Sundar Pichai, CEO of Google and Alphabet, speaks on artificial intelligence during a Bruegel think tank conference in Brussels, Belgium, on Jan. 20, 2020.
Yves Herman | Reuters
Google on Wednesday announced MedLM, a suite of new health-care-specific artificial intelligence models designed to help clinicians and researchers carry out complex studies, summarize doctor-patient interactions and more.
The move marks Google’s latest attempt to monetize health-care industry AI tools, as competition for market share remains fierce between competitors like Amazon and Microsoft. CNBC spoke with companies that have been testing Google’s technology, like HCA Healthcare, and experts say the potential for impact is real, though they are taking steps to implement it carefully.
The MedLM suite includes a large and a medium-sized AI model, both built on Med-PaLM 2, a large language model trained on medical data that Google first announced in March. It is generally available to eligible Google Cloud customers in the U.S. starting Wednesday, and Google said while the cost of the AI suite varies depending on how companies use the different models, the medium-sized model is less expensive to run.
Google said it also plans to introduce health-care-specific versions of Gemini, the company’s newest and “most capable” AI model, to MedLM in the future.
Aashima Gupta, Google Cloud’s global director of health-care strategy and solutions, said the company found that different medically tuned AI models can carry out certain tasks better than others. That’s why Google decided to introduce a suite of models instead of trying to build a “one-size-fits-all” solution.
For instance, Google said its larger MedLM model is better for carrying out complicated tasks that require deep knowledge and lots of compute power, such as conducting a study using data from a health-care organization’s entire patient population. But if companies need a more agile model that can be optimized for specific or real-time functions, such as summarizing an interaction between a doctor and patient, the medium-sized model should work better, according to Gupta.
Real-world use cases
A Google Cloud logo at the Hannover Messe industrial technology fair in Hanover, Germany, on Thursday, April 20, 2023.
Krisztian Bocsi | Bloomberg | Getty Images
When Google announced Med-PaLM 2 in March, the company initially said it could be used to answer questions like “What are the first warning signs of pneumonia?” and “Can incontinence be cured?” But as the company has tested the technology with customers, the use cases have changed, according to Greg Corrado, head of Google’s health AI.
Corrado said clinicians don’t often need help with “accessible” questions about the nature of a disease, so Google hasn’t seen much demand for those capabilities from customers. Instead, health organizations often want AI to help solve more back-office or logistical problems, like managing paperwork.
“They want something that’s helping them with the real pain points and slowdowns that are in their workflow, that only they know,” Corrado told CNBC.
For instance, HCA Healthcare, one of the largest health systems in the U.S., has been testing Google’s AI technology since the spring. The company announced an official collaboration with Google Cloud in August that aims to use its generative AI to “improve workflows on time-consuming tasks.”
Dr. Michael Schlosser, senior vice president of care transformation and innovation at HCA, said the company has been using MedLM to help emergency medicine physicians automatically document their interactions with patients. For instance, HCA uses an ambient speech documentation system from a company called Augmedix to transcribe doctor-patient meetings. Google’s MedLM suite can then take those transcripts and break them up into the components of an ER provider note.
Schlosser said HCA has been using MedLM within emergency rooms at four hospitals, and the company wants to expand use over the next year. By January, Schlosser added, he expects Google’s technology will be able to successfully generate more than half of a note without help from providers. For doctors who can spend up to four hours a day on clerical paperwork, Schlosser said saving that time and effort makes a meaningful difference.
“That’s been a huge leap forward for us,” Schlosser told CNBC. “We now think we’re going to be at a point where the AI, by itself, can create 60-plus percent of the note correctly on its own before we have the human doing the review and the editing.”
Schlosser said HCA is also working to use MedLM to develop a handoff tool for nurses. The tool can read through the electronic health record and identify relevant information for nurses to pass along to the next shift.
Handoffs are “laborious” and a real pain point for nurses, so it would be “powerful” to automate the process, Schlosser said. Nurses across HCA’s hospitals carry out around 400,000 handoffs a week, and two HCA hospitals have been testing the nurse handoff tool. Schlosser said nurses conduct a side-by-side comparison of a traditional handoff and an AI-generated handoff and provide feedback.
With both use cases, though, HCA has found that MedLM is not foolproof.
Schlosser said the fact that AI models can spit out incorrect information is a big challenge, and HCA has been working with Google to come up with best practices to minimize those fabrications. He added that token limits, which restrict the amount of data that can be fed to the model, and managing the AI over time have been additional challenges for HCA.
“What I would say right now, is that the hype around the current use of these AI models in health care is outstripping the reality,” Schlosser said. “Everyone’s contending with this problem, and no one has really let these models loose in a scaled way in the health-care systems because of that.”
