Bitcoin ATMs are a rapidly growing presence in the United States and, some experts say, a rapidly growing cybercrime menace. ATMs dealing in bitcoin are similar to their cash cousins: there are PINs to punch and withdrawal fees, just like any other ATM.
Unlike cash ATMs, though, the high value of crypto makes them prime targets for hackers. So, while a cash ATM tucked away between the snack cakes and energy drinks at a gas station may not draw much attention, a bitcoin ATM gets more scrutiny from bad actors.
“It’s clear that these machines are particularly vulnerable to both physical and cyber threats, making them a prime target for hackers and thieves,” said Timothy Bates, clinical professor of cybersecurityat the University of Michigan’s College of Innovation and Technology.
Bitcoin ATMs can be susceptible to attacks where hackers install malware on the machines to capture private keys, steal funds, or manipulate transactions, which Bates said is “especially concerning for ATMs that may not receive regular software updates or security patches.” Network vulnerabilities are also a weak spot. “If the machine’s network communications are not adequately secured, attackers can intercept data transfers between the ATM and the server, leading to data theft or unauthorized access,” Bates said.
Whether it’s hackers or scammers, the government is sounding the alarm about bitcoin ATMs. The Federal Trade Commission reported this week that scam incidents have risen by 1,000% since 2020.
Ironically, a bitcoin ATM’s risks are directly related to its strengths, according to Joe Dobson, principal analyst at Mandiant, a Google Cloud-owned cybersecurity company. Bitcoin is decentralized, permission-less, and immutable. “A transaction cannot be reversed or recalled if funds are deposited to the wrong address,” Dobson said. And while many crypto bulls find bitcoin’s lack of governance appealing, that can be problematic in ATMs. “There is no governing body within bitcoin dictating who can or cannot run a bitcoin ATM, hence many independent organizations operate the ATMs,” Dobson said.
There are also old criminal tricks that might be reversible in a traditional banking situation, but in the world of bitcoin, that is not so. For example, someone could maliciously slip their personal deposit slips into the stack at the bank, tricking folks into depositing money into their account. “A similar attack can happen with bitcoin ATMs,” Dobson said. “If an attacker compromises a bitcoin ATM, they may change the receiving wallet address (or ‘account number’), effectively stealing user funds.”
But in addition to old tricks, there are newer threats bitcoin ATMs introduce that cash ATMs do not face. Many bitcoin ATMs require personally identifiable information, such as an ID or even a Social Security number to comply with financial industry Know Your Customer (KYC) requirements. This information could be at risk if a bitcoin ATM is compromised.
In Middletown, Ohio, at the Middletown Food Mart in a hollowed-out end of town, a Bitcoin Depot ATM sits opposite a regular cash ATM, blending in among the potato chips, bottled water, and beer. Middletown’s claim to fame lately is as the hometown of Donald Trump’s running mate Ohio Senator J.D. Vance, who has refashioned himself, similar to Trump, as a pro-cryptocurrency warrior. The Middletown Food Mart sits across the street from where Vance grew up.
‘Elon Musk told me to do it.’
Sai Patel, whose family owns Middletown Food Mart, says the bitcoin ATM isn’t very busy.
“Maybe once a month someone comes in to use it,” Patel said. And if it is someone new, Patel will patiently explain how the machine works. He also keeps an eye out for unusual activity. Although the bitcoin ATM isn’t exactly drawing crowds, Patel says a surprising number of senior citizens show up at the kiosk, alarming given the rise of bitcoin ATM scams targeting seniors.
“Elderly people come in and use it,” Patel said.
He described one encounter where an elderly woman entered his shop and headed for the bitcoin ATM, then attempted to send a lot of money somewhere but had questions about using the machine. When Patel asked the woman a few questions as to why, she said, “Elon Musk told me to do it.” Patel quickly realized she had fallen prey to a scam. “I told her, no, no, no, it’s a scam,” Patel said, and he stopped her from dumping her life savings into the machine.
Alice Frei, head of security and compliance at blockchain communications & consulting agency Outset PR, says bitcoin ATM fraud is costly, enhanced by the sometimes shadowy world of crypto.
“Cryptocurrencies are easily exchanged online, often without clear identification of the parties involved. Criminals exploit this anonymity and move money almost invisibly, often employing techniques such as cross-blockchain ‘bridges’ to further obscure transactions,” she said.
