise before panel selling prices fully adjust. Near-term catalyst: Year-end procurement negotiations, panel pricing, and 2027 margin commentary are the relevant events. This is implementation confirmation of an announced increase, not a new price announcement on October 27. Corning’s end-market statistics also do not establish an equivalent decline in BOE shipments because panel area, product mix, and market share can offset weaker device units. Longer-duration implication: Concentrated, technically demanding upstream materials can retain pricing leverage even when downstream manufacturing is competitive. The geographic signal is that substantial production and customer relationships in China do not insulate panel makers from supplier currency economics and globally negotiated input prices. SOLAR MANUFACTURING PROVEN US MODULE CAPACITY BECOMES MORE VALUABLE WHEN TRADE PROTECTION COINCIDES WITH SLOW DOMESTIC RAMPS (READ-THROUGH 8) Affected company: First Solar (FSLR: United States). Directional impact and magnitude: Positive, with moderate significance for future booking economics and competitive positioning, but limited direct support for an immediate quarterly earnings upgrade. Call evidence: Corning reported solar revenue growth of 37%, but production ramped “at a slower pace than anticipated,” and the segment remained loss-making. Management said Section 232 action improved the pricing environment and demand for US-made solar products. It also described profitability improvement as gradual through 2027. The transmission mechanism combines a higher effective competitive price for imports with slower-than-planned expansion of domestic alternatives. Trade protection increases the value of qualified US supply, while manufacturing delays limit how rapidly new domestic entrants can satisfy that demand. Existing productive capacity can therefore become more valuable than announced nameplate capacity. First Solar is a relevant beneficiary because its thin-film technology does not depend on the crystalline-silicon polysilicon-to-wafer production chain that Corning is ramping. Corning’s operating difficulties therefore do not imply the same manufacturing problem at First Solar. First Solar also publicly supported the August Section 232 action, confirming that management viewed the policy as favorable to its competitive position. The incremental insight is not simply that tariffs help domestic manufacturers. It is that favorable policy and strong demand do not guarantee usable supply or profitable production. Corning’s delayed recovery strengthens the distinction between established production capability and future capacity commitments. This favors proven execution rather than indiscriminate exposure to domestic solar announcements. Near-term catalyst: First Solar reports Q3 results on October 29. The most relevant indicators are new bookings, realized contractual economics, cancellation behavior, domestic production execution, and customer discussions for future delivery periods. Its previously reported 45.1 GW contracted backlog means stronger market pricing should not be applied automatically to all existing contracted sales. Longer-duration implication: The benefit is principally stronger demand for future delivery slots and potentially more resilient booking prices during 2027 and later periods. It is not an automatic repricing of the existing backlog. The counterweight is eventual competitive supply: successful Corning and other domestic ramps could reduce scarcity over time. The positive conclusion applies to current qualified capacity and future contracting economics, not to a permanent shortage or immunity from First Solar’s own execution risks. SOURCE BASIS Corning’s October 27, 2026 earnings call and accompanying results provide the principal evidence. Company disclosures used to verify exposures and distinguish new information from previously reported developments include Amphenol’s July earnings commentary; Fujikura’s high-density cable announcement; Arista’s August results and October product announcement; Lumentum’s and Coherent’s August results; Lumen’s customer relationship and investor-day disclosures; HP’s August results; Micron’s September results; Corning’s May BOE collaboration and September display-pricing announcement; and First Solar’s July results and August trade-policy statement. All read-throughs are limited to information available by October 27, 2026.

