-
Nvidia’s next AI chip, Blackwell Ultra, will be revealed next month.
Nvidia is hosting its GTC keynote on March 18th, and its keynote speaker has actually simply exposed his talk. “Come to GTC and I’ll talk with you about Blackwell Ultra, Vera Rubin, and after that reveal you the one click after that,” Nvidia CEO Jensen Huang informed experts on the financial Q4 2025 revenues call.
He states Blackwell Ultra will can be found in the 2nd half of next year, with brand-new networking, brand-new memory, brand-new processors, however on the exact same system architecture as Blackwell.
-
OpenAI is apparently getting closer to releasing its internal chip
Image: The Verge
OpenAI stays on track to begin producing its internal AI chip next year, according to a report from ReutersSources inform the outlet that OpenAI prepares to complete its style over the next couple of months before sending it to the Taiwan Semiconductor Manufacturing Co (TSMC) for fabrication.
By making a chip of its own, OpenAI will not need to utilize Nvidia’s chips as much to train and run AI designs. TSMC will produce the chip utilizing the more effective 3-nanometer innovation, with “high-bandwidth memory” and “substantial networking abilities,” according to Reuters
-
Intel is canceling Falcon Shores, its next huge AI chip.
States Holthaus, Intel will “streamline our roadmap and focus our resources” by canceling Falcon Shores. We prepare to utilize Falcon Shores as an internal test chip just without bringing it to market.” It’ll concentrate on Jaguar Shores, a “system-level service at rack scale,” rather.
-
Intel cancels AI chip, talks agonizing past and streamlined future
Image: Alex Castro/ The Verge
-
Nvidia’s market cap come by nearly $600 billion amidst DeepSeek R1 buzz.
As Chinese AI start-up DeepSeek draws attention for open-source AI designs that it states are less expensive than the competitors while offering comparable or much better efficiency, AI chip king Nvidia’s stock rate dropped today.
CNBC stated that after closing at $118.58, down 17 percent, this was “the most significant drop ever for a U.S. business.”
-
Elon Musk, White House consultant, states OpenAI deal revealed at White House is a sham
Illustration: Cath Virginia/ The Verge; Photo: Getty Images
Elon Musk does not miss out on a chance to take a dig at OpenAI– even when the news product in concern is expected to be beneficial to President Trump. Simply a couple of hours after the other day’s White House presser on The Stargate Project concluded, Musk published on X that “they do not really have the cash.”
Softbank, OpenAI, Oracle, and MGX have actually devoted to “release” $100 billion now and $500 billion towards the AI information center business over the next 4 years.
-
An AI supercomputer you can bring around.
This is Nvidia CEO Jensen Huang’s intro of Project Digitsa GB10 Grace Blackwell Superchip-powered system with 128GB of RAM that costs about $3,000 and can run advanced AI designs in a plan little sufficient to rest on your desk.
-
PlayStation and AMD are collaborating to instill video games with AI
Sony is enhancing its collaboration with AMD so they can produce more AI-powered innovation to make video games look and play much better– and not simply on PlayStation hardware. The 2 business are developing a “much deeper cooperation” to deal with “Machine Learning-based innovation for graphics and gameplay,” lead designer of the PS5 and PS5 Pro Mark Cerny revealed on Wednesday
The 2 currently partner on the PS5 and PS5 Pro’s GPUs, which are based upon AMD’s RDNA 2 architecture, and the PS5 Pro utilizes a function called PSSR (PlayStation Spectral Super Resolution) to enhance image clearness and frame rates.
-
China opens an antitrust examination into Nvidia
Cath Virginia/ The Verge
China is examining Nvidia over antitrust offenses, supposedly over claims the chipmaker stopped working to follow conditions set throughout China’s approval for its $6.9 billion acquisition of Israeli network hardware business Mellanox in 2020.
While revealing the DGX A100 GPU after getting Mellanox, Nvidia CEO Jensen Huang stated this while describing its significance to his business:
-
What took place to Intel?
