Peter Koerte

They contrast heavy US investment with lower European investment, then argue that industrial AI needs something money cannot buy directly: operational history.

• Sharing customer data is key, they say, and requires trust earned over decades.
• This challenges compute-first narratives about AI leadership.

In a post summarizing an interview with Fortune, Koerte cites $286 billion of US AI investment in the previous year against $21 billion in Europe, notes that Siemens is investing more than €1 billion in industrial AI over three years, and names customer data and trust as the binding factors.

Read original

Compute can be ordered quickly, while the electrical and physical infrastructure that keeps a data center running takes longer to build and can determine whether projects open on time. They also point to regional manufacturing capacity for the electrical equipment data centers depend on.

At a Siemens switchgear factory in Shanghai, Koerte asked the team how they knew the data center wave was coming.

• They had moved resources two years before the orders arrived.
• They cite China moving from 30 percent to 35 percent electrification by 2030, while the US sits around 22 percent and Europe around 21 percent.
• Grids, switchgear, cooling and buildings are the slower parts, they say, and note two new US plants worth more than $200 million in Georgia and Texas making electrical equipment for AI data centers.

Read original · Original 2

Sovereign compute needs power you control, they say, while a grid absorbing large new demand needs AI to run it.

• Countries should stay globally connected while developing their own strategies for energy, digital and critical infrastructure.
• Koerte calls full technological sovereignty an illusion and says the path forward is collaboration, not isolation: the argument is about managed interdependence, not self-sufficiency.

Koerte relays a conversation with Jake Sullivan in support of a connected-but-sovereign position.

• In another post, they write that full sovereignty is an illusion because supply chains are global by design.
• They tie the sovereignty argument to two new US plants in Georgia and Texas and to the claim that countries sorting this out early will be able to host infrastructure others are announcing.

Read original · Original 2

Before-and-after numbers and named people make automation concrete and retellable. They also position these changes alongside the medium-voltage products the same teams build, linking factory-floor gains to the infrastructure that keeps data centers and factories powered.

At the Siemens Shanghai factory, Koerte writes that workers used to walk nearly 30,000 steps a day collecting parts; now robots bring the parts and pick their own routes.

• An AI check takes five minutes instead of thirty.
• They thank LiJun Dai and the Shanghai team and note that their medium-voltage switchgear keeps data centers and factories powered.

Read original

Tangible objects, named individuals and hands-on stories make relays, switchgear and engineering careers interesting rather than abstract. They use the same approach for internal recognition, introducing award finalists through the technical problem each team solved.

In Amberg, apprentices built a miniature industrial world, including a tiny 3D-printed version of Munich's Allianz Arena, powered and controlled entirely by the relays, safety devices, contactors and power supplies they manufacture.

• They built it on the factory floor, where colleagues stopped by on breaks and some invited the apprentices to their model-building clubs.
• In a separate post, Koerte presents Werner von Siemens Awards finalists working on electric locomotives that run on overhead wires or battery power, electromechanical products for direct current built for data centers and AI, and an AI-powered logistics solution described as a warehouse where an algorithm plays Tetris.

Read original · Original 2

It answers “what is the core AI thesis?” in one piece.

Peter Koerte argues that companies over 100 years old may have the biggest advantage in industrial AI because the scarce input is not money but decades of production data, sensor readings and engineering knowledge.

• Citing a Fortune interview, the post contrasts last year's AI investment — $286 billion in the US against $21 billion in Europe — and notes that Siemens is investing more than €1 billion in industrial AI over three years.
• It identifies trust as the constraint on using customer data, since that trust took decades to earn, and calls full technological sovereignty an illusion because the supply chain is global by design.

For brands planning a campaign

A strong anchor for a keynote or thought-leadership piece on industrial AI: it counters “compute wins” narratives with a Europe-versus-US funding contrast and a data-trust line that suits industrial audiences.

A natural follow-up to the industrial AI post.

After a conversation with Jake Sullivan, Peter Koerte argues that sovereignty is as much about the tech stack — compute and AI — as the energy stack of generation and grids.