And then there’s the fact that an ATM scam probably doesn’t originate in the town where it occurs. “Many crypto exchanges involved in these activities are based offshore, beyond the reach of regulators, making it difficult to trace and recover stolen funds,” Frei added.
Basic steps to avoid bitcoin ATM scams
To protect against these scams, users should be cautious and skeptical of any request to pay through a bitcoin ATM. Legitimate businesses rarely, if ever, demand payment in bitcoin through a machine.
“Verifying the legitimacy of a transaction, particularly checking the recipient’s wallet for connections to questionable entities is crucial,” Frei said, adding that users should also use licensed ATMs from reputable operators to reduce the risk.
Frei said there are steps that users can take to verify the ownership and legitimacy of a bitcoin ATM or parties involved in transactions.
“You can verify the recipient address by checking for flagged activity on platforms like Chainabuse and running an AML check on the address using available tools,” she said, If these tools show the risk score above 70%, it’s advisable to avoid sending money. “Instead, contact the ATM operator or the person who provided the address to clarify the situation,” Frei added.
According to Frei, data shows that nearly 74% of ATMs globally are managed by just 10 operators.
The largest operator of bitcoin ATMs, Bitcoin Depot, operates over 8,000 ATMs. Its CEO Brandon Mintz says the company’s machines are designed to deter hackers. But he also disputes the claims that bitcoin ATMs are major hacking targets.
“Bitcoin ATMs aren’t typically high-priority targets for cybercriminals due to the separation of the hardware and the bitcoin wallet environments,” Mintz said. Bitcoin Depot does not store any bitcoin locally at a bitcoin ATM, and there are many layers of verification and approval processes that prevent unauthorized access to the Bitcoin Depot wallet, he said.
Additionally, Mintz said, most bitcoin ATMs, including Bitcoin Depot’s, only accept cash, so this removes the ability for criminals to use card skimmers like they can install on traditional cash ATMs. However, he says users do need to be aware of scams, and some of the same basic protocols that protect consumers from old-fashioned financial scams apply to the world of cryptocurrency as well.
“Customers of bitcoin ATMs should never send bitcoin or other cryptocurrencies to unknown digital wallets or individuals they don’t know and trust. It’s important to remain vigilant and skeptical of anyone asking for cryptocurrency payments, especially if the request comes with a sense of urgency or threat,” Mintz said.
As the market leader, Bitcoin Depot has been a target of litigation and the company disclosed in its S-1 filing before going public that its users “have been and could be targeted in cybersecurity incidents like an account takeover.” A South Carolina woman sued Bitcoin Depot after falling victim to an alleged cryptocurrency scam. In another instance, authorities in Texas intervened to return money from a Bitcoin Depot ATM after a woman fell victim to a scam.
And that points to a central irony of bitcoin and the bitcoin ATM, products of technology, but ones where the most powerful weapon against fraud isn’t more technology but responsibility, Dobson said. “User responsibility is paramount in cryptocurrency. There is little recompense if something goes awry. The onus is largely on the user to take steps.”
