AI News

Latest updates from the world of artificial intelligence

HIGGSFIELD $5.4B $400M Series B • $700M run rate BITSMINDS.COM
Companies

Higgsfield Raises $400M at $5.4B, 4× Its Series A

The AI video startup says annualized revenue hit $700 million this month, up from $20 million a year ago. DST Global led the $400 million Series B, with Goldman Sachs Alternatives and Intel Capital joining.

PR Newswire

Read more →
GEMINI 3.7 Flash 34.4% → 43.6% FRONTIERCODE 1.1 BITSMINDS.COM
Models

Gemini 3.7 Flash Ships 3 Weeks After 3.6, at Half Price

Google shipped Gemini 3.7 Flash just three weeks after 3.6 Flash, claiming a jump from 34.4% to 43.6% on FrontierCode and from 49% to 65.3% on DeepSWE — and priced the introductory tier at half what 3.6 Flash cost.

Google

Read more →
An AI Store Manager Recommended Firing a Human
Industry

An AI Store Manager Recommended Firing a Human

Luna, the Claude-based agent running a San Francisco shop, parted ways with a worker late for 17 of 23 shifts — but only after its creators reminded it of the attendance policy it had written months earlier and then forgotten.

The Next Web

Read more →
ACQUIRED FOR $7B+ BITSMINDS.COM
Companies

Stripe Buys OpenRouter for $7B, 5× Its May Price

Bloomberg reports the payments company has finalized a deal for the AI gateway that routes 25 trillion tokens a week across 400-plus models — three months after OpenRouter raised at a $1.3 billion valuation.

TechCrunch

Read more →
SPACEX $21B INTEL $30B $63.4B IN 8 POSITIONS BITSMINDS.COM
Industry

Nvidia's $63B Stock Portfolio Is Mostly Its Own Customers

A routine 13F filing revealed Nvidia holds $21 billion of SpaceX and $30 billion of Intel — 80% of an eight-position book, in two companies that have both committed to buying its chips.

The Next Web

Read more →
BREAK EVEN $11.5B Q2 2026 REVENUE BITSMINDS.COM
Companies

Anthropic Posts Its First Operating Profit on $11.5B

Preliminary Q2 revenue came in above $11.5 billion — 14× a year ago — and Anthropic says it booked positive adjusted operating income for the first time since it was founded. The asterisk is on the word “adjusted”.

CNBC

Read more →
AI GENERATED AI GENERATED AI GENERATED WATERMARK NOW OPTIONAL GEMINI · FLOW SYNTHID STAYS BITSMINDS.COM
Products

Google Makes Its Visible AI Watermark Optional

Users can now switch off the visible watermark on images and video from Nano Banana, Omni and Lyria. The invisible SynthID signal and C2PA metadata stay embedded either way — and Google is open-sourcing a library so anyone can check them.

TechCrunch

Read more →
31.4% ON 50K TOKENS GLM-5.3 SAME 743B BASE BITSMINDS.COM
Models

GLM-5.3: 6× on Terminal-Bench, Same 743B Base Model

Z.ai shipped GLM-5.3 on August 14 without retraining its base model. Every gain comes from scaled post-training — and on the company’s own coding benchmark it edges past Claude Opus 4.8 while spending roughly 60% fewer output tokens.

MarkTechPost

Read more →
Claude Opus 5 vs Fable 5 vs Sonnet 5: One Prompt Each
ModelsLab

Claude Opus 5 vs Fable 5 vs Sonnet 5: One Prompt Each

We gave Claude Opus 5, Fable 5 and Sonnet 5 the same three build briefs — an animated SVG fairground, a falling-sand physics sandbox and a self-solving Rubik’s cube — one run each, no retries, no browser. All nine builds run live inside the article. Opus 5 takes it 8–6–3, and the biggest surprise is the clock: the largest model was the fastest on every round, by a factor of nearly three.

