Amazon's Nova2 trails frontier leaders, but its AI chief signals an aggressive catch-up roadmap within 12 months.
The AI Daily
Daily · June 18, 2026
The week's dominant story is the Trump administration's export-control order forcing Anthropic to yank its most powerful models offline — a move that spooked world leaders at the G7 summit into demanding guarantees that American AI can't be switched off overnight. Meanwhile, the AI model market is fragmenting fast: ChatGPT's share has fallen below 50% as Claude and Gemini surge, and Amazon's AI chief openly admits Nova2 is still chasing the frontier pack.
🧠 Foundation models
go deeper →The chatbot market is visibly fragmenting as Claude and Gemini eat into ChatGPT's majority share, while Amazon's frontier ambitions and a novel AI-chemist deployment signal how competitive model-layer development has become.
Amazon has lagged OpenAI and Anthropic, but AI chief sees path to catch up in 'coming year'
ChatGPT's AI market share drops below 50% as Gemini and Claude gain users: Report
ChatGPT loses majority chatbot share for first time as Claude's audience grows 452% YoY — competitive dynamics are fundamentally shifting.
A near-autonomous AI chemist improves a challenging reaction in medicinal chemistry
GPT-5.4 powers a near-autonomous chemistry agent that materially improves a hard drug-synthesis reaction, pushing AI into real lab work.
New research shows how AMIE, our medical AI, could help manage health conditions.
Google's AMIE medical AI matches primary care physicians in complex disease management in a Nature-published study.
🏗️ Infrastructure
go deeper →Capital continues to flood into AI compute infrastructure globally — from Nvidia's $25B bond sale to new data centers in Hong Kong, Seoul, and India — as the physical backbone for the AI boom keeps scaling.
Nvidia to raise $25bn in corporate bond sale
Nvidia's first bond sale in five years raises $25B, signaling massive capex appetite to sustain AI chip dominance.
Nvidia supplier SK Hynix ships samples of next-generation chips to major customers
SK Hynix is sampling next-gen memory chips to top customers, keeping the AI hardware supply chain moving forward.
Equinix launches sixth data center in Hong Kong
Equinix's new Hong Kong HK6 facility targets AI and high-density workloads in a key Asia-Pacific market.
STT GDC launches data center in Seoul, South Korea
STT GDC opens its first Korea facility in Seoul, expanding Asian AI infrastructure capacity.
Canada's CPP Investments forms joint venture with Indian data center firm CtrlS
CPP Investments commits up to $741M into Indian data centers via a CtrlS JV, betting big on India's AI infrastructure growth.
Amazon AI exec predicts first 'commercially useful' quantum computers in 5-7 years
Amazon's AI chief sets a 5-7 year horizon for commercially useful quantum compute, tempering near-term hype.
💰 Funding & deals
go deeper →Capital formation around AI remained brisk, with a $400M materials-science raise, a $100M cyber-warfare unicorn, and several earlier-stage rounds signaling that investors see no slowdown across the AI stack.
Pramaana Labs raises $27M to make AI prove its answers
Pramaana raises $27M seed for formal-verification AI that checks model outputs against domain rules before returning results.
AI material discovery startup CuspAI reportedly raising $400M round
Jeff Bezos-backed CuspAI is closing a $400M round to apply AI to materials discovery, a capital-intensive frontier science bet.
AI cyber warfare startup Twenty raises $100M at $1B valuation
Twenty hits unicorn status with a $100M Series B for AI-powered cyber warfare tools sold to U.S. military and intelligence.
Hydra Host raises $100m in Series A funding round
Hydra Host raises $100M Series A to expand global AI compute capacity, joining a crowded neocloud race.
Conduct raises $60M to speed up software modernization projects
Index and ICONIQ co-lead a $60M Series A in Conduct AI's code-modernization platform, with SAP as a strategic backer.
Convey closes on $38M round to help companies automate repetitive work with AI teammates
a16z backs Convey's $38M raise for a no-code agentic platform that lets non-technical workers deploy autonomous AI teammates.
🔧 Middleware & platforms
go deeper →Developers got a flurry of new tooling this week — from AWS's autonomous security scanner and agentic context graph to Vercel's agent infrastructure framework — as the race to build reliable plumbing between models and enterprise apps intensifies.
Key takeaways from day two of the Databricks Data + AI Summit
Databricks' second summit day deepened its data-plus-AI platform story with product announcements relevant to every enterprise AI stack.
AWS enters the context layer race with a graph that learns from agents, not manual curation
AWS launches a self-updating knowledge graph that agents maintain automatically, removing the biggest manual bottleneck in enterprise RAG.