On Monday, Intel CEO Pat Gelsinger suddenly chose to retire after less than 4 years on the task. That was the main story, anyhow. Within hours, Reuters Bloomberg and The New York Times had a various one: the board of directors pressed him out.
3 and a half years earlier, Gelsinger revealed an enthusiastic strategy to reverse the distressed chipmaker within 4 years– now, he’s apparently been tossed out of the business before he might persevere. It took place so quickly that Intel does not have actually a prepared follower in mind, therefore totally that Gelsinger will not even remain as an advisor. He’s gone.
-
Intel’s CEO is out after just 3 years
Image: Laura Normand/ The Verge
-
Nvidia states its Blackwell AI chip is ‘complete steam’ ahead
Nvidia has actually ended up being the world’s most important business on the back of AI chips, passing Microsoft and Apple along the method, and in today’s Q3 2025 incomes, the business recommended its record AI earnings and earnings are just the start.
While The Information just recently reported that its brand-new flagship Blackwell AI servers may have cooling concerns, the business didn’t deal with that on today’s call– rather, Nvidia ensured financiers that Blackwell remains in “complete production,” is “complete steam” ahead, which the business would continue to provide more of the chips each quarter from here on out.
-
Nvidia simply made almost $20 billion in pure revenue in a single quarter.
$14.8 billion earnings in Q1, $16.6 billion in Q2, and now $19.3 billion in Q3 of financial 2025 — that’s revenuenot incomes. (Earnings were $35.08 billion, up from $30.04 billion last quarter.)
The large bulk is from AI information center, obviously– however video gaming did have a 14 percent bump. It’s a $3B-a-quarter company, while information center is a $30B one.
-
Intel’s Gaudi AI chips are far behind Nvidia and AMD, will not even strike $500M objective
-
OpenAI will begin utilizing AMD chips and might make its own AI hardware in 2026
Image: OpenAI
OpenAI is supposedly dealing with Broadcom to establish brand-new custom-made silicon created to manage its big AI work for reasoning and protected production capability with TSMC, according to sources talking to ReutersOpenAI has actually supposedly constructed a chip advancement group of about 20 individuals, consisting of lead engineers who formerly dealt with Google’s Tensor processors for AI
Still, on its existing timeline, the custom-made hardware might not begin production up until 2026.
-
“We had a style defect in Blackwell,” confesses Nvidia CEO.
“It was practical, however the style defect triggered the yield to be low. It was 100% Nvidia’s fault,” Nvidia’s Jensen Huang informs Reuterssuccessfully validating The Information‘s report from August about why its brand-new flagship AI chips will not deliver in big quantities right now.
He states it’s now repaired, however the timeline remains the very same: Q4 for very first deliveries.
-
AMD’s AI chips are coming for Nvidia– however how rapidly?
AMD states the MI325X, delivering Q4, will beat Nvidia’s H200. Nvidia appears an action ahead; it’ll deliver “a number of billion dollars” of its next-gen Blackwell B200 GPU in Q4, too. AMD states its Blackwell rival, the MI355X, will not get here till 2H 2025.
AMD isn’t talking cost, however informed us it’ll damage Nvidia when it concerns overall expense of ownership.
1/19
Tap through our gallery for information about the upcoming MI355X, too. Images: AMD -
Samsung and TSMC have actually apparently gone over structure AI chip “megafactories” in the UAE.
We’ve heard reports about AI chip production tasks in the Middle East connected to OpenAI Sam Altman and Elon Musk
Now, the WSJ states Samsung and TSMC officers have actually checked out the United Arab Emirates “just recently,” talking about tasks worth approximately $100 billion in spite of issues about water sources and developing regional engineering skill.
-
Qualcomm wishes to purchase Intel
Illustration by Alex Castro/ The Verge
On Friday afternoon, The Wall Street Journal reported Intel had actually been approached by fellow chip giant Qualcomm about a possible takeover. While any offer is referred to as “far from specific,” according to the paper’s unnamed sources, it would represent an incredible succumb to a business that had actually been the most important chip business worldwide, based mostly on its x86 processor innovation that for several years had actually thrived over Qualcomm’s Arm chips beyond the phone area.
