Shoppers looking for gadgets and gizmos powered by generative AI technology to gift to their loved ones won’t have many options to choose from this holiday season.
Generative artificial intelligence has taken Silicon Valley by storm since the launch of OpenAI’s ChatGPT chatbot in November 2022. Although startups have raised billions to build new GenAI tools and tech giants have bought millions of Nvidia processors to train AI models, few companies have delivered new hardware built with the new-age tech as its focal point.
There was a lot of optimism over the potential of GenAI gadgets at the CES trade show in January, said Paul Gagnon, vice president for analyst firm Circana. In particular, products from high-profile startups such as Humane and Rabbit, which were marketed as being able to translate, answer questions, take voice memos and set alarms, were drawing buzz, Gagnon said.
But many of these new GenAI devices didn’t work as well as people expected, with reviewers saying that the gadgets were too slow and too prone to failure.
“As we’ve gone through the year, and those kinds of promises — which I’ll be honest, were pretty nebulous to start with — there’s been a bit of a struggle with communicating that to consumers,” Gagnon said.
A key reason GenAI hardware hasn’t had a breakthrough is that current devices are “compute restrained,” meaning they require more powerful silicon chips and related components to perform better, particularly when compared with smartphones, said Ben Bajarin, CEO of Creative Strategies, a market research firm.
Additionally, consumers may find current GenAI devices too expensive, and they may be confused about what the devices can actually do, he said.
GenAI devices, such as the Ray-Ban Meta smart glasses, also typically require a smartphone connection for an accompanying app as well as strong internet access, because a bad internet connection can lead to performance delays that frustrate people, Bajarin said.
While companies such as Microsoft, Apple, Intel, Dell and Lenovo have also heavily marketed new lineups of personal computers capable of performing GenAI tasks, consumers have yet to perk up to the sales pitch, said Ryan Reith, an IDC program vice president for mobile devices.
“I don’t think that there’s actually a need for consumers to go out and get one of these more expensive PCs,” Reith said, noting that people may be confused about why they need beefier computers when they can already access tools such as ChatGPT through their current PCs.
The reality is that while GenAI has captivated Silicon Valley, it’s still “inning zero” in regard to widespread adoption, Bajarin said.
“Even though I can rattle off all these productivity stats of how people are using AI today, it’s a very small number of people,” he said. “This is not mainstream.”
It may not be until 2025 that consumers see a “big explosion” in GenAI computers, smartphones and new gadgets, said Steve Koenig, vice president of research at the Consumer Technology Association, which produces CES.
Despite Silicon Valley not having a breakout year for GenAI hardware, here are a few GenAI devices early adopters can buy.
Ray-Ban Meta glasses
Meta released the second generation of its Ray-Ban smart glasses in 2023, but the company began rolling out GenAI features for the device earlier this year and announced several new AI capabilities at its Connect event in September.
The glasses don’t offer users augmented reality capabilities, but people can use the device to take photos, listen to music and ask the Meta AI digital assistant for information about the things within their field of view.
With the help of the device’s mics and camera, for instance, users can ask the Meta AI digital assistant to recommend a recipe when they walk through a grocery aisle and scan the shelves, the company said in a blog post.
Meta, which makes Facebook and Instagram, is selling certain versions of the glasses for 20% off through Dec. 2. This means that a pair of the Ray-Ban Meta Skyler style of glasses will cost $239.20 instead of $299 if bought online.
Rabbit r1
The Rabbit r1 is a $200 gizmo that looks like an orange, miniaturized tablet with a playful aesthetic that’s more Nintendo Switch than Apple iPad.
Outfitted with a camera and dual mics, the r1 can record audio clips and set timers or perform more advanced tasks, such as helping users recall details from past conversations, search results and voice recordings. After the device began shipping in March, reviewers criticized the r1 for stumbling at various tasks and failing to outshine smartphones that can do many of the same functions.
The startup “has used that feedback to rapidly make very significant improvements to the user experience” and has released scores of updates to improve, Rabbit CEO Jesse Lyu told CNBC in a statement.
Despite the harsh reviews, Rabbit has “sold more than 100,000 r1 devices when we originally expected to sell only 3,000” and the company is “seeing a return rate of less than 5%, which is very solid for a first-generation product,” Lyu said.
Rabbit is currently running a deal that gives shoppers free shipping, or $15 off, if they order an r1 by Dec. 4.
Bee
After raising $7 million in funding in July, the startup Bee AI will begin selling its GenAI device, the Bee, on Friday.
The Bee looks like an internet-connected smartwatch and functions like an advanced digital assistant. Its dual mics allow it to listen and analyze people’s voice memos and conversations to provide summaries and to-do lists, Bee AI CEO Maria de Lourdes Zollo told CNBC.
The Bee can also be integrated with health-care tools and people’s Google and Gmail accounts to help generate personalized summaries and action items, Zollo said. Although the startup offers a Bee app for the Apple Watch for people who don’t want to buy another hardware device, she said the core Bee device is better at understanding voices in loud environments.
