S Navani

PhD Economics & Finance | AML-CFT | Energy | Data Center | EU Policy

[Elon Musk @elonmusk]

SpaceX and Tesla are each building 100GW/year of solar production capacity as fast as possible, but natural gas will still be needed to supplement and bootstrap solar for several years.

The limiting factor for nat gas turbine production is casting the blades & vanes. By doing in-house casting at SpaceX, we can accelerate nat gas turbines coming online by up to 18 months, which is a profound game-changer.

[International Energy Agency @IEA]

The tech sector isn’t just consuming energy – it’s also shaping supply

It made up around 40% of all corporate power purchase agreements for renewables signed in 2025

Data centre deals alone covered about half of the sector’s current electricity use 👉 https://t.co/oPlklGL7zE

[Judith Dada @DadaJudith]

Doubling AI compute by 2030 won’t nearly be enough

[The Indian Matrix @indianmatrix]

India’s external debt is among the lowest in the G20, at just 19.4% of GDP.

While several major economies carry external debt exceeding 100% of their GDP, India remains at the lower end of the table.

Here’s how India compares with the world’s major economies.

[World Nuclear News @W_Nuclear_News]

A consortium led by France-headquartered innovative reactor developer @newcleoenergy has been awarded funding to develop the conceptual design of a #nuclear barge able to host multiple 200 MWe small modular reactors https://t.co/MqGdYPqxlw

[Chris Meder @EVCurveFuturist]

In Australia, under proposed national standards, new data centres must bring their own 100% renewable energy supply, firmed by batteries or gas. Crucially, this means new, additional renewable generation, not simply claiming certificates against generation already on the grid.

And the scale is enormous. AEMO is already tracking 225 data centres in development, with data-centre electricity demand forecast to rise from around 3% of NEM consumption today to 13% by 2036, potentially around 34 TWh a year.

Climate Change & Energy Minister Chris Bowen puts the challenge bluntly: “Data centres are huge consumers of electricity. They are whales.” In 2024, US data centres consumed as much electricity as the entire country of Sweden. Australia wants to get ahead of this.

Australia isn't alone. Similar policy moves are underway in China, Germany and Ireland, but Australia's proposed framework is among the most aggressive: 100% renewable energy, tied to new additional generation and backed by firming, with the new rules intended to begin from 2027.

Instead of allowing this enormous new load to compete with households and businesses for existing generation, Australia wants it to help finance the new supply needed to meet it.

AI growth → electricity demand → renewable PPAs → new solar & wind → more batteries & transmission → stronger grid.

This has always been part of my thesis: AI will become a major new demand engine for Australia's solar, wind and battery buildout, rather than a brake on the energy transition, potentially creating a blueprint the rest of the world can replicate.

AI and renewables don't have to compete. Done right, each can accelerate the other. #SWB #Bettrification

https://t.co/VsPqEZdi17

https://t.co/iwxnZUV02g

[Manish Singh @refsrc]

India has reached the income level at which the U.S., Europe and China all went on chain-hotel building sprees.

[Niels Harksen @HarksenNiels]

Germany would be wise to refurbish and restart its nuclear power plants, using the uncommitted power capacity to attract AI inference data centers and have sovereign compute on its own soil.

[Nuklearia e. V. @Nuklearia]

Polen baut sein erstes Kernkraftwerk. Der @ndr stellt es jedoch so dar, als würden die Polen 45 Milliarden Euro für etwas ausgeben, das ein Windpark für fünf Milliarden genauso gut kann.

Dabei vergleicht er bei der Leistung einen Reaktor, bei den Kosten aber alle drei. Richtig gerechnet ist Windkraft in diesem Vergleich schon bei den Investitionskosten teurer. Systemkosten noch nicht eingerechnet.

Deshalb wurde Programmbeschwerde eingelegt. Im Blog dokumentieren wir sie und zeigen, wie die polnische Umweltorganisation @fota4climate den Bau unterstützt.

https://t.co/THjJ9u4czs

[pv magazine @pvmagazine]

Chile’s hybrid heroics: In 2024, Chile passed a law incentivizing battery development paired with utility-scale solar. This move quickly led to a flood of big battery project announcements, as the capacity… https://t.co/brRSXZgJxS #Photovoltaics #EnergyStorage #RenewableEnergy

[S Navani @DrNavani]

1. Money : Training needs $1.8B to $2.5B in capital. European investor groups rarely give this amount of money for just one high-risk model. A federal funding will be needed, which is an unlikely path as rising public expenditure and debt levels already put long-term pressure on fiscal budgets.

2. Electricity & Nuclear Phase Out: Germany phased out its last remaining nuclear power plants, leaving industrial electricity as some of the most expensive in the world at €0.22 per kWh (compared to just 6 cents per kWh in the US). High energy costs are already forcing traditional manufacturing industries to leave Germany, making a massive 350 GWh training run costing €55M to €75M in electricity alone economically unviable. 

3. Compute : Fable 5 needs 50,000 to 100,000 top tier GPUs connected in one fast network. Leading German AI companies like Aleph Alpha and top universities only have access to smaller, separate GPU clusters. No single datacenter in Germany has a single cluster of 100,000 GPUs today.