Nvidia on Wednesday evening delivered better-than-expected quarterly results, with a guide that should impress even those with the highest of expectations. Revenue in the company’s fiscal 2026 third quarter grew 62% year over year to $57.01 billion, outpacing the $54.92 billion the Street was looking for, according to estimates compiled by data provider LSEG. Adjusted earnings per share for the three months ending Oct. 26 increased 67% to $1.30, also exceeding the consensus estimate of $1.25, per LSEG data. NVDA YTD mountain Nvidia YTD Talk about a strong showing. In addition to solid beats on the top and bottom lines, management guided current quarter sales to a level not only above consensus estimates but also above the so-called whisper number that was floating around. For those unfamiliar with the term, the estimates that most market watchers and participants, like the Club, cite come from sources like LSEG, FactSet, or Bloomberg – all market data platforms. These estimates are compiled from sell-side analysts, who work at the banks and firms that sell research. The whisper number, however, is what the buy-side – those who run money, like hedge funds, asset management firms, pension funds, and so on – is believed to be looking for. It sometimes happens that a stock can beat the consensus estimate and miss the whisper number, resulting in a stock move lower. Beating the whisper number, however, is an important feat as it means the company is doing even better than the ones running money and risking it on the company, expected – a very bullish sign. Nvidia shares jumped 5% in after-hours trading to $196, a step in the right direction back toward their record-high close of $207 on Oct. 29 and back toward a $5 trillion market cap. We’re reiterating our hold-equivalent 2 rating but bumping up our Nvidia price target to $230 per share from $225. Bottom line Management not only has visibility on just about 100% of the revenue the Street is modeling for next year, but appears to have indicated on the call that the $500 billion number CEO Jensen Huang called out in October is already growing. Helping to drive the growth, Huang explained that the world is currently undergoing three computing transitions simultaneously. First, Huang said there has been a shift from CPU-based general computing to GPU-based accelerated computing. (CPUs are central processing units, long seen as the brains and workhouses of traditional computers. GPUs are graphics processing units, which have become the heart and soul of AI workloads because they can complete many calculations at the same time. That parallel processing is a key advantage over CPUs.) Second, he said that AI is at a “tipping point,” transforming existing applications and enabling new ones. “For existing applications, generative AI is replacing classical machine learning in search ranking, recommender systems, ad targeting, click through prediction, to content moderation. The very foundations of hyperscale infrastructure.” Third, he said, is so-called agentic AI systems “capable of reasoning, planning, and using tools.” (Agentic AI is a type of system that can complete tasks without human supervision — for example, instead of just looking up a flight, it could book it for the user.) Why we own it Nvidia’s high-performance graphics processing units (GPUs) are the key driver behind the AI revolution, powering the accelerated data centers being rapidly built around the world. But Nvidia is more than just a hardware story. Through its Nvidia AI Enterprise service, Nvidia is building out its software business. Competitors : Advanced Micro Devices and Intel Most recent buy : Aug 31, 2022 Initiation : March 2019 At the center of it all is Nvidia. Huang said, “As you consider infrastructure investments, consider these three fundamental dynamics. Each will contribute to infrastructure growth in the coming years. Nvidia’s chosen because our singular architecture enables all three transitions, and thus so, for any form and modality of AI across all industries, across every phase of AI, across all of the diverse computing needs in the cloud, and also from cloud to enterprise to robots – one architecture.” Commentary Coming into the earnings print, we highlighted five questions posed by Ben Reitzes of Melius Research that we hoped Huang would address. The CEO and other company executives answered four of them. The first question from Reitzes was whether the capital expenditure growth could continue through the end of the decade. While time will tell, we said that it was largely going to depend on end market demand, which itself depends on the ability of Nvidia’s customers to monetize the spend. As far as demand goes, Huang got straight to the point on the earnings release, stating “Blackwell sales are off the charts, and cloud GPUs are sold out,” adding that “compute demand keeps accelerating and compounding across training and inference — each growing exponentially.” (Blackwell is the current chip platform from Nvidia) Another question Reitzes raised was: What will Nvidia do with all its free cash flow? Buybacks are clearly still in play, with the company exiting the quarter with $62.2 billion remaining of its share repurchase authorization, even as the company has already returned $37 billion to shareholders this year, through its fiscal third quarter via dividends and buybacks. On the call, Huang said that in addition to buybacks, which will continue, the cash is going to be used to fund further growth and make strategic investments. Nvidia has been on a tear, making “strategic investment” after “strategic investment” – from committing to a $100 billion multiyear investment and partnership with ChatGPT creator OpenAI to taking stakes in rival Claude creator