Skip to content

Infineon Technologies OktoberTech 2026: Michio Kaku on AI innovation, edge AI, SDVs and 800 VDC data center power

Infineon Technologies held its 10th annual OktoberTech event in Mountain View, California, on October 22, featuring briefings on physical AI, automotive technology and power systems. Power and Sensor Systems division president Adam White said the first 800 V DC products, whether prototypes

Share:
Infineon Technologies OktoberTech 2026: Michio Kaku on AI innovation, edge AI, SDVs and 800 VDC data center power
Photo: www.indexbox.io

Infineon Technologies held its 10th annual OktoberTech event on October 22 at the Computer History Museum in Mountain View, California. Alongside exploring the museum's exhibits, attendees could participate in a series of executive briefings devoted primarily to physical AI, including sessions on robotics, automotive technology and power systems. Between the briefings and keynote speeches, Michio Kaku, a theoretical physicist, professor at the City College of New York and bestselling author, discussed waves of innovation. The concept is closely associated with economist Joseph Schumpeter, who proposed the theory of creative destruction, in which new inventions make earlier ones obsolete. Kaku traced these waves from steam power in the Industrial Revolution through transistors and the information age to the current AI revolution, considering how other advances, including quantum computing and brain-computer interfaces, could create entirely new industries and reshape everyday life.

OktoberTech covered automotive semiconductors, power electronics and embedded processing, with discussions highlighting the growing demands AI infrastructure places on semiconductor technology and manufacturing capacity.

Lessons from IoT and the case for edge AI

Asked about this, Stephan Zizala, president of Infineon Technologies' Edge Systems division, described IoT not as a failure but as an unfulfilled promise. The change, Zizala argued, is that microcontrollers with greater performance and more memory can now run AI models locally, removing the need to send all information back to the cloud. AI processing takes place at the edge, with the results sent to the cloud, a shift Zizala called conceptual.

Zizala said this transition is already taking place in everyday products such as power tools and home appliances. During the session, Zizala noted that machine-learning algorithms are now being applied to motor control for predictive maintenance, citing cloud token costs, limited connectivity, latency and data protection as additional reasons to keep processing local.

Software-defined vehicles

Schiffer pointed to another setback: Western automakers had tied SDV launches to new electric vehicle platforms, so when those platforms were canceled, the SDV launches were also delayed. In China, by contrast, electric vehicle penetration has reached approximately 50%, giving Chinese automakers a competitive advantage. Schiffer concluded that established automakers have no choice but to control their own software if they are to manage increasing complexity, requiring them to consolidate approximately 100 distributed control boxes, each with dedicated software, into a central computer. Schiffer said this was the only way traditional automakers could survive, and that they now recognize this and need to catch up.

Data center power

These include the first 900 V GaN bidirectional switch intended for Enphase's IQ solid-state transformer, and 800 V-to-50 V, 12 V and 6 V GaN intermediate bus converters developed in cooperation with NVIDIA. These investments in data center power also extend to manufacturing. In July, Infineon Technologies opened a €5 billion, approximately $5.58 billion, Smart Power Fab in Dresden, a 300 mm facility for power and analog/mixed-signal devices and the largest single investment in the company's history. Adam White, president of Infineon Technologies' Power and Sensor Systems division, presented the facility as capacity for AI deployment, saying it was currently empty but pledging that demand would drive its ramp-up and adding that it was open to the industry. Volume production is scheduled to ramp up at the end of 2026.

Asked when 800 V DC would reach the market, White said the first products, whether prototypes or mass-produced units, would definitely appear in 2028. White described this as the eventual configuration in a longer progression from single-phase to three-phase power and, ultimately, from high-voltage alternating current to medium-voltage direct current at the facility level, targeting 1 MW racks. This differs from today's sidecar approach, which generates 800 V locally from three-phase alternating current. White also warned that substantial barriers to entry remain because the industry must maintain UL certification and safety from the outset.

For IBCs, Infineon Technologies launched 800 V-to-50 V and 800 V-to-12 V IBCs in March 2026, both using 650 V CoolGaN on the primary side. The October announcement strengthens GaN's role in IBCs, an approach also pursued by other power companies, and confirms that Infineon Technologies is actively developing an IBC for 800 V-to-6 V conversion. These receive 800 VDC from the sidecar and convert it closer to the core, potentially allowing converters to convert 6 V directly to core voltages below 1 V.