Even so, Schlosser said providers’ initial response to MedLM has been positive, and they recognize that they are not working with the finished product yet. He said HCA is working hard to implement the technology in a responsible way to avoid putting patients at risk.
“We’re being very cautious with how we approach these AI models,” he said. “We’re not using those use cases where the model outputs can somehow affect someone’s diagnosis and treatment.”
Getty Images
Google also plans to introduce health-care-specific versions of Gemini to MedLM in the future. Its shares popped 5% after Gemini’s launch earlier this month, but Google faced scrutiny over its demonstration video, which was not conducted in real time, the company confirmed to Bloomberg.
In a statement, Google told CNBC: “The video is an illustrative depiction of the possibilities of interacting with Gemini, based on real multimodal prompts and outputs from testing. We look forward to seeing what people create when access to Gemini Pro opens on December 13.”
Corrado and Gupta of Google said Gemini is still in early stages, and it needs to be tested and evaluated with customers in controlled health-care settings before the model rolls out through MedLM more broadly.
“We’ve been testing Med-PaLM 2 with our customers for months, and now we’re comfortable taking that as part of MedLM,” Gupta said. “Gemini will follow the same thing.”
Schlosser said HCA is “very excited” about Gemini, and the company is already working out plans to test the technology, “We think that may give us an additional level of performance when we get that,” he said.
Another company that has been using MedLM is BenchSci, which aims to use AI to solve problems in drug discovery. Google is an investor in BenchSci, and the company has been testing its MedLM technology for a few months.
Liran Belenzon, BenchSci’s co-founder and CEO, said the company has merged MedLM’s AI with BenchSci’s own technology to help scientists identify biomarkers, which are key to understanding how a disease progresses and how it can be cured.
Belenzon said the company spent a lot of time testing and validating the model, including providing Google with feedback about necessary improvements. Now, Belenzon said BenchSci is in the process of bringing the technology to market more broadly.
“[MedLM] doesn’t work out of the box, but it helps accelerate your specific efforts,” he told CNBC in an interview.
Corrado said research around MedLM is ongoing, and he thinks Google Cloud’s health-care customers will be able to tune models for multiple different use cases within an organization. He added that Google will continue to develop domain-specific models that are “smaller, cheaper, faster, better.”
Like BenchSci, Deloitte tested MedLM “over and over” before deploying the technology to health-care clients, said Dr. Kulleni Gebreyes, Deloitte’s U.S. life sciences and health-care consulting leader.
Deloitte is using Google’s technology to help health systems and health plans answer members’ questions about accessing care. If a patient needs a colonoscopy, for instance, they can use MedLM to look for providers based on gender, location or benefit coverage, as well as other qualifiers.
Gebreyes said clients have found that MedLM is accurate and efficient, but it’s not always great at deciphering a user’s intent. It can be a challenge if patients don’t know the right word or spelling for colonoscopy, or use other colloquial terms, she said.
“Ultimately, this does not substitute a diagnosis from a trained professional,” Gebreyes told CNBC. “It brings expertise closer and makes it more accessible.”
Tencent on Thursday posted 15% year-on-year revenue growth, with AI boosting the Chinese tech giant’s performance in advertising targeting and gaming.
Here’s how Tencent performed in the third quarter of 2025, per earnings released on Thursday:
Revenue: 192.9 billion Chinese yuan ($27.12 billion), surpassing the 189.2 billion Chinese yuan expected analysts, according to data compiled by LSEG.
Operating profit: 63.6 billion yuan, versus 58.01 billion yuan expected by the street.
Tencent boosted its capital expenditure earlier this year as it ramped up AI and eyed European expansion for its cloud computing services, which would compete against market leaders Amazon Web Services, Google Cloud and Microsoft Azure. It has its own AI foundational model in China called Hunyuan, however it also uses DeepSeek in some products.
Tencent shares are up 56.7% year-to-date.
This is a breaking news story. Please refresh for updates.
Traders work on the floor of the New York Stock Exchange (NYSE) on Nov. 12, 2025 in New York City.
Spencer Platt | Getty Images
The divergence between the performance of the Dow Jones Industrial Average and Nasdaq Composite on Wednesday stateside reinforces the suggestion that there are two markets operating in the U.S.: one of an artificial intelligence and another of “everything else.”
Not only did the Dow rise, it also secured its second consecutive record high and closed above the 48,000 level for the first time.
The index, which comprises 30 blue-chip companies, is typically seen as a marker of the “old economy.” That is to say, it is mostly made up of large, well-established companies driving the U.S. economy, such as banks, healthcare and industrials, before Silicon Valley became a minisun powering everything.
