Nvidia is one of the world’s biggest computer chip companies, best known for its line of graphics processing units or GPUs. Although the firm had its start in the world of consumer gaming, in recent years it’s grown into a true tech titan with diverse investments in self-driving cars, cloud computing, supercomputing, and artificial intelligence. The parallel processing power of Nvidia’s GPUs has proven to be particularly good at machine learning tasks, and its chips are in high demand not only from AI researchers but any business with an interest in artificial intelligence. From 2015 onwards, Nvidia’s share price grew sharply, allowing the company to make some key acquisitions, including UK chip designer ARM, which it announced it would purchase in September 2020 for $40 billion. Nvidia was founded in 1993 by Jensen Huang, who is currently the firm’s CEO. Known for his leather jackets and upbeat corporate presentations, Huang is a familiar figure to anyone interested in tech.
He’s expected to show a new high-end AI chip dubbed the B100, Reuters reported last week; apparently the new server-grade Blackwell GPUs can consume 1,000 watts each. Nvidia’s GTC keynote kicks off at 1PM PT / 4PM ET — just 15 minutes after I first published the words you’re reading.
Diablo IV is getting a ray tracing upgrade later this month. Armor, water, windows, and other reflective surfaces will all have ray-traced reflections for more detail and lighting accuracy. Shadows throughout the game will also feature ray tracing.
Nvidia is also bringing DLSS 3.5 to Portal with RTX. This should improve the fidelity of full ray tracing in the game, and you can enable Frame Generation on 40-series RTX cards to boost frame rates.
Nvidia updated the service’s status page around 4PM ET to say it had identified the issue and is working on a fix.
The authors, Brian Keene, Abdi Nazemian, and Steward O’Nan, allege that Nvidia trained natural language aspects of its NeMo platform on a massive training dataset containing their work, according to Reuters. Their class-action suit, if certified, would cover anyone in the US with work involved in NeMo’s training.
Although Reuters doesn’t specify, the suit may be referencing the “Books3” dataset that other author lawsuits against OpenAI and Meta are based on.
JEDEC Solid State Technology Association has released details about its new standard, which it says is better positioned to handle the demands of gaming, networking, and AI.
JESD239 GDDR7 is the first JEDEC standard DRAM to use the Pulse Amplitude Modulation (PAM) interface for high frequency operations. Its PAM3 interface improves the signal to noise ratio (SNR) for high frequency operation while enhancing energy efficiency. It also says GDDR7 has double the bandwidth over GDDR6, up to 192 GB/s per device, and double the number of channels.
We don’t expect to see GDDR7 out in the world until Nvidia and AMD release next-gen GPUs, which could come before the end of 2024 but is a long way from being confirmed.
[www.jedec.org]
As we saw at CES last month, Asus’ G16 and G14 latops are sleeker and thinner for 2024, featuring a new aluminum build, OLED screens, larger trackpads, and Intel Core Ultra 9 Meteor Lake chips to go with Nvidia 40-series GPUs.
The G16 price goes from $1,999.99 to $3,299.99 and it’s available directly from Asus or Best Buy. We haven’t reviewed this one yet, but hopefully, that new look hasn’t ruined this iteration of one of our most recommended gaming laptops.
Note: If you buy something from these links, we might get affiliate revenue.
As Nvidia reports its Q4 2023 earnings, CEO Jensen Huang says:
Accelerated computing and generative AI have hit the tipping point. Demand is surging worldwide across companies, industries and nations
That demand, and Nvidia’s dominance in AI chips, powered the company’s record earnings of $60.9 billion for the full year 2023 — a 126 percent increase from last year. Its Q4 revenue of $22.1 billion is an increase of an astounding 265 percent over the same period last year.
The open-source Gemma models are optimized for “the installed base of over 100 million Nvidia RTX GPUs,” installed in PCs around the world, in addition to Nvidia’s ubiquitous AI chips like the H100.
The models will also be part of Nvidia’s Chat with RTX demo, which lets AI models run locally and access users’ files to generate answers to prompts.
The Information reports Microsoft’s server network cards, which are currently in development, could replace the ones offered by Nvidia and save the company cash. Server cards move data quickly between servers. These new server cards should not only improve the performance of Nvidia chips, but Microsoft’s upcoming Maia AI chips as well.
[The Information]
Reuters reports on Nvidia’s new unit that will design custom AI chips, even as its customers plan for a future less reliant on Nvidia’s dominant H100 AI chips. Those can start at $16k each, and power the majority of large language models in use today, which has pushed its market value well beyond $1 trillion.
Nvidia officials reportedly met with OpenAI, Meta, Google, Amazon, and Microsoft to offer its “bespoke” AI chip services, as well as businesses in other industries like gaming and telecom.
For context, here’s how the Wall Street Journal describes OpenAI’s once-again leader’s trillion dollar effort to “reshape the global semiconductor industry:”
Such a sum of investment would dwarf the current size of the global semiconductor industry. Global sales of chips were $527 billion last year and are expected to rise to $1 trillion annually by 2030.