BitsMinds Lab

Read more →
$2B DEAL, UNWOUND BITSMINDS.COM
Industry

Manus Goes Independent as Meta's $2B Deal Unwinds

Beijing ordered the acquisition withdrawn in April. The separation is now executing — and the bill lands on users, who have until August 23 to download data that will otherwise be deleted.

South China Morning Post

Read more →
LOVABLE AT $13.3B BITSMINDS.COM
Companies

Lovable Raises $400M at $13.3B, ARR Nears $600M

The Stockholm vibe-coding company doubled its valuation in eight months on a revenue line that nearly tripled over the same stretch. At roughly 22 times forward revenue, it is one of the more defensible prices in AI coding — which says more about the comparison set than about Lovable.

TechCrunch

Read more →
$0.87 / 1M OUTPUT V4-PRO-0813 GENERALLY AVAILABLE BITSMINDS.COM
Models

DeepSeek Shipped V4-Pro Without Telling Anyone

No blog post, no changelog, no press release — just an edit to the API pricing page. DeepSeek-V4-Pro-0813 is now generally available at $0.87 per million output tokens, with benchmark gains that nobody outside DeepSeek has reproduced.

OpenRouter

Read more →
LEGAL AI VALUATION $1.8B $5.6B $10B+ OCT '25 MAR '26 AUG '26 BITSMINDS.COM
Industry

Legora Seeks a $10B Valuation, Double March’s Price

The Swedish legal AI startup is raising again at more than $10 billion, months after closing at $5.6 billion. At roughly $100 million of ARR, that is about 100 times revenue — and Legora is still the cheaper of the two leaders.

The Next Web

Read more →
$6B FOR DECART BITSMINDS.COM
Companies

Anthropic’s Biggest Deal Yet: $6B for Decart

Anthropic is in early talks to buy Israeli startup Decart for about $6 billion — roughly 50% above the valuation Decart set three months ago. The prize is not the video models everyone knows it for. It is the chip-efficiency software underneath them.

Fortune

Read more →
COGNITION · DEVIN · IN TALKS AT A $40B VALUATION BITSMINDS.COM
Industry

Cognition in Talks at $40B, Three Months After $26B

Bloomberg reports the Devin maker is in early talks on a round valuing it at $40 billion or more, up from $26 billion in May. Annualized revenue is approaching $1 billion, double the figure at the last raise — which means the valuation is rising while the revenue multiple compresses from about 53x to 40x.

Bloomberg / TechCrunch

Read more →
SPACEXAI Grok 4.6 BITSMINDS.COM
Models

Grok 4.6 vs Opus 5 vs GPT-5.6 Sol: Frontier, 60% Cheaper

Artificial Analysis scored Grok 4.6 at 61 on its Intelligence Index — level with GPT-5.6 Sol at max, behind Fable 5 at 62 and Opus 5 at 63. It costs $2/$6 per million tokens against $5/$25 and $5/$30, and finishes long agentic tasks in ~53 turns where Opus 5 takes ~103. It also trails both rivals by eight points on the hardest software-engineering evals.

Artificial Analysis

Read more →
RIVER AI · $1.1B SEED ROUND · OPEN AI STACK BITSMINDS.COM
Companies

River AI Raises $1.1B Two Months After Leaving Stealth

General Catalyst and AMP PBC led a $1.1 billion first round into Igor Babuschkin’s nine-week-old startup, with Nvidia, AMD Ventures, Y Combinator and Temasek alongside. River sells an API for reinforcement-learning fine-tuning of open-weight models — a complex run in 15 to 20 minutes, it claims, at two to four times lower cost than closed alternatives.

TechCrunch

Read more →
PIXEL 11 · TENSOR G6 · GEMINI INTELLIGENCE BITSMINDS.COM
Products

Pixel 11’s 2nm Tensor G6 Runs On-Device AI 3.5× Faster

Google built the Tensor G6 on TSMC’s 2nm N2 process and spent almost all of it on inference: 50% more TPU compute, double the memory bandwidth, and on-device AI tasks up to 3.5× faster at up to 3.5× less energy. Gemini Intelligence now ships on hardware designed for it — agentic task automation, phone calls placed on your behalf, and on-device dubbing for podcasts and video.

