01
Financial Times ↗
About half a dozen Anthropic investors say they are targeting a $2 trillion valuation for the company's planned October listing, which would make it the largest initial public offering in history and push it past SpaceX. The figure rests on projections that Anthropic's annualized revenue will reach $100 billion to $120 billion by the end of 2026, more than ten times its level at the start of the year. One investor argued $2 trillion was conservative and floated $3 trillion instead. The targets come from shareholders rather than the company, and senior Anthropic executives have reportedly not set a valuation goal even privately.
02
TechCrunch AI ↗
OpenAI has opened a preview of Ultrafast, a new API service tier that runs its flagship GPT-5.6 Sol model at up to 14 times normal speed, delivering as many as 750 output tokens per second. The tier is powered by Cerebras wafer scale hardware rather than conventional GPU clusters, a speed no GPU cloud has publicly matched. OpenAI is aiming the mode at enterprise workloads where latency matters most, naming incident response, customer support, financial analysis and e commerce as target use cases. Access is limited for now, and the company says it will widen availability as capacity grows.
03
Ars Technica ↗
Google released Gemini 3.7 Flash on August 13, only three weeks after Gemini 3.6 Flash shipped, an unusually short gap that Google attributes to algorithmic gains in the model's reasoning core rather than a larger training run. Coding scores moved sharply in that window, with its main frontier coding benchmark rising to 43.6 percent from 34.4 percent and its software engineering benchmark jumping to 65.3 percent from 49.0 percent, while enterprise automation and complex document scores roughly doubled. The model keeps a 1 million token context window with 64K token output and handles text, images, audio and video. Introductory pricing holds at $0.75 per million input tokens and $3.75 per million output tokens, half what Google says the rate becomes on January 1, 2027.
04
TechCrunch AI ↗
Databricks closed a $5 billion round at a $190 billion valuation after setting out to raise just $1 billion, with roughly two dozen investors putting forward some $15 billion of interest once the fundraise leaked. Coatue led the round, joined by Blackstone, MGX, T. Rowe Price accounts and Sixth Street Growth. Chief executive Ali Ghodsi said he took more than planned because artificial intelligence is expensive, pointing to multibillion dollar cloud commitments, a 100 person research team and continued acquisitions. The company is now at a $7 billion annualized revenue run rate growing 80 percent year over year, with its Lakebase AI database at $100 million since launching in June 2025.
05
TechRepublic ↗
Foxconn reported second quarter net profit of NT$59.97 billion, up 35 percent from NT$44.4 billion a year earlier and ahead of the NT$58.8 billion analysts expected, while revenue rose 41 percent to NT$2.526 trillion. Cloud and networking products, the division that contains AI servers, accounted for 51 percent of revenue and cleared half for the first time, overtaking the smart consumer electronics unit that builds the iPhone at 29 percent. The contract manufacturer is building AI server plants in Mexico and Texas to supply Nvidia and expects capital expenditure to rise about 30 percent in 2026. Management flagged advanced chip packaging capacity as the main constraint on AI server output heading into 2027.
06
TechCrunch AI ↗
Microsoft is folding its consumer Copilot app and the enterprise Microsoft 365 Copilot into a single application, saying personal and professional uses of AI increasingly overlap and that the split product line had become too fragmented. Starting August 18, 2026 the company is retiring Group Chats, AI generated podcasts, Copilot Labs experiments, the Mico animated character, and Deep Research, which is replaced by a Researcher tool for paying professional users. Files from the standalone app will migrate to Microsoft's cloud storage. An internal memo from executive vice president Jacob Andreou said each feature had to earn the right to exist, echoing similar consolidations by Anthropic and OpenAI.
07
Financial Times ↗
Swedish legal AI company Legora is in talks to raise new funding at a valuation above $10 billion, roughly double the $5.6 billion it was worth only four months ago in a round led by Accel. The company crossed $100 million in annual recurring revenue in April, about 18 months after launching, and now serves some 100,000 users across more than 1,200 law firms and legal teams in over 50 markets. The raise lands alongside reports that rival Harvey is negotiating at about $15.5 billion, putting a combined $25.5 billion of paper value on the two leading legal AI platforms. Investor appetite reflects how quickly professional services firms have moved artificial intelligence tools into billable workflows.
08
TechRepublic ↗
Apple is negotiating multiyear licensing deals with news publishers to supply current news and information to its rebuilt Siri assistant, according to a Wall Street Journal report that Apple declined to comment on. The structure under discussion would pay publishers variably based on how often their content is actually used, with a budget reportedly reaching into nine figures. The new Siri combines web retrieval with onscreen awareness, personal context and general world knowledge to answer questions across almost any topic. Developer testing is already live across the version 27 releases of Apple's phone, tablet, Mac and headset operating systems, with a public beta planned for later in 2026, though Apple has not said how publisher sources will be cited or linked in answers.
09
Unite.AI ↗
Israel's National AI Directorate published the implementation plan for Government Resolution 4255 on August 13, committing the country to a national computing base of at least 100,000 advanced AI accelerators over five years, starting with a dedicated 5,000 chip buildout. The plan also targets a domestically built quantum computer, a chip fabrication facility at 5 nanometers or below, a dedicated energy strategy including geothermal supply, and a feasibility study on orbital AI computing. Rather than compete on foundation model scale, the strategy pushes deployment into healthcare, education and government, with a new national institute running five regional acceleration reactors and an upskilling goal of four million people. Regulators are being told to apply a right of way principle, under which a task an AI system performs better than a human should not be blocked unless a specific material risk is proven, and public consultation on the plan closes September 6, 2026.
10
Fortune AI ↗
Self described tech prankster Riley Walz coordinated 50 people to summon Waymo robotaxis to the same dead end San Francisco street at once, gridlocking the block and forcing the company to suspend service there until the following morning. The stunt worked as a crude denial of service attack, showing that the fleet had no defense against a coordinated surge of ride requests. California already requires autonomous vehicle operators to prove they can monitor, update and maintain fleets in line with federal vehicle cybersecurity guidance, yet the prank succeeded anyway, and neither Waymo nor the state DMV responded to requests for comment. Safety researchers argue cybersecurity should be treated as a primary vehicle safety system rather than an afterthought, noting that Arizona and Michigan have written it into their autonomous vehicle rules.