Semiconductor
Resource hub

The rapid growth of AI has created a surge in the global energy consumption at a rate never seen before. Today, data centers account for approximately 415 terawatt-hours (TWh) of electricity globally. To put this into perspective, the annual energy consumption of the United Kingdom in 2023 measured at 309 TWh. The International Energy Agency (IEA) projects data centers’ energy consumption will more than double to nearly 945 TWh by 2030 [1]. A single generative AI query can consume up to ten times the power of a traditional search [1]. Meanwhile, data center energy usage in the U.S. is projected to leap from 4.4% to as much as 12% of the national grid by 2028 [2]. This creates a stark reality for the semiconductor industry. Traditional monolithic scaling has hit its physical and economic limits, leaving advanced packaging and heterogeneous integration to define the industry’s trajectory [3].To meet these escalating compute demands, the industry is rapidly shifting toward multi-die architectures, chiplets, and 3D stacking to decrease the amount of energy needed for advanced computing. This transition is fueling explosive growth in the advanced packaging market, which the Yole Group projects will reach $79.4 billion by 2030 [4]. However, stacking chiplets to bypass Moore’s Law exposes massive systemic bottlenecks. Engineers are now fighting interconnect parasitics, navigating complex power delivery architectures, and battling extreme thermal density.In a 3D-stacked architecture, pulling heat away from vertically integrated dies is one of the most pressing engineering challenges of our time. As compute density rises, issues like die warpage and localized thermal hotspots threaten both reliability and yield. The shift toward sustainable AI systems for energy-efficient computing requires breakthroughs in everything from hybrid bonding process flows to advanced thermal interface material (TIM) strategies and liquid cooling integration [6].These are not challenges that any single company can solve in isolation. Whether you are a foundry, OSAT, material supplier, or equipment provider, overcoming these bottlenecks requires pre-competitive, industry-wide collaboration. Foundational capabilities must be built collectively before competitive differentiation occurs.This is the core mission of the SEMI Advanced Packaging and Heterogeneous Integration (APHI) Technology Coalition. By collaborating on common standards, shared research frameworks, cross-vendor interoperability models, and collective technology roadmap congruency, APHI is actively dismantling the barriers to next-generation computing.The APHI community is already tackling these issues head-on. Monthly chapter meetings identify and address these and other issues facing heterogeneous integration. The most recent chapter meetings showcased in depth review of these challenges. Jonathan Abdilla from BESI detailed the technical challenges and collaborative research required for global hybrid bonding process flows. Similarly, Dr. Jie Geng from Indium Corporation led a deep dive into crucial TIM strategies for AI and HPC, exploring hybrid stacking evaluation methods and liquid cooling options to combat GPU die warpage.The future of advanced manufacturing will be defined by how effectively we manage power and heat in heterogeneous systems. We invite you to join this critical conversation at the upcoming SEMIEXPO Heartland (April 29-30 in Detroit, MI) Day 2 will feature dedicated sessions on Thermal Management Power Delivery in Advanced Packaging: From TIMs to Warpage Control, as well as strategies for securing the advanced packaging supply chain.To help shape the standards and shared roadmaps that will power the AI revolution, explore our initiatives and get involved with SEMI Advanced Packaging and Heterogeneous Integration (APHI) Technology Coalition.Rafael Tudela is Senior Technical Marketing Manager at SEMI References[1] International Energy Agency (IEA). (2024). Energy and AI Report. [2] U.S. Department of Energy (DOE) Lawrence Berkeley National Laboratory (LBNL). (2024). Report on U.S. Data Center Electricity Demand and Grid Impact.[3] Semiconductor Packaging News. Advanced Packaging and Heterogeneous Integration. Retrieved from: https://www.semiconductorpackagingnews.com/articles/92402.html [4] Yole Group. (2025). Status of the Advanced Packaging Industry 2025.

