科技热门推文——2026年8月29日

今日科技领域的讨论主要聚焦于人工智能安全、基础设施与自动化：X表示已瓦解一个由20万个账号组成的中国影响力网络，多家领先人工智能实验室和云服务提供商则呼吁加强全球网络防御。开放权重模型以及面向操控科学硬件的智能体标准相继出现，凸显工程技术的快速进展；另据报道，OpenAI正考虑利用人形机器人参与数据中心建设。与此同时，美国或将收紧芯片获取规定，安德森·霍洛维茨（Andreessen Horowitz）推出规模达11亿美元的“机器时代基金”，重点投资芯片、网络、电力、机器人及系统软件领域。

1. Polymarket (Group Score: 545.9 | Individual: 47.7)

Cluster: 26 tweets | Engagement: 22369 (Avg: 1879) | Type: Tech

BREAKING: 𝕏 reveals it uncovered a Chinese bot farm of roughly 200,000 accounts, including hundreds used to influence U.S. debate over AI data centers & energy policy.

See 25 related tweets

@perrymetzger: RT @GlobalAffairs: The X Safety team conducted an investigation into suspected Chinese inauthentic a...
@XFreeze: 𝕏 just uncovered a massive Chinese bot farm of roughly 200,000 accounts

Inside that network, 𝕏 iden...

@nic_carter: If you’re participating in the anti datacenter moral panic you are either
a literal Chinese bot ...
@aakashgupta: X found 200,000 Chinese bot accounts, including hundreds built to turn Americans against AI data cen...
@TaylorLorenz: As I was just saying, grifters and bad actors are currently flooding platforms with anti AI slop in ...

2. Mononofu (Group Score: 292.2 | Individual: 41.9)

Cluster: 13 tweets | Engagement: 3880 (Avg: 700) | Type: Tech

RT @OpenAI: We have a limited window to strengthen cyber defenses, and together with organizations including @AnthropicAI, @awscloud, @Google, @Microsoft, and @Oracle, we're calling for a global effort to give defenders the tools, resources, and support to protect the infrastructure we all depend on.

If we act decisively, we can turn today's AI advances into lasting improvements in security and make our digital world safer for everyone.

https://t.co/f33JRVCiJb

See 12 related tweets

@VaibhavSisinty: OpenAI wrote a warning letter about AI-enabled cyberattacks. Anthropic, Google, Microsoft, and 100+ ...
@godofprompt: For the first time in history, OpenAI, Anthropic, and 100+ other companies in the sector just asked ...
@yacineMTB: RT @sama: this is a critically important moment for cyber defense with AI; there is not much time to...
@jpatel41: 👇🏾\n\nQT @ChuckRobbins: I've said for years that we can create a stronger cybersecurity ecosystem by...
@pstAsiatech: Today’s AI advances are already giving defenders new ways to fix weaknesses that have accumulated fo...

3. Nas_tech_AI (Group Score: 281.5 | Individual: 40.4)

Cluster: 9 tweets | Engagement: 168 (Avg: 45) | Type: Tech

Tencent Hy’s pace of evolution is getting wild.

Three major releases in just six months — basically a big move every two months.

The new Hy4 preview brings major upgrades in model size, context length, and training data. With both pre-training and post-training getting serious attention, it’s clearly pushing into the top tier of open-source models.

What I like most is the “preview first, iterate fast” approach, with real-world feedback directly driving development.

Very curious to see what comes next. @TencentAI_News\n\nQT @TencentAI_News: Introducing Hy4 preview

◆ Open weights: 770B MoE, 49B active, 1M context ◆ Built for real work: code, docs and analysis, scientific research ◆ Hy4 preview helped optimize its own training pipeline, kernels included ◆ We ship upgrades, not price hikes

Free in @WorkBuddy_AI for the next two weeks

Blog：https://t.co/aooJ4zvF7c GitHub：https://t.co/dPAu5TkRhN Get WorkBuddy：https://t.co/bsGBeiE5ox

See 8 related tweets

@scaling01: RT @TencentHunyuan: 🚀 Hy4 preview is here.

770B, 49B active, 1M context. Built for productivity. Op...

@haider1: YOU WAKE UP

and there's somehow a new chinese open-weight model every day

today it's Tencent's Hy4...