Shoppers can buy the Bee for $49.99 and get its basic tasks, but they will have to pay a $15-per-month subscription for more features such as “better memory or better capabilities,” Zollo said.
For Black Friday, Bee is offering shoppers three free months of the device’s subscription service. The device should ship in time for Christmas, Zollo said.
Startup Figure AI is developing general-purpose humanoid robots.
Figure AI
Figure AI, an Nvidia-backed developer of humanoid robots, was sued by the startup’s former head of product safety who alleged that he was wrongfully terminated after warning top executives that the company’s robots “were powerful enough to fracture a human skull.”
Robert Gruendel, a principal robotic safety engineer, is the plaintiff in the suit filed Friday in a federal court in the Northern District of California. Gruendel’s attorneys describe their client as a whistleblower who was fired in September, days after lodging his “most direct and documented safety complaints.”
The suit lands two months after Figure was valued at $39 billion in a funding round led by Parkway Venture Capital. That’s a 15-fold increase in valuation from early 2024, when the company raised a round from investors including Jeff Bezos, Nvidia, and Microsoft.
In the complaint, Gruendel’s lawyers say the plaintiff warned Figure CEO Brett Adcock and Kyle Edelberg, chief engineer, about the robot’s lethal capabilities, and said one “had already carved a ¼-inch gash into a steel refrigerator door during a malfunction.”
The complaint also says Gruendel warned company leaders not to “downgrade” a “safety road map” that he had been asked to present to two prospective investors who ended up funding the company.
Gruendel worried that a “product safety plan which contributed to their decision to invest” had been “gutted” the same month Figure closed the investment round, a move that “could be interpreted as fraudulent,” the suit says.
The plaintiff’s concerns were “treated as obstacles, not obligations,” and the company cited a “vague ‘change in business direction’ as the pretext” for his termination, according to the suit.
Gruendel is seeking economic, compensatory and punitive damages and demanding a jury trial.
Figure didn’t immediately respond to a request for comment. Nor did attorneys for Gruendel.
The humanoid robot market remains nascent today, with companies like Tesla and Boston Dynamics pursuing futuristic offerings, alongside Figure, while China’s Unitree Robotics is preparing for an IPO. Morgan Stanley said in a report in May that adoption is “likely to accelerate in the 2030s” and could top $5 trillion by 2050.
Concerns about stock valuations in companies tied to artificial intelligence knocked the market around this week. Whether these worries will recede, as they did Friday, or flare up again will certainly be something to watch in the days and weeks ahead. We understand the concerns about valuations in the speculative aspects of the AI trade, such as nuclear stocks and neoclouds. Jim Cramer has repeatedly warned about them. But, in the past week, the broader AI cohort — including real companies that make money and are driving what many are calling the fourth industrial revolution — has been getting hit. We own many of them: Nvidia and Broadcom on the chip side, and GE Vernova and Eaton on the derivative trade of powering these energy-gobbling AI data centers. That’s not what should be happening based on their fundamentals. Outside of valuations, worries also center on capital expenditures and the depreciation that results from massive investments in AI infrastructure. On this point, investors face a choice. You can go with the bears who are glued to their spreadsheets and extrapolating the usable life of tech assets based on history, a seemingly understandable approach, and applying those depreciation rates to their financial models, arguing the chips should be near worthless after three years. Or, you can go with the commentary from management teams running the largest companies driving the AI trade, and what Jim has gleaned from talking with the smartest CEOs in the world. When it comes to the real players driving this AI investment cycle, like the ones we’re invested in, we don’t think valuations are all that high or unreasonable when you consider their growth rates and importance to the U.S., and by extension, the global economy. We’re talking about Nvidia CEO Jensen Huang, who would tell you that advancements in his company’s CUDA software have extended the life of GPU chip platforms to roughly five to six years. Don’t forget, CoreWeave recently re-contracted for H100s from Nvidia, which were released in late 2022. The bears with their spreadsheets would tell you those chips are worthless. However, we know that H100s have held most of their value. Or listen to Lisa Su, CEO of Advanced Micro Devices , who said last week that her customers are at the point now where “they can see the return on the other side” of these massive investments. For our part, we understand the spending concerns and the depreciation issues that will arise if these companies are indeed overstating the useful lives of these assets. However, those who have bet against the likes of Jensen Huang and Lisa Su, or Meta Platforms CEO Mark Zuckerberg, Microsoft CEO Satya Nadella, and others who have driven innovation in the tech world for over a decade, have been burned time and again. While the bears’ concerns aren’t invalid, long-term investors are better off taking their cues from technology experts. AI is real, and it will increasingly lead to productivity gains as adoption ramps up and the technology becomes ingrained in our everyday lives, just as the internet has. We have faith in the management teams of the AI stocks in which we are invested, and while faith is not an investment strategy, that faith is based on a historical track record of strong execution, the knowledge that offerings from these companies are best in class, and scrutiny of their underlying business fundamentals and financial profiles. Siding with these technology expert management teams, over the loud financial expert bears, has kept us on the right side of the trade for years, and we don’t see that changing in the future. (See here for a full list of the stocks in Jim Cramer’s Charitable Trust, including NVDA, AVGO, GEV, ETN, META, MSFT.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.