9to5Google

Read more →
1B MONTHLY USERS BITSMINDS.COM
Products

Gemini Hits 1 Billion Users, Two Months After ChatGPT

Sundar Pichai says the Gemini app has passed a billion monthly active users, up from about 950 million in July and roughly two months behind ChatGPT. Voice is now the interface for 63 percent of users, and the app generates over 150 million images a day. The figure counts the standalone app only — not Gemini inside Search, Workspace or Android.

TechCrunch

Read more →
WATERMARKED BITSMINDS.COM
Companies

Anthropic Watermarks Claude’s Text — Worldwide, Not Just EU

Text from every Claude model released after 2 August now carries an imperceptible watermark, and generated files ship with signed C2PA metadata. Article 50 of the EU AI Act forced it; Anthropic is applying it everywhere Claude is sold. The catch: a mark proves the text passed through Claude, not that Claude wrote it — and the tools to check one are not public yet.

TechCrunch

Read more →
A 160-Year-Old Bound, Moved Share of Riemann zeta zeros proven to satisfy the hypothesis 41.6% previous best 67.2% Claude's result 0% 100% 650 ideas that failed ~60 subagents 31M output tokens Lean formally verified BITSMINDS.COM
Research

Claude Raised a Riemann Zeta Bound From 41.6% to 67.2%

An unreleased research version of Claude combined two existing results to lift the proven share of Riemann zeta zeros satisfying the hypothesis from 41.6% to 67.2%. It took 650 failed ideas, about 60 subagents, 31 million output tokens and a Lean proof — and Anthropic says it is not a route to proving the hypothesis itself.

Anthropic

Read more →
Anthropic Gets the Data Centers. Someone Else Owns Them. Macquarie AM + GIC own the platform equity THESEUS INFRASTRUCTURE builds, operates, leases lease Anthropic anchor tenant No capital figure, no gigawatts, no sites, no timeline disclosed BITSMINDS.COM
Companies

Anthropic Gets New Data Centers It Won't Own

Macquarie Asset Management and GIC will own Theseus Infrastructure, a new platform created to build and lease data centers to Anthropic. Anthropic is the anchor tenant, not the owner — and no capital figure, site list or timeline was disclosed.

Macquarie Group

Read more →
ANTHROPIC · COMPUTE Claude's new landlord mines bitcoin 191 MW by June 2028 $9.1B base, 20-year term 96 MW first power, Dec 2027 ROCKDALE, TEXAS Riot Platforms · 20-year lease to June 2048 · tenant reported by Bloomberg, unconfirmed BITSMINDS.COM
Companies

Anthropic Leases 191 MW From a Bitcoin Miner for $9.1B

Riot Platforms signed a 20-year, 191-megawatt data-center lease worth $9.1 billion at its Rockdale, Texas campus. Riot did not name the tenant; Bloomberg reported it is Anthropic.

The Block

Read more →
OPENAI · DAYBREAK PROGRAM Two tiers, one very sharp model DAYBREAK BLUE GPT-5.6 Sol, safeguards tuned for authorized defensive work 2.0% advanced cyber completion DAYBREAK RED GPT-5.6-Cyber, for exploit validation under monitoring 95.0% advanced cyber completion Unmodified GPT-5.6 Sol: 1.5% · Last year's GPT-5.5-Cyber: 57.3% · Rated High, below Critical BITSMINDS.COM
Models

OpenAI’s GPT-5.6-Cyber Found Two Chrome Zero-Days

OpenAI split its Daybreak security program into Blue and Red tiers and handed the Red tier a purpose-trained model that clears 95% of advanced cyber tasks — against 1.5% for the same frontier model with its guardrails intact.

The Decoder

Read more →