The physical relationship is clear: as voltage falls and current rises, I2R losses become dominant. The 20 kW, 800 V-to-6 V converter demonstrated at GTC 2026 delivers more than 3,000 A at its output. The closer the voltage gets to core voltage, the shorter the electrical path to the xPU needs to be, creating a need for vertical power delivery to AI processors, namely GPUs.

Discussing power delivery to different xPUs, one panelist noted that lateral power delivery would continue to be used in CPU-heavy agentic racks. While GPUs remain the primary processors for training at a GPU-to-CPU ratio of 8:1, inference at approximately 4:1 and agentic workloads at approximately 1:1 shift the balance further toward CPUs. White said interactions with customers showed that some large companies had a clear intention to move down to a 1:1 ratio, a dramatic change from the previous 8:1 ratio.

This report provides an in-depth analysis of the global semiconductor market, covering market size, growth, demand, supply, trade flows, prices, the competitive landscape and the outlook through 2035.

The study is designed for manufacturers, distributors, importers, exporters, investors, procurement teams, advisers and strategy teams that need a consistent, data-driven view of market dynamics and a transparent analytical definition of product coverage.

Product coverage

The scope covers semiconductors, encompassing solid-state electronic devices manufactured from semiconductor materials to switch, amplify, emit light or convert electrical signals. It includes discrete devices, integrated circuits, optoelectronic devices and power semiconductor devices across silicon and wide-bandgap materials. It covers finished semiconductor devices supplied to electronics manufacturing.

Inclusions

Exclusions

Report coverage and analytical modules

This report combines a foundation of standard market statistics with strategic chapters useful for commercial planning, procurement decisions, market entry, competitor monitoring and portfolio prioritization.

Segmentation framework

The market is segmented into decision-relevant categories so that demand drivers, pricing logic, supply constraints and competitive positions can be compared within the same analytical framework.

Classification coverage

Statistical calculations use HS codes 854110, 854121, 854129, 854130, 854140, 854150, 854160, 854190. However, the analysis follows the product definition above, and other goods within the same tariff headings are not automatically included.

HS codes (framework)

Geographic coverage

The research covers global totals and individual profiles for all 50 countries and territories covered by the report: the United States, China, Japan, Germany, the United Kingdom, France, Brazil, Italy, Russia, India, Canada, Australia, South Korea, Spain, Mexico, Indonesia, the Netherlands, Turkey, Saudi Arabia, Switzerland, Sweden, Nigeria, Poland, Belgium, Argentina, Norway, Austria, Thailand, the United Arab Emirates, Colombia, Denmark, South Africa, Malaysia, Israel, Singapore, Egypt, the Philippines, Finland, Chile, Ireland, Pakistan, Greece, Portugal, Kazakhstan, Algeria, Czechia, Qatar, Peru, Romania and Vietnam. This is the same country list shown in the table of contents.

Data coverage

Units of measurement

Methodology

This report combines official statistics, trade records, corporate disclosures, product-level evidence and analyst validation. Data are standardized, reconciled and cross-checked to ensure that market size, trade flows, prices and forecasts are comparable across countries and periods.

All indicators are mapped to a consistent product definition and reviewed according to the segmentation framework used in the table of contents.

1. Introduction

Report scope and analytical framework

2. Executive summary

A concise overview of market direction

3. Market size and development trajectory

Market size, growth and scenario framework

4. Category scope, definitions and boundaries

Commercial and technical scope

5. Category structure, segmentation and product matrix

How the market is divided into decision-relevant segments

6. Demand, customers and consumer structure

Sources of demand and demand behavior

7. Production, supply and value chain

Supply coverage, trade and value capture

8. Trade, procurement and import dependence

Trade flows and external dependence

9. Pricing, promotions and commercial models

Price formation and revenue logic

10. Competitive landscape and portfolio capabilities

Which companies succeed and why

11. Geographic landscape and country roles

Where growth and supply are concentrated

12. Growth strategies and market entry

Commercial entry and expansion priorities

13. Where to enter next: the most attractive growth opportunities

Areas with the strongest rationale for expansion

14. Key company profiles

Key companies and strategic types

15. Country profiles

Detailed analysis of major national markets

16. Methodology, sources and disclosures

How the report was prepared


Original Korean article: indexbox.io

English brief by Korea Wind Intelligence, koreawindintel.com

Found this useful?

Share this article with your network.

Tags: