Daily at 8AM KST · Summaries and takeaways from 10 AI articles
We cross-check industry press like TechCrunch and The Decoder, official announcements from OpenAI and Google DeepMind, and community signal from Hacker News.The point isn't what happened but why it matters, tied into one read on the day.
120 days running
1173 stories · 319 terms explained · since 2026-07-23
·what we picked →
2026-11-19
· ~9 min read
· Past edition
10 picked from 109 candidates · ordered by significance
Today's Insight
Nvidia's profit keeps growing, but it halted China sales, delayed its quarterly report and limited neocloud guarantees
Nvidia earned $65 billion last quarter and projects $130 billion in revenue next quarter. On the same earnings call, interim CEO Colette Kress said Nvidia had stopped shipping to China and would file its quarterly report, the , later than usual. A UBS analyst says management told him the delay relates to the Jensen Huang investigation and whether to add risk factors.
Continue reading (3)Show less
Nvidia also attached conditions to its guarantees for , newer companies that buy Nvidia chips and rent out computing capacity. If a company cannot get its facility running, the guarantee disappears. The Verge's writer relayed reports that Kress as CFO pushed to reduce this kind of financing. Employees were promised raises and equity in December.
Anthropic reportedly priced its IPO at $83, above its initial range. The $1.5 trillion valuation is higher than May's but short of the $2 trillion it had sought. The pricing came right after Huang's departure from Nvidia shook confidence in the industry, and Anthropic starts trading on Thursday.
Google released Gemini 4 Neon and widened the Gemini app to low-memory Android phones. The Genie 4 and temporary threads in AI Mode also arrived the same day. Google is expanding models, devices and its search screen all at once.
Signal to watch
Watch which new risk factors appear in the 10-Q Nvidia says it will file within weeks.
01Google DeepMind Blog
Official
Covered by 3 more outlets
·2026-11-18·~50s read
Models & Research
Google DeepMind on November 18 released Gemini 4 Neon, a model built to finish complex coding, knowledge-work and science tasks quickly. Google says it ranks third on Artificial Analysis's overall Intelligence Index with a score of 61, second on the Vals Index at 70.9%, and first on Artificial Analysis's Coding Agent Index at 58.6. It costs $3 per million input and $15 per million output tokens, with a 1 million-token context window. It is available today in the Gemini API, Google AI Studio, Gemini for business, and the Gemini app for paid Pro and Ultra subscribers.
Why it matters
Why It Matters
Google is now selling Gemini 4 as several models named after elements. It says Carbon handles the hardest long-running projects, Hydrogen fast everyday work, and Neon balances depth and efficiency between them. The Decoder noted that this is the fourth new Gemini model in a month. The rankings come from outside benchmarks that Google cites, so whether it fits your own work only shows up once you try it.
Google has opened the Gemini app to the hundreds of millions of people who use lower-memory Android phones. It re-architected the app so the core experience runs smoothly on Android (Go edition) hardware with less memory. Until now those phones only had basic voice commands. They can now use Gemini Live in more than 45 languages, point the camera at notes or forms to translate them, and create images with Nano Banana. On supported phones the app launches with a long press of the power button, and it will come preloaded on upcoming devices from partners such as Transsion.
Why it matters
Why It Matters
Until now these phones got only basic voice commands, and Google had to re-architect the app to change that. The new features reach supported devices, not every Go phone. Partners such as Transsion will preload the app on upcoming phones, so many buyers will find Gemini there out of the box.
What to do now
If you use an Android Go phone, search for the Gemini app in Google Play and check whether it installs on your device.
Google DeepMind has opened Genie 4, its most advanced (a model that generates 3D spaces an AI can move through), to Google AI Ultra subscribers. It had previously been available only as a research preview. Typing text changes the weather, characters or storyline in real time, and the model now generates sound that matches the scene for the first time. The older Genie 3 also gained sessions of up to three minutes, 360-degree panoramas and chainable worlds. A new agent builds worlds from spoken descriptions, a community gallery lets people share and remix them, and all output carries a SynthID watermark. This follows The Decoder's report.
Why it matters
Why It Matters
Genie 4 has moved from research preview into a subscription product, but only for top-tier Ultra subscribers; Pro members must join a waitlist. The gallery and remix features also give people a place to share what they make.
Get it in your inbox every day
Daily headlines, a weekly synthesis on Sundays, and a monthly report. Sent at 8AM KST, the evening before in the US.
Free · no ads · one-click unsubscribe
We use your email only to send the briefing, stored and sent via services in Japan and the US. Privacy policy
Check your inbox
Find this subject in your inbox and press the link to start your subscription.
Confirm your AI News Briefing subscription
It usually arrives within a minute
If you don't see it, check your spam folder too
The link is valid for 7 days
04Wired AI
Press
Covered by 5 more outlets
·2026-11-18·~50s read
Business & Funding
Interim Nvidia CEO Colette Kress told employees just minutes into Wednesday's first all-hands that pay raises and more equity would arrive in December, months ahead of schedule. Compensation was the most common topic among 700 questions submitted in advance, she said, followed by a request for transparency from the board. Board chair Harvey Jones repeated that Jensen Huang resigned over personal conduct and a lack of candor, and said the company would not look for a "Jensen clone." Thousands of employees have reportedly lobbied for Huang's return and a petition has about 200,000 purported backers, but the board is searching inside and outside the company. Nvidia shares rose about 5% on Wednesday.