• Countries will stay globally connected while each needs a national strategy for energy, digital and critical infrastructure.
• Sovereign compute needs power you control, while a grid absorbing this much new demand needs AI to run it.
• The post connects this argument to two new US plants worth more than $200 million in Georgia and Texas, making the electrical equipment AI data centers run on.
• It closes by saying countries that sort this out early will be able to host the infrastructure others are announcing.
• The post describes Sullivan as Kissinger Professor at the Harvard Kennedy School and a former US National Security Adviser, and thanks Torsten Ende and his team for arranging the conversation.

For brands planning a campaign

A ready frame for a sovereignty-themed panel or article: “connected and sovereign at the same time,” backed by a concrete plant investment and a named former official.

The best starting point on data center power demand.

Peter Koerte asked the team at a Siemens switchgear factory in Shanghai how they knew the data center wave was coming.

• Their answer — they had moved resources two years before the orders arrived — anchors the post.
• It gives electrification figures: China going from 30 percent to 35 percent by 2030, while the US sits around 22 and Europe around 21.
• The post lays out the chain that determines whether a data center opens on time — grids that carry the load, switchgear that handles it, cooling that keeps up, and buildings designed for the job — and notes that this takes longer to build than compute.
• It thanks Bin Lin, LiJun Dai and the Shanghai team by name.

For brands planning a campaign

Strong evidence for a brief on data center power, grid capacity or electrification. The “two years before the orders” detail supports a specific story about demand forecasting rather than a generic growth claim.

The Shanghai team used to walk nearly 30,000 steps a day collecting parts; now robots that pick their own routes bring the parts to them.

• Peter Koerte notes that colleagues there build the medium-voltage switchgear that quietly keeps data centers and factories powered, and describes an AI check that takes five minutes instead of thirty.
• The post thanks LiJun Dai and the whole Shanghai team.

For brands planning a campaign

A usable anecdote for talks on factory automation, AI inspection or workforce change: the before-and-after figures are easy to retell and verify against the post.

Peter Koerte presents the Werner von Siemens Awards finalists in the Technology category: Lars Loewenstein and team for electric locomotives that run on overhead wires when available and battery power when not; Zhi Gang Jin and team for electromechanical products for direct current, built for data centers and scaling AI; and Stephan Schlauß with the Erlangen team for an AI-powered automated logistics solution in a Siemens warehouse, where an algorithm plays Tetris. The post names all three, congratulates them and mentions a meeting at the October awards event.

For brands planning a campaign

Pairs a recognisable internal-awards format with three technology themes — rail electrification, DC infrastructure for AI and logistics automation — that could supply case studies or co-created content.

Read it if a campaign needs a human story.

Apprentices at the Siemens Amberg factory built a miniature industrial world, including a tiny 3D-printed version of Munich's Allianz Arena, powered and controlled entirely by the relays, safety devices, contactors and power supplies they manufacture.

• They built it on the factory floor, so colleagues stopped by on breaks to share ideas and tips; some, who were model builders themselves, invited the apprentices to their clubs.
• Peter Koerte frames what started as a training project as something the whole factory created together, comparing it to Hamburg's Miniatur Wunderland.

For brands planning a campaign

Fits employer-brand, apprenticeship or “industrial work is interesting” campaigns, balancing the infrastructure and AI posts with a workplace story.

In an interview with Fortune, they contrast US AI investment of $286 billion in the previous year with Europe's $21 billion, then argue that industrial AI needs something money alone cannot buy: decades of production data, sensor readings and engineering knowledge. They say Siemens is investing more than €1 billion in industrial AI over three years, with customer data as the key ingredient, and that sharing that data requires trust earned over decades.

A counterintuitive thesis, backed by an interview reference, an investment comparison and an argument about data and trust.

Koerte says full technological sovereignty is an illusion and the path forward is collaboration, not isolation. A campaign built around national self-sufficiency should adjust that framing.

For brands planning a campaign

A strong keynote or thought-leadership angle: industrial AI depends on domain data and customer trust, not just model access or capital. A brand evaluating Koerte for an industrial AI brief can ask for examples of how production data, engineering knowledge and customer trust shape deployment choices.