Every weekday, the CNBC Investing Club with Jim Cramer releases the Homestretch — an actionable afternoon update, just in time for the last hour of trading on Wall Street. Markets: The S & P 500 bounced back Friday, recovering from the prior session’s sharp losses. The broad-based index, which was still tracking for a nearly 1.5% weekly decline, started off the session a little shaky as Club stock Nvidia drifted lower after the open. It was looking like concerns about the artificial intelligence trade, which have been dogging the market, were going to dominate back-to-back sessions. But when New York Federal Reserve President John Williams suggested that central bankers could cut interest rates for a third time this year, the market jumped higher. Rate-sensitive stocks saw big gains Friday. Home Depot rose more than 3.5% on the day, mitigating a tough week following Tuesday’s lackluster quarterly release. Eli Lilly hit an all-time high, becoming the first drugmaker to reach a $1 trillion market cap. TJX also topped its all-time high after the off-price retailer behind T.J. Maxx, Marshalls, and HomeGoods, delivered strong quarterly results Wednesday. Carry trade: We’re also monitoring developments in Japan, which is dealing with its own inflation problem and questions about whether to resume interest rate hikes. That brings us to the popular Japanese yen carry trade, which is getting squeezed as borrowing costs there are rising. The yen carry trade involves borrowing yen at a low rate, then converting them into, say, dollars, and investing in higher-yielding foreign assets. That’s all well and good when the cost to borrow yen is low. It’s a different story now that borrowing costs in Japan are hitting 30-year highs. When rates rise, the profit margin on the carry trade gets crunched, or vanishes completely. As a result, investors need to get out, which means forced selling and price action that becomes divorced from fundamentals. It’s unclear if any of this is adding pressure to U.S. markets. We didn’t see anything in the recent quarterly earnings reports from U.S. companies to suggest corporate fundamentals are deteriorating in any meaningful way. That’s why we’re looking for other potential external factors, alongside the well-known concerns about artificial intelligence spending, the depreciation resulting from those capital expenditures, and general worries about consumer sentiment and inflation here in America. Wall Street call: HSBC downgraded Palo Alto Networks to a sell-equivalent rating from a hold following the company’s quarterly earnings report Wednesday. Analysts, who left their $157 price target unchanged, cited decelerating sales growth as the driver of the rerating, describing the quarter as “sufficient, not transformational.” Still, the Club name delivered a beat-and-raise quarter, which topped estimates across every key metric. None of this stopped Palo Alto shares from falling on the release. We chalked the post-earnings decline up to high expectations heading into the quarter, coupled with investor concerns over a new acquisition of cloud management and monitoring company Chronosphere. Palo Alto is still working to close its multi-billion-dollar acquisition of identity security company CyberArk , announced in July. HSBC now argues the stock’s risk-versus-reward is turning negative, with limited potential for upward estimate revisions for fiscal years 2026 and 2027. We disagree with HSBC’s call, given the momentum we’re seeing across Palo Alto’s businesses. The cybersecurity leader is dominating through its “platformization” strategy, which bundles its products and services. Plus, Palo Alto keeps adding net new platformizations each quarter, converting customers to use its security platform, and is on track to reach its fiscal 2030 target. We also like management’s playbook for acquiring businesses just before they see an industry inflection point. With Chronosphere, Palo Alto believes the entire observability industry needs to change due to the growing presence of AI. We’re reiterating our buy-equivalent 1 rating and $225 price target on the stock. Up next: There are no Club earnings reports next week. Outside of the portfolio, Symbotic, Zoom Communications , Semtech , and Fluence Energy will report after Monday’s close. Wall Street will also get a slew of delayed economic data during the shortened holiday trading week. U.S. retail sales and September’s consumer price index are scheduled for release early Tuesday. Durable goods orders and the Conference Board consumer sentiment are released on Wednesday morning. (See here for a full list of the stocks in Jim Cramer’s Charitable Trust.) As a subscriber to the CNBC Investing Club with Jim Cramer, you will receive a trade alert before Jim makes a trade. Jim waits 45 minutes after sending a trade alert before buying or selling a stock in his charitable trust’s portfolio. If Jim has talked about a stock on CNBC TV, he waits 72 hours after issuing the trade alert before executing the trade. THE ABOVE INVESTING CLUB INFORMATION IS SUBJECT TO OUR TERMS AND CONDITIONS AND PRIVACY POLICY , TOGETHER WITH OUR DISCLAIMER . NO FIDUCIARY OBLIGATION OR DUTY EXISTS, OR IS CREATED, BY VIRTUE OF YOUR RECEIPT OF ANY INFORMATION PROVIDED IN CONNECTION WITH THE INVESTING CLUB. NO SPECIFIC OUTCOME OR PROFIT IS GUARANTEED.