Anthropic priced above its range and upsized to a record $81 billion - ANT starts trading today. The Trump-Xi-Putin trilateral produced a four-point peace declaration with AI nowhere in it, 15 security figures urge the DOJ to probe what Nvidia knew about chip smuggling, and OpenAI answered Microsoft's Ignite with GPT-6.2. In the afternoon: ANT opened 34% above its IPO price and topped a $2 trillion market value, Trump rejected global AI rules at APEC while Xi called for cooperation, and a short seller's China claim knocked AMD down about 8%.
The debut
1Anthropic prices at $83 - above the range, upsized, and the biggest IPO ever
Anthropic priced its IPO at $83 a share on Wednesday night, above the $70 to $78 range, and upsized the offering to 980 million shares from 825 million - raising about $81 billion at a valuation of roughly $1.5 trillion, which AP reports puts it ahead of SpaceX's June offering as the largest IPO ever. Most of the shares are sold by the company, with a small slice from existing stockholders and a 147-million-share overallotment option still with underwriters. Trading begins this morning on the Nasdaq under the ticker ANT; the offering is expected to close Friday. The debut lands against the week's jitters - Nvidia's leadership crisis, bubble talk, and a public backlash - with one analyst calling the pricing "a big win for the tech sector after a white-knuckle time with Nvidia".
In plain English
The biggest listing in history just got bigger - and it opens for trading today.
Why it matters
An above-range, upsized pricing on ~8x-covered books is the clearest verdict yet that public markets will fund frontier AI at scale; the first-day print will set the tone for every AI company waiting behind it, OpenAI included. Watch where ANT opens relative to the $83 line.
2The Shenzhen trilateral delivered a four-point peace declaration - AI was eclipsed
Trump met Xi and Putin separately, then together for 90 minutes in the first-ever trilateral of the three leaders, ending with a "Declaration Toward Peace": UN reforms, increased refined-petroleum supply, a warning that the ICC and similar bodies should not "abuse" their jurisdiction, and a pledge that a nuclear war must not be fought. Trump pushed nuclear limits - "the world's three largest nuclear powers... are uniquely positioned" to cut arsenals - while Xi called for communication and "not to challenge the red lines of one another". The AI safety agenda previewed for the bilateral - turning September's crisis hotline into a working channel - appears nowhere in the declaration; AP's wrap says big-power politics eclipsed the summit's talk of open trade and AI, even as Xi told leaders that countries must unite in keeping technology development open. US lawmakers from both parties publicly doubted the meetings would deliver.
In plain English
The three leaders made history and a joint statement - but on nukes, oil and the UN, not on AI.
Why it matters
The working AI crisis hotline between the two AI superpowers remains unbuilt, and chip tensions sit unresolved, at the exact moment the industry's biggest company is enmeshed in a China-linked scandal. Any AI-specific readout from the Trump-Xi bilateral would change this picture; none had been published at writing.
3Fifteen security and AI figures urge the DOJ to probe what Nvidia knew about chip smuggling
A group of 15 national security experts and AI policy leaders has asked the Justice Department to investigate what Nvidia knew about the alleged smuggling of its advanced AI chips to China, NBC News reports exclusively - including whether the company kept selling to customers or middlemen after evidence emerged that they were illegally diverting the technology, or "deliberately blinded itself to export control violations". The letter, sent Tuesday to Attorney General Todd Blanche, argues that smuggling "threatens to undermine" the US lead in what the administration calls super intelligence. It lands one day after Nvidia's board created a National Security and Export Control Oversight Committee, and as Huang prepares to meet US lawmakers on December 1.
In plain English
The Nvidia scandal is moving from the boardroom to the Justice Department's inbox.
Why it matters
A DOJ probe would put the AI buildout's central supplier under criminal-justice scrutiny over the exact issue - chips reaching China - that now shapes US trade policy and the company's own guidance. Nvidia has not publicly responded to the letter; a probe is requested, not opened.
4OpenAI answers Ignite: GPT-6.2 Astra and Sol, as Microsoft prices its MAI line to undercut
OpenAI released two new flagship variants this week, GPT-6.2 Astra - the large, most-capable version - and GPT-6.2 Sol, the everyday model, both available now to paying ChatGPT and Codex subscribers and both beating their siblings in OpenAI's own benchmarking. The release answers Microsoft's Ignite, where seven in-house MAI models debuted - MAI Code 2, MAI Cyber 2 Flash, MAI Image 2.7 and Dragon Copilot Transcribe among them - positioned as significantly cheaper enterprise alternatives to OpenAI's and Anthropic's flagships, alongside a new licensed-weights program that lets companies customize the models.