Examples: Source 1

In Shanghai, they asked a Siemens switchgear factory team how they knew the data-center wave was coming; the team said they had moved resources two years before the orders arrived.

• They cite China's plan to go from 30 percent to 35 percent electrification by 2030, while the US sits around 22 percent and Europe around 21 percent.
• The post lists grids, switchgear, cooling and buildings as the slower parts that determine whether a data center opens on time.
• They also write that two new US plants, representing more than $200 million in Georgia and Texas, make electrical equipment AI data centers run on.

Factory-floor anecdotes, macro electrification statistics and named regional investments connect AI demand to physical infrastructure.

The statistics and plant investments are specific to Siemens. A brand should adapt the framing to its own regions and supply-chain footprint, rather than treat the numbers as a general market forecast.

For brands planning a campaign

Koerte can speak concretely to data-center power, grid readiness and electrification constraints. A brand can use this material to build a keynote around the infrastructure decisions that determine AI capacity and ask for operational examples from switchgear and factory teams.

Examples: Source 1 · Source 2 · Source 3 · Source 4

They write that sovereign compute needs power you control, and a grid absorbing new demand needs AI to run it, making digital and energy sovereignty part of the same conversation. They say Siemens maintains a global technology base while placing manufacturing, engineers and supply chains inside the regions it serves.

A conversation recap with an external authority links company strategy to regional investment examples.

Koerte describes full technological sovereignty as an illusion and favors collaboration. A campaign should not overclaim self-sufficiency or treat sovereignty as a simple on-off choice.

For brands planning a campaign

This topic fits policy-adjacent thought leadership and localization narratives. A brand can build a session around the tradeoff between connected and sovereign, using regional manufacturing and infrastructure examples without presenting isolation as the goal.

Examples: Source 1 · Source 2 · Source 3

Workers used to walk nearly 30,000 steps a day collecting parts; now robots bring the parts and pick their own routes.

• An AI check takes five minutes instead of thirty.
• They also highlight Werner von Siemens Awards finalists in the Technology category, including battery-powered locomotives that run on any rail network, electromechanical products for direct current built for data centers and scaling AI, and an AI-powered automated logistics solution in a warehouse.

Before-and-after operational anecdotes, worker-centered detail and quantified time savings accompany award recognition.

The examples are Siemens-specific and mostly internal. A brand should ask for transferable lessons rather than assume the same results apply unchanged in another industrial setting.

For brands planning a campaign

Koerte can address manufacturing productivity, robotics and AI inspection with specific before-and-after examples. A brand can request a plant-floor case study or keynote segment on how automation changes daily work, not just headline efficiency numbers.

Examples: Source 1 · Source 2 · Source 3

The Werner von Siemens Awards post names finalist teams working on electric locomotives, DC infrastructure for AI and automated logistics.

• In Amberg, apprentices built an industrial miniature world on the factory floor, using relays, safety devices, contactors and power supplies to control a tiny 3D-printed Allianz Arena.
• Colleagues stopped by on breaks, shared ideas and invited the apprentices to their model-building clubs.

Award and apprenticeship storytelling names teams, technical challenges and factory-floor participation.

The posts center on recognition rather than market analysis. To work as thought leadership, they need a commercial or buyer-relevant thread.

For brands planning a campaign

This material supports employer brand, innovation culture and talent campaigns. For buyer audiences, a brand should pair the recognition story with the industrial AI and power themes so the content does not read as internal celebration alone.

Examples: Source 1 · Source 2

They note that most electrical products live in control cabinets or panels, quietly doing their job behind closed doors, then describe apprentices building a miniature world powered by the products they manufacture. Elsewhere, they describe colleagues who build medium-voltage switchgear that keeps data centers and factories powered, using robots and worker stories to show how the work has changed.

Human-interest openings and visual, concrete details turn technical products into accessible stories.

The anecdotal style may need stronger proof points for technical buyers. Brands should add data, specifications or independent evidence for engineering-led audiences.

For brands planning a campaign

Koerte is useful for campaigns that need to explain hidden infrastructure to general or mixed audiences. Brands can use this approach for event storytelling, social content or a keynote that makes electrification and data-center power concrete rather than abstract.

