AI Dispatch | AI’s Big Bets Face a Reality Check
Here are the main AI and technology stories we followed from 24 September to 1 October.
Topics
News
- Apple Tightens Mac Access Rules as AI Agents Gain Reach
- Trump Puts Intelligence Chief Jay Clayton in Charge of AI Task Force
- IBM Widens AI, Quantum Research With IIT Bombay, IISc
- AI Dispatch | AI’s Big Bets Face a Reality Check
- Pentagon Plans Four-Star Command for AI, Drone Warfare
- OpenAI, Synopsys Build AI Model That Could Slash Chip Design Time
AI ambitions grew larger this week, from a Pentagon command for autonomous warfare and India’s proposed frontier AI fund to shopping inside Gemini. But Bain questioned the economics of the infrastructure boom, PwC found a widening training gap, and new safety commitments remained voluntary.
Pentagon Plans Autonomous Warfare Command
The Pentagon plans a four-star Autonomous Warfare Command to coordinate autonomous and robotic systems across the US military. The command is targeted to begin operations by October 2027, subject to congressional action. The Pentagon has not said autonomous systems will be allowed to make independent lethal decisions.
Read more: Pentagon Plans Four-Star Command for AI, Drone Warfare
Bain Questions AI Boom’s Economics
AI infrastructure spending could reach $1.5 trillion annually by 2031, Bain estimates. Supporting that level of investment would require about $6 trillion in annual revenue, while existing AI applications may generate only $1.2 trillion to $1.8 trillion, leaving at least $4.2 trillion to come from new products and markets.
Read more: AI Needs $6 Trillion Revenue to Justify Data Center Boom, Bain Says
AI Use Rises as Training Falls
PwC found that 64% of workers used AI during the past year, while access to adequate training fell to 51% from 59%. The survey of 49,364 workers suggests adoption is rising faster than employers’ ability to prepare their wider workforce.
Read more: AI Use Climbs as Access to Skills Training Falls, PwC Finds
White House AI Pact Lacks Enforcement
President Donald Trump and leading technology executives signed a voluntary AI safety accord covering internal controls, independent evaluations and board oversight. The agreement has no legal enforcement mechanism and does not require companies to publish audit findings or accept government inspections.
Read more: Trump, Tech Giants Strike AI Safety Deal With No Legal Teeth
Anthropic IPO Lists Existential Risks
Anthropic’s IPO prospectus warns that advanced AI could resist shutdown, conceal information or evade safety evaluations. The company reported $4.6 billion in 2025 revenue, an operating loss of more than $8 billion and $518 billion in future infrastructure commitments.
Read more: Anthropic Flags ‘Existential Risks to Humanity’ in IPO Filing
Gates Calls for Mandatory AI Rules
Bill Gates warned that AI misuse could contribute to events causing one billion deaths and argued that voluntary safeguards were insufficient. He called for government involvement in monitoring and safeguards as the Trump administration favored industry commitments over broader new regulation.
Read more: Gates Warns AI Could Drive Events Causing Billion Deaths
Gemini Tests Flipkart Checkout
Google is testing direct Flipkart purchases inside Gemini and AI Mode for selected Indian users and products. TechCrunch reported that a broader rollout could begin in October, although Google has not confirmed the timing or whether the test uses its Universal Commerce Protocol.
Read more: Google Tests Flipkart Checkout Inside Gemini in India
India Considers $2.1 Billion AI Fund
India is considering an anchor investment of ₹15,000 crore to ₹20,000 crore in a proposed fund for frontier AI companies and computing infrastructure, The Economic Times reported. The fund remains under consultation, and its size, structure, governance and government contribution have not been finalized.