To be sure, new and flashy names, such as Nvidia and Salesforce, constitute the Dow too. But as the index is price-weighted, meaning that companies with higher share prices influence the Dow more, tech companies don’t exert as much gravity on it.
That’s in contrast to the Nasdaq, which is weighted by companies’ market capitalization, and dominated mainly by technology firms. The tech-heavy index fell as shares like Oracle and Palantir slipped — even Advanced Micro Devices’ 9% pop on its growth prospects couldn’t rescue the Nasdaq from the red.
It’s not necessarily a warning sign about overexuberance in AI.
“There’s nothing wrong, in our view, of kind of trimming back, taking some gains and re-diversifying across other spots in the equity markets,” said Josh Chastant, portfolio manager of public investments at GuideStone Fund.
But what investors would really like is if fork in the road merges into one. That tends to be the safer path to take.
What you need to know today
And finally…
People walk by the New York Stock Exchange (NYSE) on June 18, 2024 in New York City.
Private equity firms are facing a new reality: a growing crop of companies that can neither thrive nor die, lingering in portfolios like the undead.
These so-called “zombie companies” refer to businesses that aren’t growing, barely generate enough cash to service debt and are unable to attract buyers even at a discount. They are usually trapped on a fund’s balance sheet beyond its expected holding period.
Cisco Systems shares spiked higher Wednesday evening after the networking company delivered a quarterly beat and outlook raise. Another quarter of double-digit order growth proves Cisco is an underrated winner from the AI infrastructure buildout. Revenue in the company’s fiscal 2026 first quarter, which ended Oct. 25, increased 8% year over year to $14.88 billion, exceeding the LSEG-complied analyst consensus estimate of $14.76 billion. Non-GAAP earnings increased 10% on an annual basis to $1 per share, beating expectations of 98 cents, LSEG data showed. GAAP stands for generally accepted accounting principles. CSCO YTD mountain Cisco Systems YTD Look at the shares of Cisco go. They surged more than 7% in after-hours trading to just about $80 per share. That’s on top of a 3% move in regular trading hours. If the stock can take out $80.06, it will make its first all-time high since March 2000. Shares, as of Wednesday’s close, rose roughly 25% year to date. Bottom line It’s a deserving move after an excellent quarter, highlighted by accelerating product order growth, especially from artificial intelligence customers. During the post-earnings call, Cisco CEO Chuck Robbins attributed the strength in AI orders to a “deepening” relationship with existing customers. The company also called out that a “major multi-year, multi-billion-dollar campus networking refresh cycle” is underway. It wasn’t all perfect, however, as the security business missed estimates, with revenue falling year over year. According to management, some revenue recognition timing issues need to be sorted out. Security weakness was our main concern ahead of the quarter. The business also missed revenue estimates in the prior quarter, and we didn’t think a quick turnaround was likely. Our fear of this repeat was the main reason why we took some profits in this position Monday at around $71. Even though we were right to be cautious on security, the market was turning a blind eye to this issue because of how fast networking is growing. A rebound in security also isn’t needed for management to hit on its outlook, which was raised well above Street estimates Wednesday evening. Another concern of the bears entering earnings was that Cisco would be negatively impacted by the government shutdown due to its large federal agencies business. Despite the closed government, Robbins noted this business managed to grow orders by a high single-digit percentage in the quarter. He’s anticipating upside in orders once the government reopens. Why we own it Cisco Systems is an enterprise networking equipment provider that has made big strides to appeal to cloud customers. The company has also increased its presence in the security market through its acquisition of Splunk. In addition, Cisco’s long-term transition toward subscription software sales, which are sticky and come with higher margins, should help improve the stock’s undemanding price-to-earnings multiple. Competitors : Arista Networks , Hewlett Packard Enterprise , Juniper Networks Most recent buy : Aug. 19, 2025 Initiated : July 17, 2025 The story remains that Cisco has turned into a sleeper AI play thanks to the billions of dollars it is taking in from hyperscaler customers. That surge of orders is converting to big revenue. In fiscal year 2025, Cisco recognized roughly $1 billion of AI revenue from hyperscalers, which are the biggest of the Big Tech names, such as the major cloud companies. On the call, Robbins said he expects to recognize roughly $3 billion from hyperscalers in fiscal year 2026. Despite this accelerating growth and subscription revenue making up more than half of its total revenue, the stock still trades at a reasonable price-to-earnings multiple of about 19.5 times based on the new midpoint of management’s full-year adjusted earnings-per-share (EPS) outlook. We’re reiterating our 2 rating because we don’t like to chase stock spikes, but we are increasing our