@jenzhuscott: The @TencentHunyuan team kept their promise - Hy4 Preview is here!\n\nQT @TencentAI_News: Introducin...
@kimmonismus: Tencent just released Hy4 Preview: a 770B-parameter open-source model with 49B active parameters and...
@testingcatalog: Tencent released Hy4 Preview, a new open weight model under Apache License 2.0!

> Hy4 preview i...

4. MTSlive (Group Score: 234.1 | Individual: 32.4)

Cluster: 12 tweets | Engagement: 32 (Avg: 152) | Type: Tech

SITUATION EXPLAINED: The Trump administration is closing the loophole that lets Chinese labs rent US chips.

• A small group inside Commerce is building a slimmed-down replacement for the Biden-era AI Diffusion Rule, which the administration scrapped in 2025 • The target is Chinese AI firms accessing advanced chips remotely through data centers abroad, in countries including Thailand and Singapore • The current regime is upside down: the US now permits H200 sales to China while Beijing forbids its own firms from buying them • Nvidia's share of China's AI chip market fell from 95% in 2023 to 8% in 2026, while Huawei climbed to roughly 50-60%

@theojaffee: "It's kind of unclear to me how important access to American chips is for Chinese AI progress, especially if GLM 5.3 is served entirely on Chinese chips, which is what they claim."

See 11 related tweets

@perrymetzger: I fully understand the motivation, but it's not going to make much difference soon. The Chinese are ...
@wallstengine: TRUMP ADMINISTRATION WORKING ON NEW AI CHIP EXPORT RULE

The Commerce Department is developing a pol...

@MTSlive: SITUATION EXPLAINED: Would China be ahead on AI if it could buy US chips?

• The controls made a rea...

@WOLF_Financial: TRUMP ADMINISTRATION WORKS TO CLOSE CHINA’S REMOTE AI CHIP ACCESS LOOPHOLE

The Commerce Department ...

@theinformation: Exclusive: The Trump administration is developing new AI chip restrictions aimed at stopping Chinese...

5. Scobleizer (Group Score: 213.6 | Individual: 37.1)

Cluster: 10 tweets | Engagement: 1364 (Avg: 182) | Type: Tech

RT @Zai_org: GLM-5.3 is now open-weight.

Our most capable model for agentic coding and cyber defense is now available to download, run, and customize.

Weights: https://t.co/v1IbWMXxg4 Tech blog: https://t.co/ekQkO83jCv https://t.co/f8XlJksKyf

See 9 related tweets

@OpenRouter: GLM-5.3 from @Zai_org is now open-weight and live on OpenRouter.

Built for complex software enginee...

@Mayhem4Markets: GLM-5.3 is now open-weight! This is going to make for a rather exciting weekend in the world of loca...
@appliedcompute: We are excited to support GLM 5.3 for training and inference on the Applied Compute Platform.

Post-...

@pingToven: more providers coming to openrouter today!\n\nQT @Zai_org: GLM-5.3 is now open-weight.

Our most cap...

@0xSero: Top open weight model runs in home at the price of a used car\n\nQT @Zai_org: GLM-5.3 is now open-we...

6. Marktechpost (Group Score: 193.4 | Individual: 47.9)

Cluster: 8 tweets | Engagement: 2108 (Avg: 176) | Type: Tech

RT @AnthropicAI: Today, we're kicking off the first phase of the research preview for Model Hardware Standard (MHS): a new standard for AI agents to safely operate physical equipment in scientific research and advanced manufacturing.

See 7 related tweets

@boxcardavid: I couldn’t overstate how important Claude and Codex have been for setting up manufacturing in the US...
@adocomplete: It’s like MCP, but for hardware! Very exciting.\n\nQT @AnthropicAI: Today, we're kicking off the fir...
@Gorden_Sun: Anthropic之前定义了Agent（让LLM调用工具自主工作）、MCP（让Agent能调用其他软件）、Skill（让Agent按SOP执行工作流），现在又定义了MHS，模型硬件标准，让Agent能...
@cryptopunk7213: of all the anthropic releases so far this is my favorite.

you can build autonomous scientific labo...

@andersonbcdefg: what's next the AI Poop Protocol so that we have a standardized schema to describe the size, color, ...

7. a16z (Group Score: 187.0 | Individual: 41.9)

Cluster: 6 tweets | Engagement: 801 (Avg: 439) | Type: Tech

Today, a16z is announcing the Machine Age Fund, a new $1.1 billion fund for founders rebuilding what intelligence runs on: chips, memory, networking, systems software, power, and the machines that bring AI into the physical world.

Ben Horowitz, Martin Casado, and Raghu Raghuram see a new economic law. A thousand engineers cannot erase a two-year software lead, but a massive GPU cluster can turn capital directly into capability. Models are improving faster than the memory, interconnect, power, and cooling beneath them.

The founder map is changing with it. Some of the strongest teams are moving from pure software into complex hardware because every constraint in the stack is now a company-building opportunity.

In this conversation with Erik Torenberg, they explain what founders can build, why the opportunity reaches all the way down to the physical stack, and why a16z created a fund for it.

00:00 Intro 01:08 Why AI needs an entirely new infrastructure 02:02 The bottleneck is no longer the model 02:40 Founders saw it before investors did 06:07 Sold out through 2028 07:14 Why this time is different from 1999 08:57 The company whose idle hardware gained value 10:51 "Why didn't this fund exist five years ago?" 13:51 "Nobody likes to use AI more than AI" 17:06 The startup law that capital just broke 21:05 What Grok Bot got right 26:43 The case for a custom chip per model 28:23 Why AC power isn't good enough 32:07 Lots of new electricians 34:14 What one gigawatt can power 35:03 Why utilities can't just build faster 37:22 The data centers that give power back 38:01 Why we should never have called it AI 39:43 Why Nvidia will willingly leave money on the table 48:08 Hardware founders are older 52:36 Why America has to lead

YouTube: https://t.co/5QdtZyFJ9T

@bhorowitz @RaghuRaghuram @martin_casado @eriktorenberg\n\nQT @a16z: https://t.co/rohq8Li7GK

See 5 related tweets

@a16z: Ben Horowitz on the oldest law in startups, and why the industry no longer has to obey it:

"The one...

@a16z: Introducing the Machine Age Fund, $1.1 billion for the physical buildout of the AI stack:

Ben Horow...

@martin_casado: RT @a16z: Today, a16z is announcing the Machine Age Fund, a new $1.1 billion fund for founders rebui...
@MTSlive: SITUATION EXPLAINED: a16z raised a $1.1 billion fund for AI hardware.

• The Machine Age Fund, from ...

@MTSlive: a16z's @jkhamehl reveals the thesis behind its $1.1B Machine Age Fund: everything below the software...

8. garrytan (Group Score: 145.5 | Individual: 31.8)

Cluster: 7 tweets | Engagement: 479 (Avg: 263) | Type: Tech

The Tampa Bay Wave of influential people in AI\n\nQT @venturetwins: Objectively insane to make a list of the most influential people in AI and include these two but not Jensen https://t.co/WyL7QOf6vO

See 6 related tweets

@stevehou: “We are all agreed that your list is crazy. The question which divides us is whether it is crazy eno...
@exec_sum: NEWS: Time has released its list of the 100 most influential people in AI that includes industry lea...
@wallstengine: RT @wallstengine: TIME just released its 2026 list of the 100 most influential people in AI.

Someho...

@LaurenGoode: lol well-played @tbpn\n\nQT @tbpn: BREAKING: TIME Magazine declares that Jensen Huang is NOT one of ...
@NYSE: Yesterday, @TIME released the 2026 TIME100 List of the World’s Most Influential People in Artificial...

9. madhavjha (Group Score: 144.8 | Individual: 57.7)

Cluster: 4 tweets | Engagement: 2272 (Avg: 209) | Type: Tech

RT @patrick_oshag: Neil Movva (@neilmovva) started his career at Nvidia, working on GPUs and kernels, and has an unusually deep understanding of inference, from software to chips to power.

We spend a lot of time on each of those layers, how they connect, and where the important tradeoffs are.

What makes this conversation special is how detailed it is (like a 401-level class), yet Neil makes it remarkably clear and easy to follow.

Today he runs Sail Research, a company building infrastructure for agents to make tokens as cheap as possible.

We discuss:

Latency versus throughput
Why there are no bad chips, only bad pricing
The end of kernel engineering
Buying chips and power no one else wants
New chip architectures
Nvidia lore + his contrarian view of the company
Open source and the frontier labs

I learned a ton. Enjoy!

TIMESTAMPS 0:00 Intro 0:38 Building a “Token Factory” 4:21 The Future of Background Agents 13:09 Nvidia and the GPU Stack 23:27 Chips, Memory, and Transformers 36:14 The Future of AI Training Data 44:32 Chip Scarcity and Compute Arbitrage 52:44 Reinventing the AI Data Center 59:01 Power and the “Scavenger Strategy” 1:10:10 Open vs. Closed AI

See 3 related tweets

@himanshustwts: One of the best conversations I have seen in a while. On culture at Nvidia, Neil mentions:

“They ar...

@patrick_oshag: Thanks for making this one an absolute monster\n\nQT @patrick_oshag: Neil Movva (@neilmovva) started...
@TheFuturumGroup: NVIDIA’s outlook points to another year of significant AI infrastructure growth, even with supply li...

10. jaltma (Group Score: 133.1 | Individual: 36.4)

Cluster: 4 tweets | Engagement: 258 (Avg: 407) | Type: Tech

Owner is the first board I ever joined after leading their A, and Adam is the founder who made me realize I wanted to be a full time investor.

Incredible execution fueled by genuinely loving their customers. So proud of these guys and they are just at the beginning.\n\nQT @adamguild: Owner has raised $240M, reached a$ 2.3B valuation, and crossed $100M ARR.

Goldman Sachs Alternatives led the round, joined by Meritech, Redpoint, Headline, & Jack Altman.

The world is racing to build AI to replace people’s jobs. Owner is building AI to do the opposite: https://t.co/U7xRIa53Cm

See 3 related tweets

@dharmesh: Congrats to the Owner team.

Amazing progress and love the focus on SMBs.

Honored to be an indie i...

@Austen: I was bullish on Adam when I first met him in 2021.

What they’ve built has far exceeded my expectat...

@APompliano: RT @adamguild: Owner has raised $240M, reached a$ 2.3B valuation, and crossed $100M ARR.

Goldman Sa...

11. XFreeze (Group Score: 131.5 | Individual: 37.9)

Cluster: 4 tweets | Engagement: 890 (Avg: 578) | Type: Tech

SpaceXAI has built an entire Grok Bot Guides library showing how people are actually using AI agents as real teammates

These are practical playbooks from people already running Grok Bot for:

• Mobile app development • Product management • Design • Enterprise GTM • Managing multiple teams of agents

And these are not just chatbot workflows

Each Grok Bot can have its own role, memory, cloud computer, connected tools, recurring routines and learned skills, then hand work directly to other specialized bots

One example runs a six-bot mobile game studio across analytics, creative, engineering, infrastructure and bug fixing

Another has a PM running a Chief of Staff, engineering manager, five engineering agents, data analyst, product agent and recruiter

There are even setups where entire projects get their own channel, roster of bots and Notion board, basically organizing AI agents like a human team

This is probably one of the most useful resources SpaceXAI has published for understanding what working with a real team of AI agents actually looks like

You can check out the entire library here: https://t.co/RAaw7VwijP

See 3 related tweets

@XFreeze: Grok Bot just got templates....and this makes it even more powerful

You can now build an entire Bot...

@Scobleizer: The Grok Bot report. Done by an AI agent that @blevlabs and I created. Reads through about a million...
@cb_doge: SpaceXAI has added a step-by-step guide showing how to run multiple teams of Grok Bots.

Learn how e...

12. MTSlive (Group Score: 130.1 | Individual: 48.3)

Cluster: 5 tweets | Engagement: 1493 (Avg: 152) | Type: Tech

SITUATION DETECTED: Sam Altman says that OpenAI plans to build its own humanoid robots, and that their first uses may be to build more robots and then to construct future data centers, per Alex Heath.

See 4 related tweets

@alexeheath: - Sam Altman thinks OpenAI will have AGI by the end of the year.
Its next model, Astra, is trained...
@edzitron: https://t.co/7WFb6cuFEN\n\nQT @MTSlive: SITUATION DETECTED: Sam Altman says that OpenAI plans to bui...
@PolymarketMoney: BREAKING: Sam Altman reveals OpenAI eventually wants to build its “own humanoid.”...
@wallstengine: In Alex Heath’s reporting from inside OpenAI,

Sam Altman said the company eventually wants to buil...