AWS launches Continuum to find and fix code vulnerabilities at machine speed
AWS Continuum uses frontier AI to discover, validate, and auto-remediate software vulnerabilities with minimal human input.
Vercel launches a new framework and enterprise controls for agentic AI infrastructure
Vercel's new agentic framework and enterprise controls aim to standardize how companies deploy and scale AI agents in production.
Anthropic ships major Claude Design overhaul with design system imports, code round-trips, and a fix for its token-burning problem
Claude Design hits 1M users then gets a major upgrade addressing its notorious token overconsumption and adding design-system imports.
AI ready data is the missing link keeping enterprise AI stuck in pilot mode
Enterprises stall on AI ROI not from lack of compute but from data silos — AI-ready data infrastructure is the new bottleneck to solve.
📱 Application solutions
go deeper →AI is landing in high-stakes domains from drug discovery to medical diagnosis, while the Anthropic export-control saga is making enterprises and governments rethink the reliability of AI access as an operational dependency.
AI will lead to labour shortages, Bezos says in optimistic talk
Bezos predicts AI-driven labour shortages rather than mass unemployment, reframing the workforce impact debate.
Two-thirds of Americans think AI is advancing too quickly
Pew finds 49% of Americans now use AI chatbots but 63% believe the technology is moving dangerously fast.
World leaders want American AI. They just don't want America to be able to turn it off.
Macron and Modi used the G7 to demand AI sovereignty guarantees after Anthropic's abrupt blackout made the kill-switch threat concrete.
AI is hurting Apple in more ways than one: it may force iPhone price increases
Apple's AI ambitions are straining its memory-chip supply chain, with Tim Cook warning that iPhone price increases are now unavoidable.
Operating a Humanoid With Your Body Is a Hot Job in China's Hardware Capital
Shenzhen workers are using VR rigs to teleoperate humanoid robots, generating training data for China's embodied AI push.
⚖️ Policy & legal
The Anthropic-White House export-control clash dominated policy news this week, exposing just how blunt and legally ambiguous the Trump administration's AI controls have become — and sparking a G7-level diplomatic row over AI sovereignty.
'A signal of where power sits': Trump and world leaders joined by OpenAI, Anthropic, Google at G7
G7 leaders and frontier AI CEOs convened to debate AI standards and sovereignty, with Anthropic's blackout casting a long shadow.
CEOs of Anthropic and Google DeepMind call for U.S.-led AI coalition in meeting at G7
Amodei and Hassabis jointly lobbied G7 leaders for a U.S.-led alliance to set global AI standards, countering Chinese influence.
Anthropic got hit by export rules nobody understands
Vague export rules forced Anthropic to block Fable 5 and Mythos 5 even for its own staff — exposing AI controls as legally incoherent.
Vibe-decoding the White House-Anthropic fight over Fable
Analysis unpacks the political dynamics behind Trump's order to revoke Anthropic model access — power play as much as security concern.
The White House Wants Anthropic to Block All Jailbreaks. That May Not Be Possible
The administration's demand for zero-jailbreak guarantees before Fable 5 can return is technically impossible, say AI security experts.
At G7, Macron says he expects progress on broadening access to Anthropic's Mythos
France's Macron publicly pressures Washington to restore Anthropic Mythos access for allied nations at the G7 summit.
The Korean Telecom Giant at the Center of Anthropic's Mythos Controversy
The White House ordered Anthropic to cut SK Telecom's access to Claude Mythos over alleged China ties, days before the wider blackout.
Legal AI startup Eve hit with patent infringement lawsuit
Legal AI startup Eve faces a patent infringement suit, flagging growing IP litigation risk in the AI applications space.
🇮🇳 India lens
India emerged as a meaningful vector in global AI geopolitics this week: PM Modi raised AI sovereignty concerns at the G7, Canadian pension capital committed $741M to Indian data centers, and Nykaa struck a multi-year OpenAI deal to bring conversational AI to Indian beauty and fashion shoppers.
Telegram CEO Pavel Durov claims Reliance is sabotaging Telegram access for users outside India
Durov's BGP-hijacking allegation against Reliance spotlights India's telecom infrastructure as a geopolitical flashpoint for global platforms.
Nykaa taps OpenAI for conversational shopping across beauty and fashion
Nykaa's multi-year OpenAI deal embeds ChatGPT-powered discovery into India's leading beauty and fashion platform.
📖 Beyond the headlines
The White House Wants Anthropic to Block All Jailbreaks. That May Not Be Possible — The administration is effectively demanding a technical impossibility as a precondition for market access — a precedent that could shape how every frontier AI company negotiates with governments for years. Any strategist thinking about AI regulatory risk, export controls, or product liability should understand why 'zero jailbreaks' is not an engineering problem but a political one with no clean solution. Read →