TECH
Trump administration defends xAI gas turbines against NAACP Clean Air Act lawsuit
Ars Technica reports that the Justice Department filed a brief urging a Mississippi federal judge to dismiss a NAACP‑filed Clean Air Act suit alleging that xAI’s data center operates dozens of unpermitted gas turbines that emit hazardous pollutants. The government argues the turbines are “mobile sources” exempt from permits and that shutting them down would threaten national‑security‑critical AI systems used by the Department of War. The NAACP, represented by the Southern Environmental Law Center, says the turbines violate permit requirements, harm a predominantly Black community, and seeks an injunction and civil penalties.
SpaceX to acquire AI coding platform Cursor for $60 billion
SpaceX announced an all‑stock acquisition of the AI‑enhanced IDE Cursor for $60 billion, a move that follows its recent IPO and the merger of SpaceX with xAI. The deal aims to marry SpaceX’s vast compute resources with Cursor’s product, addressing both firms’ gaps, SpaceX’s need for an AI coding tool and Cursor’s compute constraints. TechCrunch noted that despite revenue growth, Cursor has struggled to break even and lost market share to Anthropic’s Claude Code.
Anthropic pauses token billing for Claude Agent SDK amid IPO preparations
Anthropic has halted a planned token‑based pricing change for its Claude Agent SDK, keeping current subscription limits intact for power users. The reversal follows user pushback and mirrors recent pricing shifts at GitHub Copilot. The company, which recently filed confidential IPO paperwork, says it will refine the pricing model to better support developers. (Ars Technica)
Qualcomm's new Snapdragon Reality Elite chip powers next‑generation smart glasses
The Verge reports Qualcomm unveiled its Snapdragon Reality Elite processor at the Augmented World Expo, with the first device, a pair of Xreal Project Aura Android XR glasses, expected to ship this fall. The chip delivers up to 60% faster graphics, 30% faster CPU and up to 160% higher NPU performance, supporting 4.4K resolution at 90 fps per eye while improving battery life by up to 20% and reducing heat. These upgrades aim to solve the visual quality, AI capability, battery, and overheating challenges that have limited previous smart‑glass designs.
DOJ approves Paramount‑Warner merger amid staff dissent and state challenges
The Justice Department announced approval of Paramount Skydance’s $111 billion acquisition of Warner Bros. Discovery, saying it will boost competition in streaming and film markets. The Wall Street Journal reported that career DOJ antitrust lawyers had leaned toward filing a lawsuit, but senior officials overruled them, prompting Sen. Elizabeth Warren to allege possible political favoritism. Federal regulators, including the FCC, have signaled support pending a waiver for foreign sovereign‑wealth funding, while several states plan to sue to block the deal and EU regulators are also reviewing it.
Uber to launch premium robotaxi service in Houston by mid-2027
Uber, in partnership with electric‑vehicle maker Lucid and autonomous‑driving startup Nuro, plans to roll out a premium robotaxi fleet in Houston by mid‑2027, following a similar launch in San Francisco. The service will use Lucid’s Gravity SUVs equipped with high‑resolution cameras, lidar and radars, while Uber focuses on the in‑cabin rider experience and operates a new 50,000‑sq‑ft depot in the city. Uber has invested $500 million each in Nuro and Lucid and aims to expand the robotaxi program to dozens of U.S. cities, competing directly with Waymo.
Pentagon uses generative AI to draft congressional reports, expands to 1.5 M staff
Ars Technica reports that the Department of Defense’s GenAI.mil platform, powered by tools such as Google Cloud’s Gemini, is now used by about 1.5 million personnel to produce congressional reports, performance evaluations, and award citations, reducing drafting time from hundreds of hours to a few. The service was highlighted by Pentagon CTO Emil Michael as a way to meet the growing volume of mandated reports, and the DoD has also signed contracts with eight leading AI firms for classified‑network deployment. However, the outlet notes concerns about accuracy and oversight after other firms, like KPMG, faced criticism for AI‑generated errors, underscoring the risk of unchecked generative‑AI use.
FIFA World Cup bug let anyone modify TV stream
TechCrunch reported that a security researcher known as BobDaHacker accessed FIFA’s internal broadcasting system by registering as a player agent and exploiting an unprotected API. The flaw allowed full control of the TV stream and commentator displays for every World Cup match, potentially enabling a hijack of the broadcast. FIFA patched the issue within hours but did not publicly acknowledge the researcher’s report.
Amazon's satellite launches hampered by New Glenn and Vulcan setbacks
Ars Technica reports that Amazon has hundreds of low‑Earth‑orbit satellites built and waiting in Florida, but only about 10% of its 3,236‑satellite constellation is in orbit. The company’s reliance on new launch vehicles is stalling: the New Glenn rocket exploded during a test in May and the BE‑4 engine that powers both New Glenn and ULA’s Vulcan faces problems, delaying those launches. So far, only Arianespace’s Ariane 6 has successfully delivered Amazon payloads, and the firm says it remains on track to begin commercial service later this year while diversifying its launch portfolio.
60% of US consumers find AI branding off‑putting, survey shows
A WordPress VIP survey of 2,000 respondents found that 60% of U.S. consumers view AI references in brand messaging as a turnoff, and 86% still doubt AI answers, preferring original sources. Trust in AI‑generated content without clear attribution was rated lower than airline fees, privacy policies, and medical bills. While consumer wariness rises, 60% of enterprise respondents reported increased traffic from AI search engines, and 74% said AI discoverability and attribution are a top priority for their brands.