OPTICAL NETWORKING DAILY DIGEST, TUESDAY, OCTOBER 27, 2026 STATUS CHANGES • Broadcom: A $63 billion financing structure to enable AI compute was announced on October 26, 2026 by Bank of America, Blackstone, Citi and Morgan Stanley: a newly formed special purpose vehicle will purchase Broadcom XPU AI accelerators and networking solutions and lease them to Anthropic, under a previously announced strategic compute capacity agreement, with deliveries starting in calendar 2027 and Broadcom providing residual value support equal to the first lien portion. A financing and demand signal for Broadcom networking products, not a new product. • Celestica: Third quarter 2026 results, released October 26, 2026: revenue $5.83 billion, up 83 percent year over year, and adjusted EPS $3.30. Connectivity and Cloud Solutions segment revenue was $4.87 billion, up 102 percent, including Hardware Platform Solutions revenue of approximately $2.3 billion, up 70 percent; Advanced Technology Solutions segment revenue was $0.96 billion, up 23 percent. The 2026 revenue outlook was raised to $21.0 billion, and an initial 2027 outlook of $40.0 billion in revenue was given. No new product. • Corning: Third quarter 2026 results: GAAP sales $4,855 million, up 18 percent; core sales $5,081 million, up 19 percent; core EPS $0.91. Optical Communications segment net sales were $2,379 million, up 44 percent, with segment net income of $532 million, up 80 percent; Enterprise Networks sales grew 84 percent and Gen AI product sales grew more than 100 percent. Corning secured four long-term, multibillion-dollar customer agreements in the quarter, including with Verizon and AT&T, with no per-agreement value stated. No new product. • ZTE: Third quarter 2026 results, filed October 26, 2026: nine-month operating revenue RMB 116.84 billion, up 16.2 percent year over year; third quarter revenue RMB 38.81 billion, up 34.0 percent; third quarter net profit attributable to shareholders RMB 1.31 billion, up 394.5 percent. Computing accounted for 38 percent of revenue, and data center solutions revenue increased approximately 130 percent. No new product. ACQUISITION • Intel: Intel announced a definitive agreement on October 26, 2026 to acquire Classiq, a Tel Aviv quantum algorithm synthesis software company, to create a full-stack quantum computing platform. The acquisition is expected to close in the first half of 2027, subject to regulatory approvals. The release discloses no price; a value of about $2.3 billion reported by Reuters, attributed to Bloomberg News citing people familiar, was not confirmed by Intel. Quantum software, non-optical. SCREENING • Corning: A claimed October 24, 2026 GlobeNewswire release announcing a Corning and Samsung Electronics memorandum of understanding for glass-based optical interconnect and advanced packaging could not be verified from a primary source on the third day of checking: exact-title searches returned no matching release, GlobeNewswire listings show no October 24, 2026 Corning item, and Samsung sources show no such agreement. No product or design win was recorded. Re-check next run. • Arista Networks: The Arista security advisories source was unchecked this run after two failed fetches, so its advisory list was not verified today. An individual advisory opened directly carries a revision dated October 26, 2026 to a release first dated October 23, 2026, which is not a new advisory. Re-check next run. • Adtran: Adtran Networks SE published a voting rights notification on October 26, 2026: UBS Group AG crossed to a total position of 3.2 percent (3.19 percent in voting rights attached to shares plus 0.01 percent through instruments), with the threshold crossed on October 22, 2026, up from a previous notification of 2.84 percent. Non-product, no optical change. • Intel: Intel and Oracle plan an expanded collaboration, announced October 27, 2026: a multiyear plan to expand Xeon-based computing on Oracle Cloud Infrastructure, including planned offerings on the next-generation Diamond Rapids server processor, with Oracle Acceleron networking mentioned in passing. Compute collaboration, no optical product. • Lightwave Logic: A timing notice dated October 26, 2026 sets the third quarter 2026 financial results and business update call for Tuesday, November 10, 2026 at 4:30 p.m. Eastern Time. Timing notice only, no product content. • Microchip: Microchip announced program milestones for its PIC64-HPSC spaceflight computing series on October 27, 2026, including radiation characterization of the radiation-hardened variant and evaluation kits for select early access customers. Spaceflight computing, not an optical networking product. • NVIDIA: NVIDIA and Meta announced an expanded collaboration on October 27, 2026, covering the Muse family of models, a planned NVIDIA investment of $3 billion in a Meta equity offering, and Meta's stated intent to expand deployment of NVIDIA Spectrum-X Ethernet networking and BlueField-4 alongside Vera Rubin and Vera CPUs. Spectrum-X is an existing product line; no new optical or networking product was announced. • NEC Corporation: NEC announced on October 27, 2026 that it will contribute to AI powered vaccine design through the EU funded PREPAIR VAC pandemic preparedness project, through NEC OncoImmunity, with about EUR 6.5 million in Horizon Europe funding. Non-optical, no product change.