The New York Times proved the report on Friday night, including that “Qualcomm has actually not yet made a main deal for Intel.”
-
Apple A16 chips are supposedly being made in America.
Previous Bloomberg press reporter Tim Culpan composes on Substack:
Apple’s A16 SoC, which initially debuted 2 years earlier in the iPhone 14 Pro, is presently being produced at Phase 1 of TSMC’s Fab 21 in Arizona in little, however substantial, numbers, my sources inform me.
They’re just utilized in the iPhone 14 Pro and basic iPhone 15 today, however perhaps the American-made chips Apple registered for will wind up in a future iPhone SE sooner or later. The concern is if it’s worth the expenses
-
Intel’s huge turn-around strategy consists of spinning off its chipmaking service
Illustration by Alex Castro/ The Verge
Intel is spinning off its chipmaking organization as part of its strategies to reverse billions in losses and a toppling stock rate. In an statement on MondayIntel CEO Pat Gelsinger stated the Intel Foundry will end up being an independent subsidiary with “clearer separation and self-reliance” from Intel.
With the modification, the Intel Foundry will have its own operating board and report its monetary revenues independently from Intel. Intel will likewise quit working on the factories it’s structure in Poland and Germany for 2 years “based upon expected market need.” The business is still moving forward with its plants in Arizona, Oregon, New Mexico, and Ohio.
-
Sony apparently selected AMD over Intel for the PS6
Illustration by Alex Castro/ The Verge
What’s next after the PS5 ProA report from Reuters concentrates on Sony’s strategies beyond this fall’s brand-new $700 systemstating that the fight to win an agreement for the chip powering a future PlayStation 6 boiled down to AMD vs. Intel, with others like Broadcom got rid of previously, with AMD ultimately triumphing.
According to Reutersconsidering that AMD makes the chip in the PS5 and PS5 Pro, preserving backwards compatibility in a possible relocation became part of “months” of conversations in 2022 in between executives and engineers at Sony and Intel. Intel’s quote was obstructed since they could not concur on how much earnings Intel would make from each chip it would create as Taiwan Semiconductor Manufacturing Co (TSMC) dealt with the production procedure.
-
TikTok’s moms and dad business supposedly gets closer to making its own AI chips.
A report from The Information information the China-based ByteDance’s strategies to standardize 2 brand-new AI chips by 2026 with Taiwan Semiconductor Manufacturing Company (TSMC). The relocation would assist ByteDance conserve “billions of dollars” rather than purchasing chips from Nvidia.
-
AMD is turning its back on flagship video gaming GPUs to chase after AI very first
Image by Tom Warren/ The Verge
AMD is stating the peaceful part out loud: it’s now focusing on AI chips ahead of flagship GPUs for players. The business’s simply laid out a brand-new company technique, where it will combine its RDNA video gaming graphics and CNDA information center efforts into a single universal “UDNA” that’s intended at AI.
In 2 interviews with Tom’s Hardware (you’ll certainly wish to check out bothAMD computing and graphics manager Jack Huynh does not elude. With video gaming graphics, he discusses, the objective is now developing scale and market share at lower cost points– not the “King of the Hill” flagship GPUs that have not encouraged adequate purchasers to leave Nvidia behind.
-
The Nvidia AI antitrust examination is ‘intensifying,’ reports Bloomberg
Cath Virginia/ The Verge
An antitrust examination by the Department of Justice concentrating on Nvidia’s AI supremacy as the hardware business of option is intensifying, reports BloombergThe outlet reports that Nvidia and other business have actually gotten lawfully binding ask for details as its sources state regulators are examining whether Nvidia is “making it more difficult to change to other providers and punishes purchasers that do not specifically utilize its expert system chips.”
“Nvidia wins on benefit, as shown in our benchmark outcomes and worth to clients, and clients can pick whatever option is best for them. We have actually asked with the U.S. Department of Justice and have actually not been subpoenaed. We are delighted to address any concerns regulators might have about our company,” stated Nvidia representative John Rizzo, as reported formerly by CNBC