The semiconductor industry is hitting a structural inflection point: explosive AI‑driven demand, rapidly rising manufacturing complexity, and stringent sustainability expectations are converging at once. In this context, edge AI deployed directly on tools, sensors, and local controllers, is shifting from experimental to essential, particularly in fabs where milliseconds matter. SEMI’s timely workshop, Smarter Sensors, Smarter Fabs: AI at the Edge in Semiconductor Manufacturing, taking place March 18–19, 2026 in Milpitas, CA, will address this important topic.From sparse sensing to dense instrumentationTwo decades ago, most process tools relied on dozens of sensors per chamber. Today, leading etch, deposition, CMP, and lithography systems routinely integrate hundreds of sensing channels spanning pressure, flow, RF power, optical endpoint, vibration, and chemistry. At 3 nm and 2 nm, process windows are so tight that yield hinges on multivariate understanding of chamber conditions and tool state rather than a few independent alarms. Sensor proliferation has turned fabs into rich data environments—but also exposed the limits of traditional, centrally managed control.Why edge AI is displacing cloud‑only controlConventional architectures push heavy analytics to centralized servers or the cloud, with supervisory systems periodically updating recipes, setpoints, or dispatch rules. Across manufacturing, measured cloud round‑trip times commonly range from 800 to 2,400 ms, whereas edge systems co-located with equipment can respond in 15–45 ms, roughly 50–160× faster. For safety‑ and yield‑critical loops in semiconductor manufacturing, that latency gap is often unacceptable.At the same time, new generations of low‑power neural processing units (NPUs) and edge accelerators deliver tens of trillions of operations per second (TOPS) at single‑digit watt budgets, making always‑on inference viable inside tools, cameras, and controllers. The result is a decisive move toward edge‑native architectures: models execute where data is produced, while cloud resources are reserved for retraining and fleet‑wide learning.Edge AI on the line: control, inspection, and maintenanceIn process control, edge AI is enabling a shift from univariate threshold checks to multivariate models that understand the joint dynamics of sensor streams. Platforms today embed deep‑learning and statistical models directly at or near the tool, performing real‑time endpoint prediction and anomaly detection from high‑dimensional time series. Similar approaches are emerging in lithography and CMP, where local inference helps keep focus, overlay, and removal rate within spec before wafers drift out of control.Inspection and logistics are undergoing a similar transformation. Vision systems with embedded NPUs classify defects at line speed, often above 100 parts per minute, eliminating the need to ship large image volumes to a central cluster. Robots and autonomous mobile robots (AMRs) use local intelligence for short‑horizon planning and collision avoidance, while higher‑level systems focus on global scheduling and optimization.Predictive maintenance is one of the most mature applications: vibration, acoustic, temperature, and pressure data are analyzed locally to detect anomaly signatures hours or days before conventional thresholds trip. Reported benefits include reductions in unplanned downtime, longer component life, and lower maintenance costs when these models are integrated into manufacturing execution systems (MES) and maintenance workflows.Digital twins and agentic AI on top of edge dataDigital twins build on this sensing and edge‑analytics foundation. By maintaining virtual, live‑updated models of tools, lines, and entire fabs, they enable scenario testing, debottlenecking, and root‑cause analysis without putting WIP at risk. Vendors and early adopters report that such twins can shorten process‑node ramps and facility bring‑up by enabling thousands of “what‑if” experiments before physical changes are made.​Agentic AI is now emerging as the orchestration layer above these twins. In semiconductor case studies, agents connected to MES, advanced process control (APC), and planning systems have delivered double‑digit improvements in throughput, cycle time, and tool utilization by autonomously adjusting routing, batch sizes, and scheduling in response to live fab conditions. Other agents mine unstructured engineering notes and fault reports to accelerate root‑cause analysis, turning hard‑won lessons into repeatable, codified behavior.Sustainability as a first‑class requirementSustainability pressures are reinforcing this stack. Semiconductor manufacturing is energy‑ and resource‑intensive, and regulators and customers alike are demanding more transparency and improvement. Edge‑connected monitoring of energy, utilities, and emissions has already helped some fabs cut energy‑related costs by around 20 percent through tighter control of HVAC, process gases, and idle modes. Research initiatives such as imec’s Sustainable Semiconductor Technologies and Systems (SSTS) program are using virtual fab methods and detailed life‑cycle assessment to guide process and equipment choices for lower environmental impact.Strategic takeaways and where to learn moreThe trajectory is clear: fabs that combine dense sensing, edge AI, digital twins, and agentic AI are building toward continuously learning, self‑optimizing operations. Architectures will need to be edge‑first rather than cloud‑only. Simply adding sensors without local intelligence will not deliver competitive advantage, and environmental KPIs are likely to be optimized with the same rigor as yield and cycle time.For practitioners who want to translate these trends into roadmaps, the Smarter Sensors, Smarter Fabs: AI at the Edge in Semiconductor Manufacturing” workshop (March 18–19, 2026, Milpitas, CA) spearheaded by the SEMI Manufacturing Coalitions* will bring together experts in sensing, edge architectures, digital twins, and agentic AI to share concrete deployments and architectures tailored to semiconductor fabs.*The SEMI Manufacturing Coalitions include Smart Manufacturing, Fab Owners Alliance (FOA) MEMS and Sensors Industry Group (MSIG), Advanced Packaging Heterogenous Integration (APHI) and Semiconductor Components, Instruments, and Subsystems (SCIS). Anshu Bahadur is Senior Program Manager, Technology Communities at SEMI. Mark da Silva is Senior Director, Manufacturing Coalitions at SEMI.