13. sairahul1 (Group Score: 112.8 | Individual: 35.5)

Cluster: 4 tweets | Engagement: 12 (Avg: 65) | Type: Tech

This is INSANE, Anthropic ran its marketing with basically one person.

Austin lau, a non-technical growth lead, was running paid search, paid social, email, and seo solo.

Here’s the workflow:

export ad CSVs into Claude Code AI flags underperforming ads agents generate new headlines + descriptions Figma auto-swaps copy across 100 ad templates MCP server pulls live Meta data

The results:

ad creation went from 2 hours to 15 minutes. total marketing output grew 10×. conversion rates landed 41% above industry average.

One person doing what used to take an entire marketing team.

Now imagine doing the same with your business.

This is the complete guide to automating your entire marketing without hiring anyone ↓

Bookmark this. Start this weekend.\n\nQT @sairahul1: Grok Bot is the best AI agent right now

It gives you an army of agents that can do work for you 24/7

If you set it up correctly, you gain super powers

In this article, I cover how to automate your entire marketing using Grok Bot. https://t.co/wOFTNmKVrp

See 3 related tweets

@sairahul1: founders realising they can automate their entire marketing with Grok Bot https://t.co/kq9ZBTaRHF\n\...
@sairahul1: RT @sairahul1: This is INSANE, Anthropic ran its marketing with basically one person.

Austin lau, a...

@sairahul1: RT @sairahul1: Grok Bot is the best AI agent right now

It gives you an army of agents that can do w...

14. WSJ (Group Score: 99.6 | Individual: 25.4)

Cluster: 6 tweets | Engagement: 292 (Avg: 124) | Type: Tech

Exclusive: Andreessen Horowitz, which backed SpaceX and Anduril, raised $1.1 billion for a Machine Age fund to invest in chips, robots and other hardware as the AI infrastructure bottleneck intensifies https://t.co/C3lRtozxJY

See 5 related tweets

@MTSlive: SITUATION DETECTED: Andreessen Horowitz is launching a $1.1 billion Machine Age fund to address crit...
@business: Prominent venture capital firm Andreessen Horowitz says it has raised $1.1 billion for its newest ar...
@exec_sum: BREAKING: Andreessen Horowitz launches a $1.1B “Machine Age Fund” focused on AI infrastructure &...
@WSJTech: Andreessen Horowitz, which backed SpaceX and Anduril, raised $1.1 billion for a Machine Age fund to ...
@MSBIntel: BREAKING: Andreessen Horowitz launches $1.1 billion Machine Age fund for physical AI infrastructure....

15. rohanpaul_ai (Group Score: 97.9 | Individual: 27.4)

Cluster: 5 tweets | Engagement: 220 (Avg: 46) | Type: Tech

RT @rohanpaul_ai: The AI detector industry should be very uncomfortable reading this MIT report.

The strongest institutional rejections yet, direct from MIT.

"we recommend against relying on AI detectors. "

"As it risks an arms race in which students respond to automated detection by using increasingly powerful “AI humanizers” to remove signals that AI detectors are cued to catch. The result: a lot of effort on both sides that in the end serves no one."

The report argues that mixed human/AI writing is difficult to detect, detectors can be evaded by “AI humanizers,” false positives can harm students, and heavy policing creates distrust.

AI detection systems may also mistake the writing of non-native English speakers or neurodivergent students for text generated by AI. Even low rates of false positives can put students on edge and cause serious individual consequences."

See 4 related tweets

@seconds_0: Like clockwork https://t.co/UtuigEpNoy\n\nQT @rohanpaul_ai: If you need more proof that AI detectors...
@BrianRoemmele: BOOOOM! IT’S OFFICIAL!

MIT HAS BANNED ANY USE OF TEXT BASED AI DETECTION.

I just got off the phone...

@BrianRoemmele: RT @BrianRoemmele: BOOOOM! IT’S OFFICIAL!

MIT HAS BANNED ANY USE OF TEXT BASED AI DETECTION.

I jus...

@BrianRoemmele: RT @grok: Yes. MIT’s Aug 2026 Ad Hoc Committee report recommends against relying on AI detectors. Th...

