Closing Time: Nvidia talks Perplexity at $30 billion-plus.

• Nvidia is discussing an equity investment in Perplexity in a round that would value the AI search startup at more than $30 billion, according to people with knowledge of the talks, The Information reported.

• Alibaba launched a Hong Kong share placement worth about HK$80 billion ($10.2 billion) to fund full-stack AI, Reuters reported, and Hong Kong shares fell as much as 10 percent after the pricing, CNBC said.

• Server makers are telling customers prices for some of Nvidia’s flagship AI-server chip systems will rise about 17 percent, generally on Grace Blackwell 300 and Vera Rubin 200 systems, The Information reported.

• Nvidia said its Groq 3 LPX rack is in full production after the $20 billion Groq deal, with Nebius and SpaceX as early Vera and LPX customers, CNBC reported.

Today’s Biggest Moves

Symbol	Name	Chg %	Price
$RGTI	Rigetti Computing, Inc.	-8.60%	16.37 USD
$IONQ	IonQ, Inc.	-8.47%	41.06 USD
$QBTS	D-Wave Quantum Inc.	-8.39%	18.68 USD
$MDB	MongoDB, Inc.	-6.53%	402.69 USD
$STX	Seagate Technology Holdings PLC	-6.51%	794.65 USD
$SNDK	Sandisk Corporation	-6.45%	1493.12 USD
$MU	Micron Technology, Inc.	-5.83%	910.43 USD
$SMCI	Super Micro Computer, Inc.	-5.56%	35.17 USD
$WDC	Western Digital Corporation	-5.24%	435.38 USD
$COHR	Coherent Corp.	-4.85%	275.49 USD

---

1. Nvidia is shopping a Perplexity stake above $30 billion

Phoebe Liu, Valida Pau, and Amir Efrati at The Information reported that Nvidia is discussing an equity investment in Perplexity as part of a round that would value the AI search startup at more than $30 billion, according to people with knowledge of the discussion. The round would be worth billions of dollars and lift Perplexity’s valuation more than 50 percent from its last financing, with annualized revenue already above $750 million.

That is the same deal toolkit Nvidia has used on model and power bets, now pointed at the query layer, and the article also frames a tech-licensing or hiring path Nvidia considered alongside the equity talk. The talks are at a $30 billion-plus price, not a closed check size. Read The Information

2. Alibaba raises $10.2 billion for AI, then the stock gets hit

Kane Wu and Casey Hall at Reuters reported Sunday that Alibaba launched a Hong Kong share placement worth about HK$80 billion ($10.2 billion), with 100 percent of the net proceeds earmarked for full-stack artificial intelligence, meaning chips, infrastructure, and the development and deployment of AI models. A term sheet reviewed by Reuters put the deal at 710 million ordinary shares priced at HK$112.70, a 3.6 percent discount, and Reuters said it would be the largest-ever primary follow-on by a Hong Kong-listed company.

Two people familiar told Reuters the book saw sovereign-wealth demand and that Alibaba increased the size after it was oversubscribed. CNBC said Hong Kong shares fell as much as 10 percent Monday and were last down 8.4 percent at the placement price, with U.S.-listed ADRs off 3.4 percent premarket. China AI capex is now a priced $10 billion equity print with a dilution sticker. Read Reuters

3. Nvidia AI chips are getting about 17 percent more expensive

Phoebe Liu and Amir Efrati at The Information reported that prices for some of Nvidia’s flagship AI-server chip systems are set to rise about 17 percent, according to two people whose firms were notified by server makers, and that the hikes generally apply to Grace Blackwell 300 and Vera Rubin 200 systems due to be delivered next. That number sharpens Saturday’s Bloomberg report, from Brody Ford and Ian King, that some of Nvidia’s biggest customers had been told the prices of servers that hold its AI chips are going up more than 15 percent in many cases as memory-chip costs soar.

Bloomberg’s people familiar with the notices said the increases take effect on systems shipped early next year, with the size of each hike depending on the chip generation and the memory configuration. Contract manufacturers that build those racks for Microsoft, Google, and Oracle have already passed the notice along, and Nvidia did not comment to Bloomberg. Read The Information

4. Groq racks are in production, with SpaceX and Nebius in line

Kif Leswing at CNBC reported that Nvidia announced Monday its Groq 3 LPX rack is in full production, commercializing technology from its December $20 billion Groq asset purchase, the company’s largest acquisition. Dion Harris, an Nvidia senior director, told reporters the rack will be deployed alongside Vera CPUs and Rubin GPUs at the neocloud Nebius and will be online later this year.

Nvidia packages 256 Groq 3 chips per LPX rack and said the system can deliver 3,400 tokens per second on an Artificial Analysis benchmark, a measure of how fast a model can generate text. Phoebe Liu at The Information reported that SpaceX and Nebius are set to be early customers of Nvidia’s Vera CPUs and inference-focused Groq LPX racks. The $20 billion check is now a shipping rack with named first customers, two days before Nvidia earnings. Read CNBC

5. Taiwan indicts nine over AI servers bound for China

Reuters reported from Taipei that Taiwan prosecutors indicted nine people, including employees of Nvidia and Super Micro, over the alleged illegal export of AI servers to China. Keelung prosecutors charged eight people, including one employee of Nvidia’s Taiwan unit and two employees of Super Micro’s Taiwan unit, with breach of trust and document forgery.

The alleged scheme made it appear that 130 Super Micro B300 servers would stay in a rented Taiwan facility; 74 were ultimately exported to Chinese customers, including through Indonesia, Japan, and Hong Kong, and Taiwan customs stopped the remaining 56 after spotting irregularities. Prosecutors said the defendants were fully aware of Nvidia and Super Micro export-control rules and colluded for enormous profit. This is the same B300-to-China channel Super Micro’s board said last week it found no evidence current senior management knew about. Read Reuters

6. OpenAI shopped Lancium, then Nvidia wrote the check

Valida Pau at The Information reported that OpenAI explored an investment in or an acquisition of Lancium, the power developer behind its Stargate campus in Abilene, Texas, last quarter, before Nvidia put billions of dollars into the same firm. Stargate is the giant Texas data-center project that needs land and megawatts as much as it needs chips, and Lancium is the landlord on the power side of that site.

That sequence shows how far a model maker will go to lock the electricity under its own campus, and it shows Nvidia is willing to own a piece of the landlord when OpenAI walks away. The report does not attach a new dollar figure beyond those billions. Read The Information

7. General Intuition is shopping a $6 billion valuation

Rebecca Bellan at TechCrunch reported that General Intuition, the New York startup building a foundation model that trains generalized AI agents to move through space and time, is in talks to raise funding at a $6 billion pre-money valuation from new investors including Valor Equity Partners, Point72 Ventures, and Seven Seven Six. Existing investors including Khosla Ventures and General Catalyst are also participating, just weeks after the company raised $320 million at a $2.3 billion valuation.

CEO Pim de Witte spun the company out last October from the video-game clip platform Medal. A source close to the deal told TechCrunch the round is still being finalized and is oversubscribed, with the money aimed at robotic embodiments, more compute through a CoreWeave partnership, and hiring. In talks is not a closed check, and TechCrunch marked Valor’s participation as still being confirmed. Read TechCrunch