price target to $85 per share from $78. Commentary Total Product orders increased 13% year over year – an acceleration from 7% growth in the prior quarter – with growth across all geographies and customer markets. When we review Cisco, we always focus on orders because that’s the best leading indicator of where revenue is headed. Product revenue grew 10% year over year to $7.77 billion, beating estimates of about $7.47 billion. Starting with the Networking sub-segment, product orders increased by a high teens rate, representing the fifth consecutive quarter of double-digit growth. AI infrastructure orders from hyperscaler customers were a big driver of that growth. Cisco took in $1.3 billion of orders in the quarter, an acceleration from the more than $800 million in the prior quarter. The company also saw strong orders for enterprise routing, campus switching, wireless, industrial IoT, and servers. Credit Cisco’s close relationships with portfolio name Nvidia and Advanced Micro Devices for its recent AI success. Last month, Cisco announced the N9100, which they called the first Nvidia partner developed data center switch based on Nvidia Spectrum-X Ethernet switch silicon. “The N9100, available in the second half of fiscal year 2026, will provide the operational consistency and flexibility needed for sovereign and neocloud providers to build and manage AI at scale,” Robbins explained. Neoclouds are next-generation specialized clouds for accelerated computing. CoreWeave , which rents cloud-based Nvidia chips for AI tasks, is an example of a neocloud. Cisco is also helping G42, leading United Arab Emirates AI firm, with powering, connecting, and securing its large-scale AI clusters with AMD graphics processing units (GPUs) The enterprise AI story is starting to emerge, too. Cisco experienced strong demand for switching, routing, and wireless products, which Robbins said is an indication of customers “investing in the connectivity needed for AI deployments.” Across sovereign, neocloud, and enterprise customers, Robbins called out a growing pipeline above $2 billion for its high performance networking products. This comes after Cisco booked $200 million of orders in its fiscal first quarter from these customers. By division, Networking revenue increased 15% to $7.77 billion, beating estimates. The largest driver of this increase in sales was from service provider routing, which is mostly from AI infrastructure. Data center switches and enterprise routing were also up double digits, while campus switching revenue increased by a high single digit percentage. In the Security division, revenue fell 2% year over year and missed analysts’ forecasts again. It’s disappointing to see a sizeable miss in back-to-back quarters, but management attributed the decline to a timing issue. Robbins explained that more customers are using Splunk’s offerings through cloud subscriptions instead of on-premise deals, leading to a timing change of when revenue is recognized. Ultimately, this transition isn’t a bad thing. The company is in favor of more subscription-based revenue. Cisco completed its $28 billion acquisition of Splunk in March 2024. “We are actually pleased to see more cloud subscriptions for Splunk as they enable greater adoption and expansion, and allow us to deliver innovation faster to enable customers to unlock value from AI Now ” Robbins explained on the call. More broadly. Cisco said it continued to see order growth for some of it newer and refreshed security products, which make up about one third of the portfolio, while its order products are in decline. Importantly, management doesn’t believe Security’s stumbles will last long. They expect revenue growth to accelerate and end the year at a much higher rate. But even if that doesn’t happen and the results don’t materially improve from here, Cisco said it’s still confident in its ability to deliver on its fiscal Q2 and full year 2026 outlook. The Collaboration and Observability units saw revenue drop 3% and rise 6%, respectively, with Collaboration missing estimates and Observability matching expectations. Services revenue increased 2% year over year to $3.81 billion, slightly beating estimates. As always, we appreciate Cisco’s consistent approach to returning cash to shareholders. The company repurchased $2 billion worth of shares in the quarter at an average price of $68.28. That looks like a great trade since the stock is knocking on the door of $80 in after-hours trading. It has $12.2 billion remaining under its authorization. Cisco stock, as of Wednesday’s closing price, has a 2.2% annual dividend yield. Guidance Cisco expects fiscal 2026 second-quarter revenue of $15 billion to $15.2 billion, which is well above the consensus estimate of $14.62 billion. It also sees non-GAAP EPS of $1.01 to $1.03 cents, which is nicely above the consensus estimate of 98 cents. For full year 2026, Cisco now expects revenue of $60.2 billion to $61 billion, which is about a $1 billion increase from the prior outlook of $59 billion to $60 billion. This revised outlook exceeds the consensus estimate of $59.64 billion. On the bottom line, management raised its EPS forecast to $4.08 to $4.14 from its prior outlook of $4.00 to $4.06. This new midpoint of $4.11 is better than the consensus analyst estimate by 7 cents. (Jim Cramer’s Charitable Trust is long CSCO, 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.