16. poezhao0605 (Group Score: 96.7 | Individual: 24.9)

Cluster: 4 tweets | Engagement: 30 (Avg: 11) | Type: Tech

Short interest in MiniMax, the Hong Kong-listed Chinese AI lab, hit 20% of free float before its H1 report. A record since listing.

Then the report landed: revenue up 283%, six months of sales beating all of 2025 by nearly half. The stock rose 3.8% the next day.

Someone is badly wrong here. I graded the report to find out who. https://t.co/PaIF8cXKED

See 3 related tweets

@poezhao0605: Record short interest, then a 283% revenue print. The first-half report from MiniMax, the Hong Kong-...
@poezhao0605: The asymmetry: Zhipu's short interest sits at 6%. MiniMax trades at about a quarter of Zhipu's marke...
@poezhao0605: The detail that complicates the bull case: adjusted net loss more than doubled to $293m in the same ...

17. mardehaym (Group Score: 95.9 | Individual: 34.4)

Cluster: 3 tweets | Engagement: 138 (Avg: 55) | Type: Tech

RT @mardehaym: A PE operating partner asked us to build production AI agents inside a portfolio company's billing system, processing real healthcare claims under HIPAA.

Two people hand-wrote every rule in their claims engine across 300+ denial codes and payer logic that changes quarterly.

Four months later, seven production agents handle it with zero patient data exposure.

First month, we didn't touch a model. We mapped their data: where it sits and what's missing, so agents reason from structured facts instead of guessing.

I've watched teams skip this step across dozens of engagements. They bolt a model onto the product, watch it hallucinate over unstructured inputs, and decide AI isn't ready for their industry. The data work is what makes it ready.

We built an enrichment layer that assembles 34 dynamic variables per claim before any LLM sees it, pre-computed and versioned so the agent receives ranked facts instead of searching for context.

Every agent follows one pattern: pre-compute context, strip all patient data before the model sees it, validate output against a strict schema, let deterministic code accept or reject the action. If the output falls outside the allowlist, the system fails closed.

Seven agents, each locked to a single workflow like denied claim follow-up or billing reconciliation, each running its own enrichment payload.

Then we built the eval harness.

Every agent runs against a curated test suite before any update reaches production. When a model provider ships a new version or payer logic changes, the harness catches regression before a single live claim is affected. The flagship agent reconciles denials to the penny: 59 out of 60 on the eval set.

Most teams launch an agent and hope it keeps working. We launch one and prove it does on every deployment.

We route calls across two model providers. Swapping one changes nothing in the output because the eval harness verifies it.

Model integration was the shortest line item in the four-month build.

The operating partner now benchmarks the rest of the portfolio against this system.

That's the line between a portfolio company running AI and one still running demos.

See 2 related tweets

@alex_prompter: great lessons on enterprise AI agents:\n\nQT @mardehaym: Their CPTO expected months before an AI age...
@mardehaym: RT @mardehaym: Their CPTO expected months before an AI agent could handle real claims data.

Six wee...

18. Ric_RTP (Group Score: 94.4 | Individual: 38.1)

Cluster: 3 tweets | Engagement: 293 (Avg: 180) | Type: Tech

Jensen Huang just admitted the biggest AI labs can't borrow money like normal companies.

So Nvidia signs for them, and they spend it on Nvidia chips.

Nvidia reported Wednesday and the numbers are absurd:

Revenue of $96.2 billion, up 106%, with net income of$ 59.7 billion, the most profitable quarter any public company has EVER posted.

And Huang just told Fox Business that every chip Nvidia can make next year is already sold.

Here's why this matters the most:

Huang wrote this himself about his own customers: "Frontier AI labs have extraordinary demand for training and inference compute, but many are growing faster than their balance sheets and long-term credit profiles can support."

Then: They "still lack the decades-long infrastructure contracts and investment-grade financing capacity needed to secure the AI factory infrastructure independently."

Put simply: His customers can't get the loans.

So Nvidia signs for them.

There's a compute campus going up in Ohio with OpenAI as the tenant. Nvidia has tied roughly $105 billion in commitments to it. OpenAI's existing and planned commitments now come to about 12 gigawatts of Nvidia compute.

CFO Colette Kress told analysts Nvidia will also provide selective credit enhancement for nearly 2 gigawatts of compute at a second frontier lab. She wouldn't say which one.