At SEMICON West, Dr. Bertrand Cambou of High Entropy Security explored a fast-evolving challenge in the semiconductor ecosystem: the rise of counterfeit and cloned devices. His session introduced a technical and scientifically grounded approach to solving this problem—device fingerprinting, the use of intrinsic manufacturing variation to create unclonable identities for each chip.Cambou’s presentation illustrated how semiconductor fingerprinting combines physics, manufacturing realities, and AI-driven validation to establish authenticity in ways traditional methods cannot match.Shape1. Every Semiconductor Device Contains a Unique Physical SignatureCambou began by explaining a foundational property of semiconductor manufacturing: no two devices are perfectly identical. Microscopic variations in materials, lithography, doping, and process steps produce natural randomness.These variations form a unique, unclonable physical fingerprint. This fingerprint can be:MeasuredCharacterizedBound to device identityUsed for authenticationThis approach is resilient because it does not require adding new circuits or security modules—the silicon itself becomes the security feature. 2. Counterfeit Devices Pose a Growing ThreatCambou outlined several real-world scenarios where counterfeit or cloned components cause harm:Fake secure elements entering financial systemsCounterfeit microcontrollers used in identity solutionsUnauthorized clones undermining product reliabilityTrojan-modified chips entering critical systemsTraditional security measures often fall short because counterfeiters now replicate packaging, serial numbers, and superficial logic behavior. Physical fingerprinting offers an inherently more robust defense. 3. The Limitations of PUFs and the Challenge of CloningPhysical Unclonable Functions (PUFs) have long been proposed as a way to leverage hardware uniqueness. But Cambou noted that widely used PUFs—such as SRAM-based designs—can sometimes be replicated or approximated through characterization attacks.Clone resistance requires more than static measurements; it requires validation that the device is alive, not merely mimicking expected outputs. 4. AI “Liveness Tests” Strengthen Fingerprinting SystemsCambou introduced AI-based liveness detection as a second layer of defense. When a real device is measured repeatedly over time, natural noise and drift appear in the fingerprint. AI models can distinguish this authentic variability from the behavior of a simulated or cloned device.This approach makes fingerprinting:More robustHarder to spoofMore reliable under different environmental conditionsBetter aligned with real-world deployment needs 5. Fingerprinting as a Tool for Supply Chain IntegrityFingerprinting supports broader supply chain goals—many of which were highlighted throughout SEMICON West. These include:Ensuring genuine components enter critical systemsVerifying device originDetecting unauthorized substitutionsStrengthening counter-counterfeiting effortsEnhancing lifecycle trackingCombined with provenance and traceability, fingerprinting helps create a more trustworthy global semiconductor ecosystem. 6. Physics-Based Trust Will Grow in ImportanceCambou’s closing insight was that as the semiconductor ecosystem becomes more distributed and threats become more sophisticated, trust anchored in physical reality will be increasingly valuable. Fingerprinting creates a path toward that trust—one rooted not in documentation or assumptions, but in the immutable characteristics of silicon itself. Source: “Secure Together: Building Cybersecurity Resilience Through Industry Alliances,” SEMICON West 2025. Speakers: James Kaplan (McKinsey Company); Quentin Kantaris (TXOne Networks); Bradford Hegrat (Accenture); Nijaz Velic and Richard Morris (NY CREATES); Tom Palmaers and Giselle M.H. Van Tornout (imec); SZ Lin (Sun Square); Ross Mahler and Marty Wachi (Moxa); Simon Davies (Renesas); Jennifer Lynn (IBM); Prabhu Jayanna (AMD); Anusha Annapareddy (Applied Materials); Bertrand F. Cambou (High Entropy Security); Daniel O'Loughlin (Qualcomm). Panel moderator: Andrew M. Seward (Tokyo Electron America).