📈 US Market Pre-Market Intelligence | June 3, 2026 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ Sources: CNBC · Benzinga · StockTwits · The Motley Fool · TheStreet Pro · Reuters · Bloomberg · LSEG · FTSE Russell · · Yahoo Finance · The Globe and Mail Data window: Past 24h (priority: past 12h) ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【Market Snapshot】 • S&P 500 Futures: -0.3% | Nasdaq futures: +0.4% • VIX: 15.77 (1-year low, neutral-low) • Fear & Greed: 66 (Greed zone) • WTI Crude: ~$93/barrel (Iran tensions supporting) • Dollar Index: DXY 107.2 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【SECTION 1: TOP NEWS — TECH/AI/SEMICONDUCTOR】 🔥 Jensen Huang at GTC Taipei (June 2-3, 2026): The Single Biggest Catalyst ━━━━━━━━━━━━━━━━━━━━ Source: The Motley Fool · StockTwits · Bloomberg · Reuters · Yahoo Finance · BigGo Finance · Times of India · DataCenterNews Asia 1. MRVL +32.5% Tuesday → all-time high. Huang on stage: "Matt [Murrett] is building the next trillion-dollar company." MRVL premarket today: +22%. → Custom AI chips (non-NVDA) = the new AI alpha. Gary Black (top Tesla/tech fund manager): "Broadcom and Marvell are the big winners as focus shifts to custom AI ASICs." 2. AVGO + Broadcom: Hit 52-week high alongside MRVL. Custom AI networking/DPU chips gaining enterprise share. Gary Black: AVGO is a "big winner" in the custom chip shift. 3. HPE Hewlett Packard Enterprise: Hit 52-week high premarket. Analyst (StockTwits): "AI-driven quarter — shares deserve a higher multiple." Aruba/AI infrastructure backlog strong. 4. NVIDIA Vera Rubin platform: Full production announced at GTC Taipei today (June 3). Next-gen AI data center GPU. NVDA stock slightly lower premarket (-0.5%) as markets digest the rally in competitors MRVL/AVGO. 5. TSMC ADR hit record high +2.5%. Deepened NVIDIA partnership confirmed at GTC Taipei. Philadelphia Semiconductor Index (SOX) surged 5.9% to all-time peak — biggest single-day move in 2026. 6. IPG Photonics, MACOM, Amkor: All skyrocketed after Huang's GTC Taipei keynote. AI photonics + advanced packaging = next bottleneck. 7. Trump Executive Order: Government gets early access to advanced AI models. StockTwits called it "pro-AI infrastructure" signal. Microsoft + NVIDIA partnership: RTX Spark + Vera CPU for Windows AI laptops. Build 2026 event this week. 8. MU Micron: UBS upgraded to $1625 target — more than 100% upside. HBM memory demand = structural supercycle. ⚠️ Warning Signal: Michael Burry (The Big Short): AI chip rally is within 7% of 2000 dot-com bubble peak. Chart suggests caution on AI momentum names. Retail vs. institutional positioning diverging. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【SECTION 2: RETAIL SENTIMENT — STOCKTWITS/TRADINGVIEW/WALLSTREETBETS】 Source: StockTwits · TradingView News · TheStreet Pro · Yahoo Finance ━━━━━━━━━━━━━━━━━━━━ 📊 Most-discussed tickers by retail (past 24h): • MRVL: Trending #1# on StockTwits. Retail piling in after Huang shoutout. Bullish comments dominate. "MRVL to $200" comments everywhere. • AVGO: Custom chip theme driving retail interest. "This is the NVDA of 2026" sentiment growing. • ASTS (AST SpaceMobile): Snapped 2-day losing streak. Top execs buying shares = vote of confidence. SpaceX IPO buzz creating turbulence but long-term thesis intact. Retail mostly bullish. • RKLB / LUNR / RDW: Slipping on SpaceX IPO speculation. Bears say SpaceX could cannibalize launch demand. Bulls watching. • META: Retail sees buying opportunity. Stock still trails Mag 7 peers despite new subscriptions + layoffs + cloud plans. "META cheap vs. GOOGL/AMZN" sentiment. • TSLA: Dips on SpaceX merger rumors. Retail influencer: "Bull case adds $450B to valuation." SpaceX IPO terms uncertainty weighing. • INTC Intel: Surprising surge today. Investors asking "why is Intel surging?" AI PC + foundry turnaround narrative returning. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【SECTION 3: MARKET THEMES — AI ROTATION + SMALL CAP BREAKOUT】 Source: The Globe and Mail · 247WallSt · Benzinga · AInvest · TheStreet Pro · LSEG ━━━━━━━━━━━━━━━━━━━━ 🚀 AI Money Rotating from Mega-Cap → Small/Mid Cap: • Small caps are AI's big winners in 2026 (The Globe and Mail, Reuters, 10 hours ago) • Russell 2000 small-cap value has beaten growth by 9 percentage points YTD • Something rare is driving the Russell 2000 — and AI is the answer (Benzinga) • AI exuberance rotating into small caps amid sticky inflation (ActionForex) • IWM, VTWO, URTH all flashing strong buy signals (MarketsHost) 📊 Russell 2000 reconstitution underway (FTSE Russell, May 22 announcement): Total US equity market cap in Russell 3000 reached $75.6 trillion — 29% increase from prior year. Major index changes creating volatility + opportunities. 💰 Sector rotation: Energy (oil $93, Iran) + AI semis + small-cap value = today's leadership. Defensive sectors (utilities, consumer staples) outflowing. ⚡ Middle East: Oil supported at $90-95 range. US-Iran tensions escalating (missiles launched toward Kuwait/Bahrain per fxstreet). Energy stocks (XOM, CVX) get tailwind. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【SECTION 4: KEY STOCKS — PREMIER DATA FROM AUTHORITATIVE SOURCES】 Source: StockTwits · The Motley Fool · Yahoo Finance · Benzinga · TheStreet Pro ━━━━━━━━━━━━━━━━━━━━ 🔥 MRVL — Marvell Technology Price: ~$140+ (premarket +22%) | 52-week HIGH Catalyst: Jensen Huang at GTC Taipei: "next trillion-dollar company" Analyst: Gary Black — "big winner" in custom AI chip theme Analyst: Morgan Stanley sees $120+ target Narrative: AI custom ASIC for cloud. Google's TPUSnake, Amazon's Trainium = MRVL customers. Data center custom silicon = secular trend. Risk: Rich valuation. PS 20x. Momentum + Huang boost = near-term overbought. 📊 AVGO — Broadcom Price: ~$220+ | 52-week HIGH Catalyst: Custom AI networking chips + VMware synergy Narrative: #2# AI chip play after NVDA. Custom DPUs + networking for AI data centers. Gary Black: "big winner" alongside MRVL Risk: Valuation already pricing in strong growth 🚀 HPE — Hewlett Packard Enterprise Price: premarket +strong | 52-week HIGH Catalyst: AI-driven quarter beat. Aruba networking + GreenLake AI services. Analyst quote (StockTwits): "AI quarter — shares deserve higher multiple" Narrative: AI infrastructure + edge computing + hybrid cloud = multi-year growth Risk: Competition from Dell/Arista in AI networking 💡 TSMC — Taiwan Semiconductor ADR Price: $446.69 (+2.5% record high) Catalyst: NVIDIA partnership deepened at GTC Taipei. Advanced node capacity = structural moat. Narrative: Foundry king. AI compute demand = capacity fully booked for 2026-2027 Risk: Taiwan geopolitical risk premium always present ⚡ IPG Photonics / MACOM / Amkor Price: all skyrocketed after GTC Taipei keynote Catalyst: AI photonics (laser/optical interconnect) + advanced packaging (Amkor = chiplet packaging) Narrative: AI hardware bottleneck shifting from compute → interconnect + packaging Risk: Volatile momentum names ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【SECTION 5: SMALL-CAP US ALPHA 🔍】 Serenity Framework: Demand → Earnings → Small-Cap Elasticity → Verification ━━━━━━━━━━━━━━━━━━━━ 📌 IWM — iShares Russell 2000 ETF | Small-Cap Core Trigger: Russell 2000 hit strong buy signals; small-cap value beating growth by 9pts YTD; AI rotation into small caps confirmed (The Globe and Mail) Elasticity: Russell 2000 constituents with AI exposure = leverage to sector rotation Verification: FTSE Russell reconstitution underway; IWM options volume surging Catalyst: ADP jobs data today; Fed speakers this week; Russell reconstitution culminates end of June Risk: Rate sensitivity if jobs data hot → delay Fed cuts → small caps hurt 📌 FLNC — Fluence Energy | AI Data Center Cooling Trigger: Jensen Huang GTC keynote emphasized AI data center infrastructure at scale. Cooling/power = next bottleneck as compute density