Nvidia put up to $10 billion into Anthropic in November at a valuation near$ 350 billion, and Anthropic agreed to buy up to a gigawatt of Grace Blackwell and Vera Rubin systems in the same deal.

And Nvidia isn't only guaranteeing these companies. It OWNS pieces of them.

This week's filing shows $18 billion committed to equity investments for the rest of the fiscal year, and$ 47.9 billion already sitting in private companies as of late July.

Now here's where it gets really insane:

Last week, Huang sat on a CNBC set surrounded by six of Wall Street's biggest firms. Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR. They signed a memorandum to mobilise up to $500 billion in outside capital for AI data centres.

Nvidia kept the option to backstop up to a quarter of those deals.

And Huang used that stage to announce that Nvidia GPUs are now an asset class. Pension and credit funds can now lend against graphics cards the way they lend against office towers.

Kress saw the accusation coming and got ahead of it on the earnings call: "We recognise the scale of this support, and we know some will call this circular financing. We see it differently."

But look at the two things Huang says about the same companies.

On the earnings call he said AI has hit its inflection point, that the tokens are productive and profitable, and that compute is now revenue.

But he also said those same labs can't secure investment-grade financing on their own.

A business that's inflecting into profit is exactly the business a bank lends to. Banks lend against cash flow every day. But Nvidia‘s guarantee exists because something in that first story isn't landing with the people whose job is pricing risk.

Kress does have a real answer to this though. She said the second lab's credit support only complements capacity it already secured on its own, without Nvidia backing it. Vendor financing is also old and legal. Cisco did it and GE built a finance arm on it.

Huang's case is that Nvidia understands these businesses better than any lender could, and he says the risk is low and his only regret is not investing more and sooner.

He may be completely right.

But one thing is certain:

Nvidia guarantees the paper. The paper buys the chips. Nvidia books the sale. Then Nvidia tells you the order book is full for a year.

That order book is the entire argument for a $5 trillion company.

And Jensen Huang just explained, in his own words, that his customers couldn't have written those orders without him.

Isn’t this suspicious?

See 2 related tweets

@Ric_RTP: RT @Ric_RTP: This Silicon Valley insider just exposed Sam Altman and Dario Amodei of running a con o...
@gulVasikova: $CBRS One thing I’m paying close attention to in AI right now is the shift from “software eating the...

19. theinformation (Group Score: 86.6 | Individual: 27.8)

Cluster: 5 tweets | Engagement: 21 (Avg: 15) | Type: Tech

Nvidia is buying Hugging Face partly to strengthen open-source AI, which it sees as a counterweight to closed-model developers building rival chips.

Full story: https://t.co/GMHbABBMzd

See 4 related tweets

@Rachelmetz: You really must watch the whole segment, in which @EdLudlow does a great job of pressing @Thom_Wolf ...
@theinformation: Nvidia’s Hugging Face deal gives Jensen Huang control of a key distribution point for open-source AI...
@jason: AI’s Cyber Threat Is Here, NVIDIA Buys Hugging Face & Autonomous Rockets https://t.co/HbslbLnMHV...
@FastCompany: How Nvidia’s Hugging Face deal would reshape the open AI ecosystem https://t.co/Si6gtm6xNJ...

20. hey_Jessicaai (Group Score: 84.4 | Individual: 22.8)

Cluster: 4 tweets | Engagement: 156 (Avg: 112) | Type: Tech

RT @AnishJaitwar: Broke? Jobless? Feeling stuck? Google is paying $1,500 a day — and you don’t need a resume.

All it takes is: 📱 A phone 🌐 Internet ⏱ 5 minutes a day

✅ Like this post 💬 Comment “Google” 🔁 Retweet & Bookmark 🔖 👣 Make sure you’re following @AnishJaitwar so I can DM you the details

See 3 related tweets

@hey_Jessicaai: RT @JayBisen473370: Broke? Jobless? Feeling stuck? Google is paying $1,500 a day — and you don’t nee...
@hey_Jessicaai: RT @robertsmith_ai: Broke? Jobless? Feeling stuck? Google is paying $1,500 a day — and you don’t nee...
@hey_Jessicaai: RT @jihad_sameul: Google pays $1,200 per day.

But most people don't know how.

You can also earn if...