India makes AI mandatory in schools from next term, the EU's content labeling law is five weeks from enforcement, and data center energy costs are landing in your electricity bill.
AI is now a required subject for every Indian student from Class 3
From the 2026-27 academic year, CBSE is mandating AI and computational thinking from Class 3 onwards, with a 15-hour AI skill module from Class VI. India's SWAYAM platform already has over 41 lakh student enrolments across 110+ AI courses from IIT and IISc. Under PMKVY 4.0, 36,584 people have been trained in AI: 45% of them women.
Why it matters: This is one of the largest compulsory AI education rollouts attempted anywhere in the world by enrollment scale. If you teach, work in edtech, or are advising students on careers, the 2026-27 school year is the planning horizon.
Five weeks left: the EU's AI content labeling law becomes mandatory on August 2
Europe's AI Act reaches full enforceability on August 2, 2026. After that date, platforms must mark AI-generated content in a machine-readable format and clearly disclose synthetic content to users at first interaction. A separate EU Code of Practice on deepfake labeling is also being finalized this week, covering watermarking and metadata standards across images, video, and audio.
Why it matters: This is the world's first binding law requiring AI content to be labeled, not just flagged. If you use AI-generated media in your work for European audiences, compliance is no longer optional. Other countries are expected to follow within 12 months.
US data center power demand is set to nearly double, and households may pay
US data centers could reach 150 gigawatts by 2028, up from roughly 80 GW today. Over 300 state-level bills on data center energy have been introduced in 2026 alone, and 230+ environmental groups are calling for a moratorium on new data centers. The White House's voluntary Ratepayer Protection Pledge (now signed by seven major AI companies) commits them to not passing data center costs onto electricity consumers, but it carries no legal enforcement.
Why it matters: AI's energy footprint is no longer an abstract concern. Communities near data centers are already reporting local electricity price pressure. This is becoming a consumer issue, not just a tech one, and the bills are arriving.
Qualcomm is spending up to $14B to own the chips that run AI
Qualcomm is nearing a $4 billion acquisition of AI software startup Modular and separately pursuing AI chip company Tenstorrent for $8–10 billion. Together, these deals would position Qualcomm as a full-stack AI chip player across edge devices and cloud, directly challenging Nvidia's dominance in AI hardware.
Why it matters: Nvidia has set AI hardware pricing almost unilaterally. A credible challenger emerging means AI compute costs could eventually fall faster. If AI hardware diversifies, the cost of running AI tools drops, which matters for everyone, not just big companies.
AI is creating a two-speed job market, and the gap is already measurable
PwC's 2026 Global AI Jobs Barometer shows companies most integrated with AI are raising wages and headcount faster than those least exposed. But it is not an even rise: skills like judgment, leadership, and communication are being rewarded more; roles built around routine tasks face the most pressure. AI is also the top cited reason for layoffs for the third consecutive month.
Why it matters: The question is no longer "will AI take my job", it is "am I building the skills AI cannot replace?" According to PwC's data, people already ahead of that shift are earning measurably more. The gap between the two tracks is widening.
Anthropic's US-forced suspension is sparking a sovereignty debate in India
After the US government directed Anthropic to suspend its Fable 5 and Mythos 5 models globally in mid-June, CNBC reported this week that the episode has reignited India's debate about sovereign AI. Businesses and researchers who lost access overnight now see foreign AI dependence as an infrastructure risk. Critics say India's domestic AI investment is "too slow, way too small" to offer a real alternative.
Why it matters: When a government in another country can switch off an AI tool your organisation depends on (overnight) it stops being a product issue and becomes a geopolitical one. India's response to this episode will shape how quickly it builds domestic AI alternatives.
August 2: the EU's AI Act enforcement begins, and the world is watching
Five weeks from now, the EU AI Act becomes fully enforceable. The real test is not the law itself, it is the first enforcement action. When a regulator fines a platform for unlabeled AI content, it sets the price of non-compliance for every company operating in or selling to Europe. Other regulators in the UK, India, and the US will calibrate their own timelines based on what follows. The countdown is on.
DeepSeek's 1.6 trillion-parameter model + AI funding hits $242B in one quarter
China's DeepSeek released its V4 series headlined by a 1.6 trillion-parameter mixture-of-experts model: the largest open-weight AI released to date, built for heavyweight reasoning and coding. Separately, global AI startup funding reached $242 billion in Q1 2026 alone: 80% of all venture capital globally, with Anthropic's $65B Series H still the single largest private round in tech history.