Examples: Source 1 · Source 2

Their posts present them as working inside Siemens, not as an outside commentator.

View example

View example

Koerte makes a similar shift from compute to power and construction. Both explain how operational data from industrial systems is central to running AI-era facilities.

Freedman writes from a background in data infrastructure and software, focusing on real-time data layers, Postgres and system architecture. Koerte draws on experience inside an industrial automation and electrification manufacturer, using factory-floor examples and regional investment to illustrate the same constraint.

Read original · Original 2

Koerte also argues that digital and energy sovereignty have become one conversation, and that power and grids determine where AI infrastructure can run.

Rodic writes structured geopolitical and investment briefs, with board-level implications and watch-next lists. Koerte tells the same story through company operations, named teams and specific factory and plant examples, without an advisory format.

Read original · Original 2

Koerte makes a parallel case: industrial AI depends on decades of production data and engineering knowledge, while factory automation brings practical changes to daily work.

Anbarasu's posts are event recaps and corporate vision statements from Körber, with a strong Industry 5.0 and human-centric framing. Koerte's posts are more anecdotal, often built around a single factory, a named team or a before-and-after number, with more emphasis on electrification and data-center power.

Read original · Original 2

These posts argue that power has shifted from an optimization variable to the binding constraint on AI infrastructure.

• They name the physical limits — power, cooling, water, grid interconnects and backup generation — and regional power capacity limits that have blocked server expansion requests despite available hardware and demand.
• The conference post speaks directly to controls engineers and system integrators about running industrial control alongside a data layer built to hold the volume and history of physical systems.
• It outlines a planned talk on architecture, how the field got here and where it goes next.
• A technical panel, working session or engineering-focused webinar is a better fit than a main-stage keynote.

The posts use software and systems framing, not factory deployments, named equipment programs or corporate investment figures.

• Ask for a plant-level example or corporate co-branded voice if the campaign depends on either.
• Power capacity appears as a constraint on where compute can run, not as a strategic question about countries.
• Ask for follower and demographic data before forecasting who will be in the room.

Examples: Source 1 · Source 2

Both posts read as analyst briefings: they quantify announced capacity, including a 2.16-gigawatt Australian data-centre lease expected to come online from 2027 and up to 2 GW of planned AI capacity against roughly 1.6 GW of existing Australian computing capacity.

• They then work through board-level questions — where data sits, where inference happens, who controls the infrastructure, which laws apply and whether electricity supply can support it — and list who gains and who loses.
• The Europe energy post applies the same frame to gas storage levels, LNG pricing and rearmament, arguing that energy security constrains industrial strategy, defence production and AI deployment.
• This suits a board briefing, an investor-facing session or a panel on where AI capital actually lands.

The posts summarize reported deals and market estimates rather than firsthand operational experience; the specifics come from sources cited inside the posts, not the author's own programs.

• They offer no engineering detail behind the numbers or named facility the author has run.
• Ask for an operator's testimony, a plant-floor example or a product-adjacent implementation story if the campaign needs one.
• Ask for audience data before projecting reach or room composition.

Examples: Source 1 · Source 2

The conference post frames industrial AI as a defining manufacturing capability whose value comes from applying AI to efficiency, productivity, resilience and better decision-making — and only when it meets deep industrial domain expertise.

• It argues that Europe must move from ambition to execution faster, build ecosystems and think globally while staying grounded in local realities.
• The second post sets out Industry 5.0 around resilience, human centricity and sustainability, connecting hardware, software, digital and AI into one foundation for smart factories and process solutions.
• This fits a European manufacturing audience, a transformation-themed thought-leadership piece or a moderated industry panel.

Both posts are event commentary and corporate positioning rather than close analysis of specific work.

• They contain no grid, power or data-center content; sovereignty appears only as a general Europe-versus-US-and-China framing.
• They describe a company position, not a documented program with figures or results.
• Expect an approvals process and confirm what can be evidenced.
• Ask for audience data to assess fit; a campaign centred on energy infrastructure needs a different speaker.

Examples: Source 1 · Source 2