Why it matters
Why It Matters
One concern investors have raised is a loss of top talent, since many employees stayed at Nvidia for Huang and his vision. On Tuesday's earnings call Kress called stabilizing the workforce probably the most important thing leadership is working on, and she opened the all-hands with pay. Departures before a successor is named will show whether that holds people.
TechCrunch reports that Anthropic, the maker of Claude, has priced its IPO at $83 per share, above the initial range of $70 to $78. On a fully diluted basis (counting all shares expected to be issued) it is valued at $1.5 trillion, up from $965 billion in May when it raised $65 billion, but below the $2 trillion it was reportedly aiming for months ago. It begins trading on Nasdaq on Thursday under the ticker "ANT." Its seven co-founders each sold nearly 3.4 million shares for about $282 million apiece, and Amazon and Google hold nearly 15% and 14% respectively.
Why it matters
Why It Matters
TechCrunch says pricing above the range points to strong investor appetite, even after Nvidia CEO Jensen Huang's departure shook confidence in the industry. The valuation is still below the $2 trillion Anthropic once sought. TechCrunch adds that Thursday's first trades may set expectations for private rivals such as OpenAI, which has not announced IPO plans.
On Tuesday's earnings call Nvidia said it would file its quarterly report, the , "in the coming weeks" and gave no reason. The company has typically filed it the same day it reports results. UBS analyst Timothy Arcuri wrote that management told him on a callback the delay relates to the ongoing investigation of Huang and whether additional risk factors must be added. He kept a buy rating but cut his price target from $240 to $220. The same day, The Wall Street Journal reported that the Securities and Exchange Commission (SEC) is investigating whether Nvidia's disclosures misled investors.
Why it matters
Why It Matters
The 10-Q carries far more detail than an earnings release, including sections on risk factors and legal proceedings, so those are the pages to read if the delay is tied to the investigation. Nvidia already faces scrutiny from several congressional committees, the Justice Department's look at Megaspeed, a company tied to Nvidia's former CEO, and, reportedly, the SEC.
The Verge's tally puts Nvidia's latest quarterly profit at $65 billion, about $8,300 every second or roughly $700 million a day. Its profit peaks during the crypto boom and the GPU shortage were $1.24 billion and $3 billion a quarter. On Tuesday's earnings call, interim CEO Colette Kress said Nvidia intends to return all excess free cash flow, net of strategic uses, to shareholders. The baseline Huang promised five months earlier was 50%. Next-quarter revenue is projected at $130 billion.
Why it matters
Why It Matters
Market value fell from about $5 trillion in early November to $3.6 trillion after Huang's exit, but profit shows no sign of slowing. The Verge argues that gives new leadership little reason to change much. The number to watch is next quarter's gross margin forecast, which slips from nearly 75% to 71-72%.
Even Samsung, counting phones and TVs, earned less than Nvidia, as did the other memory leaders.
Nvidia added a new paragraph to its third-quarter earnings release that attaches conditions to its commitments to , newer companies that buy Nvidia chips and rent out AI computing capacity. Nvidia had offered some of them a guarantee to buy their capacity if no one else would. The new wording says the commitments take effect only if the neocloud secures financing, and the capacity is deployed, made operational and accepted by Nvidia. When Bernstein analyst Stacy Rasgon asked, Kress said a delayed buildout means the guarantee "doesn't exist anymore," adding that nothing had changed.
Why it matters
Why It Matters
According to The Information and The Wall Street Journal, Kress as CFO pushed Huang to spend less on financing the AI ecosystem and significantly scaled back a commitment to OpenAI. The Verge's writer reads the new wording as Nvidia publicly stating limits on what it will back, and notes signs that the Federal Reserve may raise rates, which would hurt debt-funded builders such as CoreWeave. Kress, for her part, said the wording is not new.
Google is adding "temporary threads" to AI Mode, the conversational interface in Search. Tapping the temporary threads icon at the top right of AI Mode detaches that conversation from your account, so it will not resurface in search history or recommendations. The AI's answers are not informed by past preferences and the chat does not affect future personalization. Closing the thread returns you to normal search. It is rolling out globally on desktop and mobile in all supported languages.
Why it matters
Why It Matters
The Gemini chatbot already had temporary chats, and AI Mode now gets the same option. It is meant for digging into a personal topic without having it harden into one of your recorded interests. The report confirms separation from history and personalization only, and does not cover how much data Google keeps on its side.
What to do now
Start a temporary thread from the icon at the top right of AI Mode. The rollout is gradual, so the icon may not be visible yet.
Nvidia has walked away from a China sales approval it fought to win. Interim CEO Colette Kress said on Tuesday's earnings call that guidance includes no licensed Hopper data center products and that Nvidia is no longer shipping them to China. She did not explain why. The US government approved H200 sales last December, and Huang had pushed for that approval. Reuters reported earlier this month that around 600,000 H200 chips were stuck in Hong Kong after Beijing withheld approval for large-scale imports. Concerns had also been raised that Chinese buyers were not properly vetted.
Why it matters
Why It Matters
Huang's interim successor dropped, on her first earnings call, the China approval he had fought for, and gave no reason. The trouble with these sales started before Huang left: Beijing had held up large-scale imports, and questions had been raised about how Chinese buyers were vetted. What is clear is that China H200 revenue is out of Nvidia's guidance.
The same briefing you just read, every day. A weekly synthesis on Sundays and a monthly report at the start of each month come with it. Sent at 8AM KST, which is the evening before in the US.
Free · no ads · one-click unsubscribe
We use your email only to send the briefing, stored and sent via services in Japan and the US. Privacy policy
Check your inbox
Find this subject in your inbox and press the link to start your subscription.