The semiconductor industry is the bedrock of modern technology, enabling everything from AI and cloud computing to electric vehicles. Yet, this critical sector is also one of the most resource-intensive globally, with a substantial dependency on water. A single fabrication plant can demand up to 10 million gallons of water daily, comparable to the consumption of a city with 300,000 residents. Much of this water is, of course, reused and recycled through sophisticated systems. This immense water usage, particularly the requirement for ultrapure water for processes like cleaning and etching, makes consistent access to high-quality water a non-negotiable for operational reliability and business continuity. The new insights report "Ripple Effects: Water Risk and Resilience Across the Semiconductor Value Chain" provides the first global baseline of water risk hotspots for the semiconductor sector, assessing water risks across 140 facilities across 89 water basins to inform future risk mitigation strategies.The analysis discusses how water risk can manifest itself as a financially material impact on business continuity by triggering idle time, recovery costs, and cascading delivery delays across global supply chains. S P Global projects that by 2050, water-related risks could cost the world's largest IT companies up to $24 billion annually. Crucially, the study identified flooding and reputational risks—such as strained relationships with local communities over water allocation—as the most significant immediate threats to the semiconductor value chain. These concerns are most acute in major hubs like Taiwan, South Korea, and parts of the U.S.While the industry is frequently criticized for its water usage, only 16% of the analyzed sites are currently affected by water scarcity. However, this metric offers a false sense of security. As climate change intensifies, the frequency and severity of water-related disruptions are set to exceed the scope of existing contingency plans. The long-term projections show that over 40% of semiconductor facilities announced since 2021 are located in watersheds projected to face high or extremely high water stress between 2030 and 2040. This underscores the urgent need to integrate forward-looking risk modeling into new site planning to ensure long-term operational resilience.Effective risk management is significantly hindered by the limited transparency surrounding supplier-level water data. While many companies perform water assessments for their direct operations, a comprehensive, industry-wide approach to supplier data and risk management is lacking. CDP data shows that 1 in 5 companies reported $77 billion under threat from supply chain water risks, yet only half of those companies engage with their suppliers on these issues. For semiconductor end users, these risks are often deep within multi-tiered networks, requiring engagement that goes well beyond Tier 1 suppliers.To manage these complex risks, the report stresses the necessity of moving toward a contextual approach that includes localized assessments. Contextual water risks are inherently location-specific, dependent on local availability, quality, and infrastructure, as well as broader catchment-level dynamics, regulatory pressures, and community expectations. Several structured methodologies support this necessary shift from basic operational management to corporate water stewardship, including the Alliance for Water Stewardship (AWS) Standard, the TNFD's LEAP framework, and the Science Based Targets for Nature (SBTN). This approach encourages companies to look beyond their own operations to safeguard regional water security.Because water is a shared resource, collective action is essential to deliver the scale and urgency needed to tackle common challenges within catchments. The semiconductor value chain is deeply interconnected, with companies often sharing suppliers within the same water basins, creating a strategic opportunity for collaborative stewardship. The report encourages companies to scale their impact by moving beyond isolated efforts to form sector-wide and cross-sector partnerships—especially at the catchment level—through public-private engagement. This collaboration, which includes proactive engagement with policymakers and local utilities, is key to aligning on water management and stewardship practices to address shared water challenges and build collective trust.Innovation and technology must play a central role in advancing water stewardship across the value chain. A major hurdle is the general undervaluation and