Anthropic, Intel, and neocloud provider CoreWeave. A third question from Reitzes dealt with the need for clarity on the $500 billion of orders for Blackwell and the next generation Rubin that Huang mentioned last month at the company’s GTC conference. On the call, CFO Colette Kress said, “We currently have visibility to a half trillion dollars in Blackwell and Rubin revenue, from the start of this year through the end of calendar year 2026.” Now, Nvidia’s fiscal year is a bit off; it’s almost a year ahead and ends in January. But if we assume that Nvidia does $212.8 billion in its current 2026 fiscal year – about what has thus far been reported, plus the $65 billion from the guidance for the current quarter – that leaves just over $287 billion to be realized in most of its fiscal year 2027, which again extends about one month past the end of calendar year 2026. We know it’s confusing, but suffice it to say, Nvidia already has visibility on nearly 100% of the sales Wall Street is looking for, with time still to go to generate even more orders as enterprise, consumer, and perhaps most exciting, sovereign adoption ramps up. In fact, based on commentary on the call, it seems there have already been announcements for new orders not included in that $500 billion figure, with Kress saying that the deal announced with the Kingdom of Saudi Arabia for 400,000 to 600,000 more GPUs over the three years is new, as is the recently announced deal with Anthropic. “So, there’s definitely an opportunity for us to have more on top of the $500 billion that we announced,” Kress stated. As for Reitzes’ question on margins, they’re clearly going to hold in for the near-term, with management guiding the current quarter to a level above expectations. “Looking ahead to fiscal year 2027, input costs are on the rise, but we are working to hold gross margins in the mid-70s,” Kress said. That’s precisely what the Steet was looking for. The one Reitzes question that Huang did not expand on was about remarks the CEO made earlier this month to the Financial Times, saying “China is going to win the AI race.” At the time, Huang softened that language in a statement, saying “China is nanoseconds behind America in AI,” adding it is vital the U.S. wins by “racing ahead.” While this particular line of inquiry was not mentioned on the call, Huang did say, “While we were disappointed in the current state that prevents us from shipping more competitive data center compute products to China, we are committed to continued engagement with the U.S. and China governments and will continue to advocate for America’s ability to compete around the world.” Nvidia has said for a while now that its forward guidance includes zero sales from China. Segment results Data center , the biggest of Nvidia’s five operating segments, saw revenue increase 66% year over year to a better-than-expected $51.22 billion in fiscal 2026 Q3, and a stunning 25% sequentially. Within the data center unit, compute revenue rose 56% to $43 billion, and networking revenue gained 162% to $8.2 billion. Gaming saw revenue jump 30% to $4.27 billion, but it did miss estimates of $4.41 billion. Professional Visualization revenue jumped 56% and was driven by the company’s recently released DGX Spark, a Grace Blackwell-based AI supercomputer small enough to fit on your desk, and Blackwell sales growth. On the call, Kress said, “Pro visualization has evolved into computers for engineers and developers, whether for graphics or for AI.” Automotive revenue was up 32% year over year as the industry continues to adopt Nvidia’s autonomous solutions. That number was, however, short of expectations. The OEM & Other segment saw revenue up 79%. This unit at Nvidia covers partnerships with original equipment manufacturers, licensing, and other things not accounted for in the other segments. Guidance Looking ahead to the current fiscal 2026 fourth quarter, management’s outlook was largely better than expected. Revenue of $65 billion, plus or minus 2%, was ahead of not only the $61.66 billion LSEG consensus estimate, but also the $64 billion whisper number that was being floated around Wall Street ahead of the release. Adjusted gross margins are expected to be 75%, plus or minus 50 basis points, better than the 74.1% estimate compiled by FactSet. Expectations for adjusted operating expenses in the fiscal fourth quarter of $5 billion are about in line with expectations. (Jim Cramer’s Charitable Trust is long NVDA. See here for a full list of the stocks.) 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.
Jensen Huang, chief executive officer of Nvidia Corp., during the US-Saudi Investment Forum at the Kennedy Center in Washington, DC, US, on Wednesday, Nov. 19, 2025.
Stefani Reynolds | Bloomberg | Getty Images
In the weeks leading up to Nvidia’s third-quarter earnings report, investors debated whether the markets were in an AI bubble, fretting over the massive sums being committed to building data centers and whether they could provide a long-term return on investment.
During Wednesday’s earnings call with analysts, Nvidia CEO Jensen Huang began his comments by rejecting that premise.
“There’s been a lot of talk about an AI bubble,” Huang said. “From our vantage point we see something very different.”
In many respects, Huang’s remarks are to be expected. He’s leading the company at the heart of the artificial intelligence boom, and has built its market cap to $4.5 trillion because of soaring demand for Nvidia’s graphics processing units.
Huang’s smackdown of bubble talk matters because Nvidia counts every major cloud provider — Amazon, Microsoft, Google, and Oracle — as a customer. Most of the major AI model developers, including OpenAI, Anthropic, xAI and Meta, are also big buyers of Nvidia GPUs.
