Guozhen AIGlobal AI field notes and model intelligence

Daily AI Brief — August 13, 2026

August 13's AI news centered on the flagship model race, agent deployment, and compute infrastructure. DeepSeek shipped V4 Pro in full via API while SpaceXAI's Grok 4.6 topped rival benchmarks at aggressive prices, and capital flowed into enterprise AI, AI coding, edge chips and embodied-data startups. On governance, OpenAI, Anthropic and Google all committed to watermarking AI-generated content, moving provenance from debate toward industry consensus.

17 Items3 Models4 Agents17 Sources
#01Models

DeepSeek releases V4 Pro official version, taking on Fable 5

On August 13, QbitAI reported that DeepSeek has released the official version of V4 Pro, with the full-capability model now directly available through the API with no staged rollout.The release is positioned against Fable 5 on multiple capabilities, signaling a head-on challenge to the current flagship leader.For developers, the complete flagship model is now the default option when calling DeepSeek, making integration costs and migration paths clearer.

DeepSeek V4 Pro正式版发布,调用API即可直接用上完全体
#02Models

SpaceXAI releases Grok 4.6, topping GPT-5.6 Sol and Fable 5 Max on agent benchmarks

SpaceXAI released Grok 4.6 on August 13, with official benchmarks showing it beating GPT-5.6 Sol and Fable 5 Max on GDPVal-AA v2 and other agentic tests.Priced at $2 per million input tokens and $6 per million output tokens, it undercuts rival flagship models while being tuned for long-horizon agent tasks, and it is live in Grok Build, Cursor, Grok Bot and the API.

#03Models

Claude clears all Hadamard matrices below order 2000, striking another open math problem off the list

QbitAI reports that Anthropic's Claude has cleared all Hadamard matrices below order 2000, removing a long-standing open problem in combinatorics from the math to-solve list.The report stresses that the breakthrough hinged on the mathematician's approach rather than the model choice, with Claude serving as the reasoning tool.The result shows large models can contribute to serious mathematical research, moving AI from a problem-solving tool toward a participant in mathematical discovery.

#04Agents

Huawei Cloud and Dipu Technology launch data intelligence solution to accelerate industry agents

Huawei Cloud and Dipu Technology have jointly released a data intelligence solution aimed at accelerating the deployment of “industry agents” across vertical sectors, combining cloud-side capabilities with data intelligence products.For such agents to run in production, they depend on high-quality, trustworthy data foundations, and the partnership targets data supply and governance in industry scenarios.The release reflects that industry agents are moving from proof-of-concept to scaled delivery, with cloud vendors now selling AI as full solutions rather than single models or raw compute.

#05Agents

Maverick Payments integrates Findustry AI agent to automate chargeback rebuttals

Payment processor Maverick Payments has integrated Findustry's AI agent to automate chargeback rebuttals, according to PYMNTS.Chargeback rebuttals require evidence collection, rule matching and strict deadlines, making them among the most labor-intensive parts of payments operations.The integration promises merchants faster dispute responses and lower operational burden, adding a concrete example of agents deployed inside payment workflows.

#06Agents

Automattic's AI contacts app Mesh comes to Android as “a CRM for everyone”

TechCrunch reported on August 12 that Mesh, Automattic's AI-powered contacts app and relationship manager, has arrived on Android, positioned as “a CRM for everyone.” The launch puts a consumer-grade personal CRM in front of Android's massive user base and completes the product's cross-platform coverage.AI apps are increasingly moving into high-frequency but long-neglected surfaces like contacts and calendars, and Mesh's Android debut is another signal of that trend.

Automattic的AI联系人应用Mesh登陆Android,主打“面向所有人的CRM”
#07Business

OpenAI-backed Thrive Holdings raises $2B at $12B valuation to bring AI to the enterprise

Thrive Holdings, an OpenAI-backed company focused on bringing AI to the enterprise, has raised $2 billion at a $12 billion valuation, with SoftBank, D1 Capital Partners and Alitmeter Capital among the investors.The round lands as enterprise AI moves from pilots to scaled deployments, showing capital still flows heavily to companies that can bridge frontier models and production workloads.How the money is spent, and whether Thrive converts its valuation into durable revenue, is the key question.

OpenAI支持的Thrive Holdings融资20亿美元,估值120亿美元押注企业级AI
#08Business

AI coding startup Cognition reportedly in talks to raise at $40B valuation

TechCrunch reports that AI coding startup Cognition is reportedly in talks for a new funding round at a valuation of around $40 billion, just months after it raised $1 billion at a $26 billion valuation.If confirmed, the jump would underscore investors' continued appetite for AI coding startups.The talks have not been officially confirmed, and final terms could still change.

