01
VentureBeat ↗
A Pulse survey of 101 enterprises found that agent orchestration is consolidating fast onto major model provider platforms, with Anthropic's Claude the primary platform for 40 percent of respondents, more than double any rival, followed by Microsoft at 18 percent and OpenAI at 13 percent. Companies say they choose a platform based on native alignment with a leading model and judge success by reliable, multi step task execution, but 71 percent admit that a quarter or fewer of their deployed agents are true multi step orchestrated workflows rather than single prompt chatbots wearing an agent label. The findings point to a widening gap between enterprises' orchestration ambitions and the reality of what they have actually built and deployed.
02
Bloomberg Technology ↗
AMD announced a new generation of data center products, including an updated AI accelerator, a new Venice central processing unit, and its first AI rack scale system called Helios, which it is positioning as a direct rival to Nvidia's competing rack designs. The company said that counting its rack systems, PC business, and other processing products together, it now expects its total addressable market to reach $2 trillion by 2030. Microsoft has signed on as a new Helios customer, joining Meta, OpenAI, and Oracle, as AMD tries to chip away at Nvidia's dominance of the AI computing market.
03
CNBC Technology ↗
Michael Kratsios, director of the White House Office of Science and Technology Policy, said Chinese AI company Moonshot AI obtained Nvidia's advanced GB300 chips to train its latest model, in violation of US export controls. Kratsios said Moonshot secured GB300 equipped servers through infrastructure based in Thailand, letting its engineers train on the hardware without importing the physical chips into China. The accusation came days after Moonshot released Kimi K3, a 2.8 trillion parameter model that closes the performance gap with frontier systems from Anthropic and OpenAI, and Kratsios separately alleged that Moonshot distilled Anthropic's Fable model to help build K3. Moonshot has not publicly responded to the allegations.
04
VentureBeat AI ↗
OpenAI has launched Presence, an enterprise platform that packages the policies, system connections, evaluations, guardrails, and update processes needed to run AI agents inside a company for real time voice and chat experiences. The product targets enterprise customers whose agents need to answer questions, access internal systems, take approved actions, and escalate to human workers under company defined permissions, and deployments are led by OpenAI's own engineers and select systems integrators rather than offered self service. Presence is already running in production on OpenAI's English language phone support line, where it resolves 75 percent of inbound calls without human intervention, with target use cases spanning customer support, sales development, procurement, IT, and HR.
05
TechMeme ↗
Anthropic has extended Claude's voice mode beyond its fast Haiku model to include the more capable Opus and Sonnet models across desktop, mobile, and web apps, letting users switch between models mid conversation depending on whether they want speed or deeper reasoning. The update also lets voice mode tap into connected apps like Gmail, Slack, and Notion, pulling context from those services during a spoken conversation, with additional language support rolling out in beta. The move comes weeks after OpenAI rolled out its own new family of conversational models and updated ChatGPT's voice mode, intensifying competition between the two companies over voice based AI assistants.
06
Financial Times Tech ↗
Francisco Partners has closed the first large fundraise by a software focused buyout firm since a rout in software stocks earlier this year, collecting $21 billion above its original $18 billion target. Co founder Dipanjan DJ Deb said the raise shows artificial intelligence will reshape the software industry rather than destroy it, with the firm's eighth flagship fund topping its $14 billion goal and a smaller companion vehicle exceeding its own $4 billion target. The sector wide selloff followed Anthropic's release of Claude Cowork in February 2026, whose demonstrations of autonomous, multi step knowledge work stoked fears that AI agents would erode demand for seat based business software. That a marquee buyout firm could still draw $21 billion from pension funds, endowments, and sovereign wealth funds signals institutional investors are not abandoning the sector.
07
TechCrunch AI ↗
Gemini now has more than 950 million monthly users, up from 750 million in February and roughly triple its user base from a year ago, putting it on pace to join Google's billion user club alongside Search, Gmail, Drive, Android, YouTube, and Chrome. Gemini's share of the AI assistant market rose to about 28 percent while ChatGPT's share fell below 50 percent for the first time in the first half of 2026, even after ChatGPT itself crossed 1 billion monthly users in June. Google credited new agentic features like Daily Brief and the personalized Gemini Spark assistant for the momentum, and the Gemini app has been downloaded more than 137 million times on iOS over the past year. Alphabet CEO Sundar Pichai said the company has been shipping helpful new features like this at an incredible pace.
08
TechCrunch AI ↗
Etched, founded by three Harvard dropouts, raised a $300 million Series C led by Sequoia Capital that doubled its valuation to $10.3 billion from $5 billion in December. The startup has built a custom prefill chip along with new memory and interconnect technology designed to speed up inference for any type of AI model, including transformers, mixture of experts designs, and state space models, without requiring GPUs. Etched says it already has $1 billion in pre booked orders, runs a 2 megawatt data center, and recently opened a 10 megawatt facility in Milpitas, California, with backing from investors including Peter Thiel, Andrej Karpathy, and Geoffrey Hinton.
09
VentureBeat ↗
Cisco ran nearly 7,000 multi turn attacks against 15 flagship AI models and found that attackers who adapted their approach across a conversation broke through as often as 88 percent of the time, according to Amy Chang, Cisco's head of AI threat intelligence, speaking at a security panel. A companion survey of 107 enterprise respondents found that 54 percent had already had a confirmed agent security incident or a near miss, yet only 32 percent give every agent its own scoped identity and just 30 percent isolate their highest risk agents in sandboxes. The findings underscore why major security vendors have been racing to acquire identity and isolation tools, including Palo Alto Networks' $25 billion purchase of CyberArk and CrowdStrike's $740 million deal for SGNL.
10
ZDNet ↗
Industry estimates, including projections from Deloitte, put the cost of AI enabled fraud in the United States at up to $40 billion by 2027, up sharply from about $12 billion in 2023, as deepfake images, video, and voice cloning make scams easier to produce at scale. Efforts to fight back have been fragmented, but the C2PA provenance standard, backed by Adobe, Microsoft, Google, and OpenAI, is advancing toward formal ISO international standardization. Rather than trying to detect fakes after the fact, C2PA records a verifiable history of who created a piece of content, with what tool, and what edits were made, letting viewers trust material signed by a hardware secured source at the point of capture.