mispricing of water, which perpetuates systemic underinvestment in water-focused technology. Despite this, leading semiconductor companies are deploying advanced solutions such as onsite recycling systems, real-time water monitoring, and utilizing alternative sources like municipal wastewater. Embracing AI-driven systems for scenario modeling and catchment-level risk forecasting further enhances adaptive capacity and resilience.The "Ripple Effects" report makes it clear that water challenges affect every segment, demanding tailored response tactics and strategies. Foundries, with their large operational footprints, must prioritize sourcing reclaimed water and expanding onsite reuse, while chemical and materials suppliers must proactively manage rising regulatory risks around water quality contaminants. The insights report also provides a practical roadmap for advancing corporate water stewardship, outlining progression from water risk assessment (Stage 1) to site-level action and collective engagement (Stage 2), and culminating in transparent validation and reporting (Stage 3). By following a structured water stewardship pathway, the semiconductor industry can build operational resilience and ensure a responsible future for the entire value chain.To learn more, download the report or watch the webinar recording. Alua Suleimenova is Senior Program and Staff Manager | Global Sustainability at Marvell Technology and leader of SEMI's ERMR Working Group.The Environmental Risk Mitigation and Reporting (ERMR) Working Group was established under SEMI's Sustainability Initiatives in January 2023, and it aims to develop a baseline and roadmap of best practices for identifying, managing, governing and reporting climate, water, and biodiversity risks across the semiconductor value chain. This insights report is a publication in SEMI’s ERMR Working Group thought‑leadership series on global environmental risks and resilience.

At SEMICON West, the SEMI MEMS and Sensors Industry Group (MSIG) showcased how advances in sensor miniaturization, integration, and edge intelligence are reshaping mobility, automation, and the environment. From synthetic perception to infrared AI and sensor-driven autonomy, the TechTALKS session revealed a clear theme: sensors aren’t just collecting data — they’re defining how the world will see and act in real time. Three Sensor Trends to WatchIntelligence at the EdgeSTMicroelectronics’ dual-accelerometer IMU illustrated how in-sensor compute delivers context-aware functionality — recognizing events like impacts, motion, or orientation without taxing system-level power or latency. Embedded finite-state machines and ML-based logic make smart sensors the first step in real-time decision loops.Synthetic Perception and New MarketsSilex Microsystems framed MEMS as the enabler of “synthetic perception” for AI systems that must see, hear, and sense their environments. Emerging applications span wearables, AR/VR, quantum photonics, and telecom infrastructure — markets accelerated by new 12-inch manufacturing capacity and heterogeneous integration of optical and mechanical functions.From Autonomy to AwarenessWaymo and Boston Dynamics demonstrated how sensor fusion and AI extend autonomy from vehicles to robots. Waymo’s foundation model for perception integrates LiDAR, RADAR, and vision to navigate dense urban domains; Boston Dynamics’ Spot records thermal, acoustic, and visual data for predictive maintenance and facility analytics. Both examples show how continuous sensor streams fuel safer, smarter operations.Why It MattersSensors define how autonomous and AI systems perceive reality.Integration of compute and context at the edge reduces latency and power.Manufacturing capacity and standardization are critical to scaling innovation from lab to market.Cross-sector applications — from mobility to sustainability — depend on the sensor ecosystem that MSIG connects.What to WatchSilex announced a new 8” MEMS Foundry Fab that will be located in the US starting in 2026. Further expansion of their MEMS foundry capacity is projected to be a 12-inch fab ramp in 2028–2029.Growth in context-aware IMUs and low-power AI sensor nodes.Advances in infrared and spectral sensing for sustainability use cases. Source: TechTALKS: Sensing Tomorrow: Emerging Applications for MEMS Sensor Technologies, SEMICON West 2025. Moderator: Paul Carey, PhD (SEMI). Speakers: Tim Brosnihan (Silex Microsystems); Matteo Fusi (STMicroelectronics); Venkataraman Chandrasekaran (Meta); James (Jae-Hyung) Lee (Stratio); Yeh-Jiunn Tung (Waymo); John Weiler (Boston Dynamics). Learn More about SEMI's MEMS Sensors (MSIG) Community ›