In plain English
The model race didn't pause for the chaos - OpenAI shipped upgrades, and Microsoft is competing on price.
Why it matters
Microsoft - OpenAI's biggest backer - is now openly pricing its own models against its partner's, on the day Anthropic's listing reprices the whole sector. The benchmark wins are OpenAI's own tests; independent evaluations will take time.
5ANT opens at $111.50, 34% above the IPO price - and tops $2 trillion
Anthropic shares opened at $111.50 on Thursday morning, 34% above the $83 IPO price, Nikkei reports, and the company's market value topped $2 trillion during the session, per CNBC's live coverage - making Anthropic the tenth publicly traded US tech company valued at $1 trillion or more, a list now led by Apple, Microsoft, Nvidia, Alphabet and Amazon. More than 640 million shares changed hands by midday, already past the 522 million that traded in SpaceX's entire first day in June. Early backer Menlo Ventures' aggregate position is around $25 billion, partner Matt Murphy told CNBC, and Melius initiated coverage at buy with a $155 target. The debut also fed the policy fight: Senator Mark Warner and other Democrats renewed calls for AI regulation - "A record-breaking IPO for a company worth $1.5 trillion... and still NO action on AI regulation".
In plain English
The verdict is in - public investors paid up, massively.
Why it matters
A 34% pop on the largest IPO ever reprices every private AI lab behind it, OpenAI included - and hands regulation advocates their biggest exhibit yet. The flip side: the gap between the $83 price and the open is money the company and its selling stockholders did not capture.
6Trump tells APEC the US opposes global AI regulation - Xi calls for cooperation
AI did surface at the summit after all - in Trump's own remarks to an APEC session, reported after this morning's edition. "We have to watch it. We have to be careful. We have to cherish it. It's almost completely good, but it's got some bad elements," Trump said, while making clear he opposes regulation that could hamper the industry: "We do not want to do anything to destroy the tremendous potential of this technology. We won't allow that to happen." The Hindu headlines the stance as US opposition to "globalist regulation" of AI. Xi, in his opening remarks, called for international cooperation on AI and an open technology ecosystem - "Open cooperation is a sure path to the development and revitalization of the Asia Pacific" - with outlets including Folha reporting his framing of keeping AI under human control.
In plain English
Both superpowers did talk AI at the summit - and described opposite governing philosophies.
Why it matters
This corrects the morning's picture: the split is now on the record - Washington publicly rejecting global AI rules while Beijing asks for cooperation - with the bilateral crisis hotline still unannounced. For labs, a US-China regulatory fork makes fragmented compliance the base case.
7AMD slides about 8% after a short seller alleges its firmware reached a blacklisted Chinese company
AMD shares dropped roughly 8% after short seller Hunterbrook alleged that the company's firmware repeatedly appeared at a blacklisted Chinese server maker and related entities, CNBC and the Globe and Mail report - ties the report argues could invite US scrutiny. AMD is both an Anthropic investor and a chip supplier to it, and the allegation lands the same week that 15 national-security and AI figures urged the Justice Department to probe what Nvidia knew about chip smuggling to China.
In plain English
The China-chip scrutiny widened from Nvidia to the industry's other AI chipmaker.
Why it matters
This is an allegation from a firm positioned to profit from the decline, not a confirmed violation - but the market move is real, and it shows how quickly China exposure now reprices AI chip stocks, with AMD down about 8% and Micron also sliding out of the trillion-dollar club this week.
Checked against original announcements and named press reports on November 19, 2026, morning and mid-afternoon. Anthropic's figures are from its own pricing announcement; the record comparison is AP's. The trilateral declaration's four points are as reported by The Hill; no AI-specific readout of the Trump-Xi bilateral had been published at writing, so the hotline's status is unconfirmed, not decided. The DOJ letter is a request for an investigation, not an opened probe; Nvidia had not publicly responded. OpenAI's benchmark claims are its own tests. ANT's opening print and session market value are as reported by Nikkei and CNBC's live coverage; intraday prices move. Trump's AI remarks are as quoted by Firstpost, with the 'globalist regulation' framing from The Hindu. The AMD move follows a short seller's allegation - claims from a party positioned to profit, not a confirmed violation.