Read more CNBC reporting on AI
Huang has deep visibility into the market, and on the call he offered a three-pronged argument for why we’re not in a bubble.
First, he said that areas like data processing, ad recommendations, search systems, and engineering, are turning to GPUs because they need the AI. That means older computing infrastructure based around the central processor will transition to new systems running on Nvidia’s chips.
Second, Huang said, AI isn’t just being integrated into current applications, but it will enable entirely new ones.
Finally, according to Huang, “agentic AI,” or applications that can run without significant input from the user, will be able to reason and plan, and will require even more computing power.
In making the case of Nvidia, Huang said it’s the only company that can address the three use cases.
“As you consider infrastructure investments, consider these three fundamental dynamics,” Huang said. “Each will contribute to infrastructure growth in the coming years.”
Reversing the slide
In its earnings release, Nvidia reported revenue and profit that sailed past estimates and issued better-than-expected guidance. Last month, Huang provided a $500 billion forecast for sales of the company’s AI chips over calendar 2025 and 2026.
The company said on Wednesday that its order backlog didn’t even include a few recent deals, like an agreement with Anthropic that was announced this week or the expansion of a deal with Saudi Arabia.
“The number will grow,” CFO Colette Kress said on the call, saying the company was on track to hit the forecast.
Prior to Wednesday’s results, Nvidia shares were down about 8% this month. Other stocks tied to the AI have gotten hit even harder, with CoreWeave plunging 44% in November, Oracle dropping 14% and Palantir falling 17%.
Some of the worry on Wall Street has been tied to the debt that certain companies have used to finance their infrastructure buildouts.
“Our customers’ financing is up to them,” Huang said.
Specific to Nvidia, investors have raised concerns in recent weeks about how much of the company’s sales were going to a small number of hyperscalers.
Last month, Microsoft, Meta, Amazon and Alphabet all lifted their forecasts for capital expenditures due to their AI buildouts, and now collectively expect to spend more than $380 billion this year.
Huang said that even without a new business model, Nvidia’s chips boost hyperscaler revenue, because they power recommendation systems for short videos, books, and ads.
People will soon start appreciating what’s happening underneath the surface of the AI boom, Huang said, versus “the simplistic view of what’s happening to capex and investment.”
C. C. Wei, chief executive officer of Taiwan Semiconductor Manufacturing Co. (TSMC), left, and Jensen Huang, chief executive officer of Nvidia Corp., during the TSMC sports day event in Hsinchu, Taiwan, on Saturday, Nov. 8, 2025.
Bloomberg | Bloomberg | Getty Images
Asian chip stocks rallied in early trading Thursday after American AI chip darling Nvidia beat Wall Street expectations and issued stronger-than-expected guidance for the fourth quarter.
South Korea’s SK Hynix popped around 4%. The memory chip maker is Nvidia’s top supplier of high-bandwidth memory used in AI applications.
Samsung Electronics, which also supplies Nvidia with memory, was also up nearly 4%. The company has been working to catch up to SK Hynix in high-bandwidth memory to land more contracts with Nvidia.
“We expect Nvidia’s results to drive higher earnings estimates across the sector, including for its primary GPU supplier TSMC, memory vendors SK Hynix and Samsung, and the broader Asian subcomponent and assembly value chain,” Rolf Bulk, equity research analyst at New Street Research, told CNBC.
In Tokyo, Renesas Electronics, a key Nvidia supplier, added about 4%. Tokyo Electron, which provides essential chipmaking equipment to foundries that manufacture Nvidia’s chips, gained 5.87%. Another Japanese chip equipment maker, Lasertec, was up about 6%.
Japanese tech conglomerate SoftBank skyrocketed nearly 7%, though the firm recently offloaded its shares of Nvidia. Softbank owns the majority of British semiconductor company Arm, which supplies Nvidia with chip architecture and designs.
SoftBank is also involved in a number of AI ventures that use Nvidia’s technology, including the $500 billion Stargate project for data centers in the U.S.
Nvidia’s sales and outlook are closely watched by the technology industry as a sign of the health of the AI boom, and its strong earnings could ease recent fears regarding an AI bubble.
“There’s been a lot of talk about an AI bubble,” Nvidia CEO Jensen Huang told investors on an earnings call. “From our vantage point, we see something very different.”