Nvidia's quarterly revenue for fiscal year 2027's first quarter hit $81.6 billion, up 85 percent year-over-year. The company reported record data center results: $75.2 billion, 92 percent higher than the prior year. These numbers describe the visible side of the AI boom. The invisible side: the data centers running these chips consume electricity so voraciously that fusion energy investment has surged 69 percent to $4.48 billion in a single year. Not because fusion energy has become commercially viable, but because the technology industry sees nowhere else to turn for power.
Data Centers as a New Energy Problem
US data center electricity consumption reached roughly 400 terawatt-hours in 2026, according to industry analysis. That represents about 1.5 percent of total global electricity demand, a metric that stood at a fraction of this five years ago. For comparison, Spain consumes roughly 250 terawatt-hours annually. American AI data centers are now an energy consumer at the scale of a mid-sized industrial nation.
Consequences emerge in corporate filings. Google's annual greenhouse gas emissions rose 37 percent by the company's own account, primarily from AI infrastructure energy consumption. The company had committed to climate neutrality by 2030, but that goal now faces pressure. AlgorithmWatch notes that data center expansion increasingly influences electricity prices for consumers as hyperscalers compete with local grids for capacity.
Pressure on power supply has mobilized the energy sector. Mergers and acquisitions in the US energy sector reached a record $203.6 billion in the first five months of 2026. The main driver is expectation that data centers will continue raising energy demand for years ahead and that traditional utilities cannot meet those volumes.
The Fusion Energy Bet: Why Now
Fusion energy stands as the long-term answer to AI's power appetite. Since July 2025, $4.48 billion has flowed into fusion projects globally, a 69 percent increase from the prior year, according to an industry study dated July 14, 2026, cited by Wirtschaftswoche and Onvista. That is remarkable for a technology still without commercial operation.
The logic: technology companies can only meet long-term climate goals if they unlock renewable power sources providing constant, concentrated supply. Solar and wind are scalable but distributed. A large data center requires hundreds of megawatts at a single location. Fusion promises exactly that. Masayoshi Son, founder of Softbank, called nuclear fusion an inevitable energy catalyst for data centers and announced corresponding investment plans.
Skeptics remain unconvinced. Fusion has yet to deliver a commercial kilowatt-hour. Whether it scales fast enough to meet data center power demands of the coming decade remains open. What current investment actually reflects is more a hedging bet than certainty. Early investors secure technology access in case fusion arrives sooner than expected.
Apple: Hardware Engineer Assumes Command
On September 1, 2026, Tim Cook hands the CEO position to John Ternus. Cook, leading Apple since 2011, steered the company as an operations expert: supply chains, margins, scaling. Ternus is Senior Vice President of Hardware Engineering, responsible for iPhone, Mac, AirPods, and iPad. Apple announced the succession in April 2026; Cook becomes Executive Chairman.
The move signals intent. Apple lagged Google, Microsoft, and OpenAI in the AI race. Siri remained unimpressive compared to ChatGPT or Google's Gemini. In hardware, however, Apple holds strong positioning through its own chip ecosystem, the M-Series and A-Series. A CEO with hardware background sends a message: Apple bets on combining proprietary silicon with local AI inference, rather than wagering on cloud services dependent on third parties. CNN Business characterized the transition as signaling that Ternus faces "high expectations."
August 2 Deadline: EU Law, AI Regulation, and Germany's Startup Boom
For German companies, summer 2026 becomes regulatory reality. On June 11, the German parliament passed the KI-MIG, the Law for AI Market Monitoring and Innovation Promotion. The law transposes the European AI Act (EU Regulation 2024/1689) into national law and names the Federal Network Agency as the central market surveillance authority. Starting August 2, 2026, large parts of the EU regulation take direct effect.
Concretely, this means high-risk AI systems in critical infrastructure, education, employment, and law enforcement must undergo conformity assessment. Companies deploying or developing such systems must meet documentation obligations. Penalties can reach seven percent of global annual turnover.
Meanwhile, recent data from the German Startup Association show Germany is in the midst of an AI-driven founding boom. In the first half of 2026, 3,053 new startups launched, 52 percent more than the second half of 2025. 1,038 had direct AI relevance. June 2026 alone saw $2.6 billion in venture capital flow to the German startup market, a month with over 600 new registrations. Six companies reached unicorn status in the half-year, bringing Germany's total to 36.
The Federal Network Agency will serve as the first regulatory contact for many of these startups. How it balances oversight with innovation promotion, the dual mandate in the KI-MIG name, will become clear in coming months. One thing is certain: as of August 2, testing in a regulation-free environment is over.