The money is needed to fuel AI’s growth and solve the scarcity of expensive AI chips required to train the large language models that underpin systems like ChatGPT. According to the WSJ, Altman is pitching a chip-making partnership to investors from the UAE, SoftBank CEO Masayoshi Son (again), and TSMC.
Technically, it began a voluntary recall in December — “STOP USING THEM IMMEDIATELY,” it wrote — but now the US Consumer Product Safety Commission is involved. Full refunds or $60 store credits are available. 25,300 units were sold, with 272 reports of loose, overheated, and melted adapters. Nvidia’s own adapter had a similar issue in 2022.
Nvidia has issued a hotfix GPU driver to address micro-stuttering with V-SYNC and stutter when scrolling in web browsers for some systems. The patch also fixes some stutter in Red Dead Redemption and stability issues with Immortals of Aveum. If you’re not experiencing issues then it’s probably best waiting for Nvidia’s next WHQL driver, but if you’re eager for the fixes you can download the latest driver from Nvidia’s website.
Nvidia has launched its RTX 4080 Super GPUs today, and Asus is ready with an update to last year’s RTX 4080 Noctua Edition. The Super model includes a four slot cooler with 120mm Noctua fans on top. Asus hasn’t confirmed pricing just yet, but given the $450 premium on the previous RTX 4080 model it’s probably going to be a lot more than the $999 starting price of the RTX 4080 Super.
Nvidia powers most of the AI projects from Microsoft, OpenAI, Amazon, and Meta, but they’re also trying to lessen their dependence on its limited supply. The New York Times explains they want to make switching between Nvidia chips and others (including their own) “as simple” as possible.
As The Verge reported, OpenAI CEO Sam Altman is interested in building chips. Microsoft’s AI-focused chip Maia 100 is expected to arrive this year, and Amazon announced the latest version of its Trainium chip.
6
Verge Score
Nvidia’s second RTX 40-series Super card isn’t as impressive as its first.
In his review, Tom Warren calls the first RTX 40-series Super card from Nvidia “one that I would seriously consider for a 1440p PC build,” and if you’ve been waiting for a PC gaming upgrade that won’t destroy your budget, this might be the one.
Along with a new Game Ready driver, 4070 Super hardware is now available at stores like Amazon, Best Buy, B&H, and Newegg for $599.99 and up, depending on the card.
Note: If you buy something from these links, we might get affiliate revenue.
8
Verge Score
Nvidia delivers more performance for the same price point inside a great 1440p GPU.
Nvidia’s Chat with RTX lets you train your own local chatbot with your own files for free, no cloud required. Seconds after I fed it the Epic v. Google legal complaint PDF, I got decent answers to questions like “What does Epic Games want” and “Which laws does Epic allege Google violated?”
But it also confidently hallucinated that Framework’s Nirav Patel is The Verge’s CEO after ingesting this YouTube transcript, LOL. Maybe I’ll just use it as memory aid for reams of old notepad files?
Monday was quite the whirlwind! Let’s take a step back and recap all of the fun surprises:
• Apple dropped the launch date for the Vision Pro.
• Samsung’s Ballie AI robot now doubles as a projector.
• MSI has a new Steam Deck competitor called the Claw.
• Nvidia revealed its RTX 4080 Super and RTX 4070 Ti Super.
There’s still more to come! Stay tuned to The Verge for more CES coverage from the show floor.
Nvidia only just announced its new RTX 40-series Super GPUs today at CES, but we’ve already got our hands on one. The RTX 4070 Super launches on January 17th at the same $599 price point as the original, but there are more CUDA cores to power 1440p gaming. It looks very similar to the RTX 4070, but Nvidia has picked a slightly darker shade of alloy frame.
RTX Remix, which lets modders update older games with visual upgrades like ray tracing and Nvidia’s DLSS, launches on January 22nd in open beta. It made Portal look incredible, and it’s being used to build a community-developed remaster of Half-Life 2.
Nvidia, Twitch, and OBS are teaming up on the new enhanced broadcasting feature that will be available in beta later this month. It allows streamers to transmit up to three concurrent streams to Twitch at different resolutions and quality. This allows non-partner / affiliate streamers to broadcast at multiple resolutions and quality levels. Twitch is also experimenting with 4K and AV1 streaming on RTX 40-series GPUs.
You can read more about Twitch’s plans in a blog post and a support page.
Update January 8th, 1:45PM ET: Added additional links from Twitch.
We’re expecting to see some new RTX 40-series Super cards today, if leaks and Nvidia’s own teasers are anything to go by. Nvidia is holding a “special address” keynote at CES 2024, starting at 8AM PT / 11AM ET / 4PM GMT today. You can watch the whole show below, and stayed tuned to The Verge for all the latest Nvidia news.
Apple CEO Tim Cook, Google CEO Sundar Pichai, Nvidia CEO Jensen Huang, and OpenAI CTO Mira Murati are all on the list to meet with European Commission executive vice president Margrethe Vestager, according to a report from Reuters. We’ll be on the lookout for what (if anything) comes out of the meetings.