explodes. Elasticity: AI data centers need liquid cooling at scale — FLNC is leader in modular battery/cooling systems Verification: Stock up strongly post-GTC keynote. Institutional buying volume rising. Catalyst: Next earnings — watch backlog. If backlog grows >40% YoY = confirm supercycle thesis. Risk: Competition from Schneider Electric / Vertiv 📌 SOXL — Direxion Daily Semiconductor Bull 3X Trigger: Philadelphia Semiconductor Index +5.9% to all-time high. TSMC record. MRVL +32%. AI chip theme = institutional inflow. Elasticity: 3x leveraged ETF = amplified move in semiconductor sector Verification: SOXL options activity spiking. Retail interest elevated (StockTwits trending). Catalyst: Any NVIDIA/AMD/MU earnings beat = SOXL pops 5-8% same day Risk: 3x leverage = decay risk. Only for short-term tactical use. ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【SECTION 6: MACRO + MARKET STRUCTURE】 Source: Bloomberg · Reuters · fxstreet · LSEG · FTSE Russell ━━━━━━━━━━━━━━━━━━━━ 📅 Today's Key Events (June 3, 2026): • ADP Private Sector Employment (May) — consensus: +180K • ISM Services PMI (May) — flash PMI showed expansion, services follow-up • EIA Weekly Crude Inventory • Fed speakers: likely to maintain data-dependent stance • GTC Taipei keynote continues (NVIDIA ecosystem announcements) ⚠️ Key Risks: 1. Iran/Middle East escalation → oil spike >$95 → inflation risk → Fed hawkish 2. AI chip rally froth (Burry warning) → sector could see sharp 5-10% correction 3. SpaceX/OpenAI/Anthropic IPO timeline: Standard Chartered warns "market oxygen being sucked out" when these hit 4. Tariff fatigue: Section 301 tariffs on 60 economies (forced labor) could hit supply chains 5. MU/TSMC: Any supply disruption from Taiwan Strait tension = semiconductor sector crash ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【SECTION 7: INSTITUTIONAL FLOWS】 Source: LSEG · FTSE Russell · Bloomberg ━━━━━━━━━━━━━━━━━━━━ • AI/Infrastructure: MRVL, AVGO, HPE, TSM — HEAVY institutional accumulation • Energy: XOM, CVX — inflows on oil geopolitics premium • Small-cap value: IWM, VTWO — first real institutional rotation signal of 2026 • Outflows: Utilities (XLU), Consumer Staples (XLP), REITs — rate sensitivity • Crypto/Fintech: COIN, SQ — mixed; Bitcoin holding $95K support ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 【INVESTMENT SUMMARY — 3 Scenarios】 🔵 Bull Case: AI custom chip theme continues → MRVL/AVGO/TSM lead → small caps catch fire → Russell 2000 breaks out → IWM $250+ Catalyst: MU earnings beat, AMD data center beat, ADP jobs soft (Fed cuts priced in) 🔴 Bear Case: Burry warning correct → AI chip rally peaks → MRVL/AVGO reverse → Nasdaq -3% correction Catalyst: Strong ADP + hot inflation → rate cut timeline pushed out → small caps get crushed 🟡 Base Case: AI infrastructure secular bull → semis consolidate at high level → small caps rotate in/out in tranches → VIX stays 15-18 Catalyst: No major catalyst → range-bound S&P 500 with sector divergence ━━━━━━━━━━━━━━━━━━━━━━━━━━━━ ⚠️ Disclaimer: This content is for informational purposes only and does not constitute investment advice. Data sourced from public English-language financial media (CNBC, Bloomberg, Reuters, StockTwits, Benzinga, The Motley Fool, TheStreet, LSEG). Historical performance does not guarantee future results. Consult a licensed financial advisor before investing.

👀 #MediaTek# confirms its AI ASIC collaboration with #Intel’s# EMIB-T advanced packaging. The chipmaker is now pushing 400G SerDes IP to target next-gen ASIC opportunities, including reportedly #Google’s# TPU v10 and Meta’s custom AI chips.💡More: 🔗

NVIDIA JUST CONFIRMED ITS NEXT-GEN CHIPS ARE ON SCHEDULE, AND MADE A DIRECT COMPARISON TO AMD IN THE SAME BREATH Nvidia $NVDA says its chips remain on track for use in AI data centers. The company also says its new Vera processor is faster than AMD's Turin. Nvidia said a major customers are already testing Vera Rubin hardware.