传 AI 编程公司 Cognition 正洽谈新一轮融资,估值或达 400 亿美元
#09Business

Edge AI chip startup Acrab raises $130M Series B as GΞLIX 1 heads to production

Acrab, a Singapore-based AI compute infrastructure company, has closed a $130 million Series B with Vertex Growth among the investors, bringing cumulative funding past $480 million.Founded in 2024, the company has moved its first edge AI chip GΞLIX 1 and the Agent Box personal AI center into customer onboarding and scaled production preparation.Acrab's bet is that agentic AI pushes high-frequency, privacy-sensitive tasks to the edge, where devices need a new compute foundation rather than just a faster NPU.

#10Business

Embodied-data startup SCALEFORCE closes two funding rounds in 40 days

Embodied-intelligence data infrastructure startup Yuanpoint Technology (SCALEFORCE) has completed a new funding round worth tens of millions of yuan, its second within 40 days, with investors including Hengxu Capital, Kailian Capital and a top domestic embodied-AI industry player.The proceeds will fund its MatrixOS physical-AI operating system, a large-scale data production network, and team expansion.Domestic embodied-AI funding reached 93.5 billion yuan in the first half of 2026, up roughly five-fold year over year, as data becomes the core asset of the physical-AI race.

具身数据公司元点科技 40 天两轮融资数千万,押注物理 AI 基础设施
#11Business

Honor unveils Robot Phone, calling it the world's first robot smartphone

Honor has unveiled the Robot Phone, which it calls the world's first robot smartphone, positioning it as leading the industry into the era of embodied interaction — the first time the phone category has been built around “robot” as its core concept.With embodied AI a hot industry theme, embedding robot-like sensing and action into the most mainstream device form is a distinctive product strategy.Details on features, configuration, availability and pricing remain to be announced.

荣耀发布全球首款机器人手机Robot Phone,智能手机迈入具身交互时代
#12Business

360 and Knowledge is Power magazine launch “Zhili Nano” AI film platform for science content

Knowledge is Power magazine and 360 Technology Group have launched the “Zhili Nano” science and sci-fi AI blockbuster creation platform, built on 360's nano film pipeline.The platform brings AI film generation into science communication, a field long dependent on text and conventional video production, potentially lowering both the barrier and cost of science video creation.The launch event also featured AI creation exchange activities, signaling an attempt to build a content ecosystem around AI-assisted science creation.

#13Business

Lenovo posts record Q1 revenue of 183.4B yuan, up 43%, on AI server boom

Lenovo reported record fiscal Q1 revenue of 183.4 billion yuan, up 43% year over year, its best-ever quarter for the period, with the AI server business entering an explosive growth phase as the core driver.With global AI compute demand surging, cloud vendors and enterprise and government customers are scaling data center investments, and Lenovo's AI server orders have ramped sharply.The results signal an accelerating shift from traditional PC maker to AI infrastructure provider; investors will watch margins and order sustainability.

#14Policy & Governance

At Ai4, Hinton, Fei-Fei Li and Ng make the case for staying open amid safety concerns

At the Ai4 conference, Geoffrey Hinton, Fei-Fei Li and Andrew Ng debated regulation, open-source access, and how the U.S.can stay competitive as China advances across Asia.With safety concerns mounting, the three pioneers made the case for staying open, arguing that openness remains the better path even as risk discussions intensify.

AI4大会:辛顿、李飞飞、吴恩达同台,呼吁在安全焦虑中保持开放
#15Policy & Governance

Claude users push back on Anthropic's watermarks, fearing detection at work and school

Anthropic's new watermarking system for Claude is drawing backlash on social media, with some users complaining it will expose their AI use at jobs and in classes, according to TechCrunch.Watermarking embeds imperceptible markers in AI-generated text so external tools can identify machine-made content, which could flag AI-assisted homework or work products.Supporters say it improves transparency and curbs cheating, while critics worry it hurts ordinary users; AI content provenance is moving from policy debate into shipped product behavior.

部分 Claude 用户不满 Anthropic 新水印:担心工作与课堂上使用 AI 被抓个正着
#16Policy & Governance

OpenAI, Anthropic and Google sign on to watermark AI-generated content, reports say

Multiple reports say OpenAI, Anthropic and Google have signed commitments to watermark AI-generated content, making synthetic output easier to identify.Coordinated action by the biggest labs would make watermark systems far more effective than today's scattered, uneven efforts.For audiences, synthetic content becomes easier to recognize and trust improves; for platforms and publishers, labeling AI content is shifting from voluntary experiments toward industry consensus.

#17Infrastructure

Report: SpaceX races toward 10GW as AI compute war enters Musk's “comfort zone”

Odaily reports that SpaceX is racing toward a 10-gigawatt compute target, arguing the AI compute war is entering Musk's “comfort zone.” Ten gigawatts would dwarf today's hyperscale clusters, which mostly operate in the hundreds of megawatts, requiring an overhaul of power, grid, land and cooling infrastructure.The report's core argument: the compute race is shifting from chips to electricity and infrastructure, where Musk's large-scale engineering experience becomes a competitive edge.

报道:SpaceX 冲刺 10GW 算力规模,AI 算力战争进入马斯克“舒适区”

More Daily Reports

All Daily Reports