At SEMICON West, industry leaders agreed: Scalable, secure data platforms are now essential for AI-driven semiconductor manufacturing. Federated data sharing, traceability, and edge-to-cloud integration form the digital twin backbone powering intelligent fabs.If you work anywhere near semiconductor manufacturing, you’ve heard the refrain: AI will optimize everything—yield, uptime, energy, supply chain. The catch? None of it works without a data backbone strong enough to feed, connect, and secure those models across thousands of tools and workflows.That’s why at SEMICON, leaders from Idaho National Laboratory, Athinia, PDF Solutions, AWS, and Multiscale Technologies aligned on one message: scalable data platforms are no longer optional—they’re the digital twin backbone that turns isolated models into intelligent manufacturing systems. From Digital Threads to a National FabricRoss Kunz (Idaho National Laboratory) kicked off with the SMART USA Initiative, which is building the connective tissue between digital twins across industry, academia, and government. His “digital backbone” vision centers on federated data management, standard semantics, and open APIs—so a process twin built in one fab can plug seamlessly into another.The payoff: a shared innovation fabric where validated models, datasets, and best practices circulate as national infrastructure, not one-off experiments. Structured Data, Trusted CollaborationFor Athinia’s Dr. Adam Schafer, the biggest barrier to scale isn’t algorithmic—it’s trust. He showed how Athinia’s multi-party SaaS platform allows device makers, material suppliers, and equipment vendors to share process data securely without exposing IP.Its 1-2-3 framework (Ingest → Transform → Analyze) normalizes messy fab data into structured, reusable datasets that power root-cause analysis and machine-learning pipelines.Obfuscation, version control, and collaborative workspaces make the data both governed and actionable—the foundation every digital twin depends on. Smart Manufacturing’s Beating HeartPDF Solutions’ Jonathan Holt described digital twins as “purpose-driven virtual representations” that enable forecasting, optimization, and control. His breakdown—component, process, and enterprise twins—anchors AI within the physical reality of manufacturing. Holt emphasized the next frontier: traceability. As fabs grow more distributed, maintaining a single, trusted record of every material and process is what will separate reactive operations from predictive ones. Scaling from Edge to CloudAWS leaders Dhara Vaishnav, Gautham Unni, and Felix David framed digital twins as living systems that thrive only when data flows freely from edge to cloud. Their AWS Digital Twin Framework combines IoT TwinMaker, SageMaker, and a spatial data lake to orchestrate real-time telemetry, simulation, and AI across global fabs.Security and scalability are baked in: multi-tenant isolation, federated governance, and GenAI-ready APIs that allow twins to learn continuously while keeping intellectual property protected. AI as the Optimization EngineSurya Kalidindi (Multiscale Technologies) closed with a look at AI-driven multiscale twins that merge physics-based models, silicon data, and expert knowledge.His Bayesian frameworks cut silicon trials in half by replacing expensive wafer experiments with calibrated simulations. A semantic layer translates complex fab data into process-aware queries—so engineers can ask questions in their own language and get trustworthy, AI-assisted answers in seconds. Why It MattersFederated data fabrics transform digital twins from silos into shared national assets.Secure, structured, and versioned data makes AI trustworthy at scale.Multiscale twins connect design, fab, and test for faster yield learning.Standards and semantics turn digital twins into the backbone of manufacturing resilience. What to WatchSMART USA interoperability pilots across multi-institution fabs.Athinia trust-layer adoption for supplier–fab collaboration.AWS TwinMaker and SageMaker integrations with fab MES systems.AI agents and Bayesian twins accelerating node development.Source: Building the Digital Twin Backbone: Why Scalable Data Platforms Are No Longer Optional, SEMICON West 2025. Moderator: Anshu Bahadur (SEMI). Speakers: Ross Kunz (Idaho National Laboratory); Adam Schafer, PhD (Athinia); Jonathan M. Holt (PDF Solutions); Dhara Vaishnav, Gautham Unni, and Felix David (AWS); Surya R. Kalidindi (Multiscale Technologies)Learn More about SEMI's Smart Manufacturing Community ›