4. Talent : Fewer than 500 researchers worldwide know how to build frontier models from scratch, and most live in San Francisco. Top US companies pay $1.5M to $8M per year, while regular German tech salaries are €150k to €250k. Moving this top talent to Germany would be extremely expensive.

5. Time (Grid and Bureaucracy Delays): Getting approvals and connecting 100 MW of continuous power to the German energy grid takes 18 to 24 months. Building the whole project in Germany takes 3 to 4.5 years, compared to 12 to 18 months for US tech companies.

So I would claim that currently it is not possible. Instead of spending $2 billion+ to pre-train, simply download top-tier open-weight models (like Kimi K3, Llama 3/4 series or Mistral Large/NeMo variants).

[CATL @catl_official]

What if a battery could build its own anode?

CATL’s self-forming anode technology removes the conventional graphite anode, increasing volumetric energy density by 60% and gravimetric energy density by 50%.
#CATL #BatteryTechnology #EVTechnology #BatteryInnovation

[International Energy Agency @IEA]

Surging EV exports from China are shielding manufacturers from declining domestic car sales.

While China's total car sales fell by more than 20% in the first half of 2026 versus a year earlier, electric car exports increased by more than 120% 👉 https://t.co/y9MSqMWIhv

[Daniel D. Eckert @Tiefseher]

Die Wachstumsschwäche in unserem Nachbarland Frankreich 🇫🇷 ist gar nicht gut. Im Gegensatz zu den deutschen BIP-Zahlen sind die französischen nach unten revidiert worden, die französische Wirtschaft ist 2026 knapp einer Rezession entgangen 😕

[International Energy Agency @IEA]

Global energy spending is on track to reach $3.4 trillion in 2026

Around $2.2 trillion of this is set to go to grids, storage, low-emissions fuels, nuclear, renewables, efficiency & electrification

And about $1.2 trillion goes to oil, natural gas & coal: https://t.co/8IajZj1gIL

[S Navani @DrNavani]

@Tiefseher Das weltweite KI-Rennen ist ein Duopol: Nur die USA und China verfügen über die nötige Rechenleistung (Compute) für Spitzenmodelle. Doch die neue Hürde ist nicht mehr der Chip, sondern der Strom.

[McKinsey Global Institute @McKinsey_MGI]

The chips get the headlines. The electricity bill decides the economics. Same data center: $200/MWh in China, $380 in London.

Explore the investment case: https://t.co/KNq13bSKU0

[Judith Dada @DadaJudith]

Danke für die Einladung zur heutigen Kabinettklausur @bundeskanzler. Deutschland und Europa können noch so viel mehr erreichen - trotz aller Herausforderungen ist transformative KI eine historische Chance für dieses Land, eine lebenswertere Zukunft für uns alle zu gestalten. Es braucht jetzt Mut für den Wandel, Ambition für das Große und den eisernen Willen, hart für den Fortschritt und unsere Sicherheit zu arbeiten.
1) Zugang zu den stärksten globalen KI-Modellen & die Ambition, eigene Spitzenmodelle aufzubauen
2) Drastischer Ausbau von Rechenkapazitäten in Europa
3) Die Expertise führender Gründer & Wissenschaftler und Anwender in Expertengremien inkludieren
4) Dynamischen Wandel ermöglichen und Existenzen anstatt von Stellen sichern - Reform des Arbeitsrechts
5) Eine konkrete Vision erarbeiten, wie KI den Alltag Bürgerinnen und Bürger spürbar verbessert

[Normal Guy @Normal_2610]

German car parts makers used to produce hundreds of millions of components yearly

Now volumes are dropping fast, Germany lost 124,000 factory jobs in 2025, auto sector hardest hit at 50,000. VW profits fell 40%, Mercedes profits halved. Suppliers are moving into defence, medical devices, and robotics.

Schaeffler now makes parts for drones and armored vehicles. VW is talking to Israeli defence firms. These moves help individual companies survive. But defence and medtech need far fewer parts than car factories.

Cannot replace lost scale

Swiss products carry premium because Swiss Made label means quality. By law, 60% of manufacturing costs for industrial goods happen inside Switzerland to use Swiss cross.

But franc keeps getting stronger, making factories more expensive than competitors. Bernina moved sewing machine production to Thailand and cannot use Swiss label anymore. On shoes got permission to use Swiss cross even though shoes are made in Vietnam, because R&D happens in Zurich.

When rules meant to protect your brand force you to either lose money or lose your label, protection becomes problem

German auto suppliers losing volumes need markets with scale demand. India offers that.

PLI subsidies worth $26bn cover 14 sectors including medtech, auto parts, and defence. Schaeffler, Bosch, and Continental already sell into India. 3 of 4 German suppliers are diversifying away from autos.

Precision components like ceramics, sensors, and bearings designed for European cars could serve Indian factories making medical devices, defence equipment, and electric vehicles.

German firms bring engineering depth. India brings volume and growth. Matching distressed supply with growing demand creates value both sides need right now.

https://t.co/usDnaq7lmM

[McKinsey & Company @McKinsey]

The demand for AI skills is showing up across more jobs, industries, and wage groups than you might expect.

As AI becomes more embedded in the workforce, organizations have an opportunity to help people build the skills to use it in their day-to-day work. https://t.co/jsqsdiNSWS