Sopa Images | Lightrocket | Getty Images

Nvidia has established itself as the undisputed leader in artificial intelligence chips, selling large quantities of silicon to most of the world’s biggest tech companies en route to a $4.5 trillion market cap.

One of Nvidia’s key clients is Google, which has been loading up on the chipmaker’s graphics processing units, or GPUs, to try and keep pace with soaring demand for AI compute power in the cloud.

While there’s no sign that Google will be slowing its purchases of Nvidia GPUs, the internet giant is increasingly showing that it’s not just a buyer of high-powered silicon. It’s also a developer.

On Thursday, Google announced that its most powerful chip yet, called Ironwood, is being made widely available in the coming weeks. It’s the seventh generation of Google’s Tensor Processing Unit, or TPU, the company’s custom silicon that’s been in the works for more than a decade.

TPUs are application-specific integrated circuits, or ASICs, which play a crucial role in AI by providing highly specialized and efficient hardware for particular tasks. Google says Ironwood is designed to handle the heaviest AI workloads, from training large models to powering real-time chatbots and AI agents, and is more than four times faster than its predecessor. AI startup Anthropic plans to use up to 1 million of them to run its Claude model.

For Google, TPUs offer a competitive edge at a time when all the hyperscalers are rushing to build mammoth data centers, and AI processors can’t get manufactured fast enough to meet demand. Other cloud companies are taking a similar approach, but are well behind in their efforts.

Amazon Web Services made its first cloud AI chip, Inferentia, available to customers in 2019, followed by Trainium three years later. Microsoft didn’t announce its first custom AI chip, Maia, until the end of 2023.

“Of the ASIC players, Google’s the only one that’s really deployed this stuff in huge volumes,” said Stacy Rasgon, an analyst covering semiconductors at Bernstein. “For other big players, it takes a long time and a lot of effort and a lot of money. They’re the furthest along among the other hyperscalers.”

Originally trained for internal workloads, Google’s TPUs have been available to cloud customers since 2018. Of late, Nvidia has shown some level of concern. When OpenAI signed its first cloud contract with Google earlier this year, the announcement spurred Nvidia CEO Jensen Huang to initiate further talks with the AI startup and its CEO, Sam Altman, according to reporting by The Wall Street Journal.

Unlike Nvidia, Google isn’t selling its chips as hardware, but rather providing access to TPUs as a service through its cloud, which has emerged as one of the company’s big growth drivers. In its third-quarter earnings report last week, Google parent Alphabet said cloud revenue increased 34% from a year earlier to $15.15 billion, beating analyst estimates. The company ended the quarter with a business backlog of $155 billion.

“We are seeing substantial demand for our AI infrastructure products, including TPU-based and GPU-based solutions,” CEO Sundar Pichai said on the earnings call. “It is one of the key drivers of our growth over the past year, and I think on a going-forward basis, I think we continue to see very strong demand, and we are investing to meet that.”

Google doesn’t break out the size of its TPU business within its cloud segment. Analysts at D.A. Davidson estimated in September that a “standalone” business consisting of TPUs and Google’s DeepMind AI division could be valued at about $900 billion, up from an estimate of $717 billion in January. Alphabet’s current market cap is more than $3.4 trillion.

A Google spokesperson said in a statement that the company’s cloud business is seeing accelerating demand for TPUs as well as Nvidia’s processors, and has expanded its consumption of GPUs “to meet substantial customer demand.”

“Our approach is one of choice and synergy, not replacement,” the spokesperson said.

‘Tightly targeted’ chips

Customization is a major differentiator for Google. One critical advantage, analysts say, is the efficiency TPUs offer customers relative to competitive products and services.

“They’re really making chips that are very tightly targeted for their workloads that they expect to have,” said James Sanders, an analyst at Tech Insights.

Rasgon said that efficiency is going to become increasingly important because with all the infrastructure that’s being built, the “likely bottleneck probably isn’t chip supply, it’s probably power.”

On Tuesday, Google announced Project Suncatcher, which explores “how an interconnected network of solar-powered satellites, equipped with our Tensor Processing Unit (TPU) AI chips, could harness the full power of the Sun.”

As a part of the project, Google said it plans to launch two prototype solar-powered satellites carrying TPUs by early 2027.

“This approach would have tremendous potential for scale, and also minimizes impact on terrestrial resources,” the company said in the announcement. “That will test our hardware in orbit, laying the groundwork for a future era of massively-scaled computation in space.”

Dario Amodei, co-founder and chief executive officer of Anthropic, at the World Economic Forum in 2025.

Stefan Wermuth | Bloomberg | Getty Images

Google’s largest TPU deal on record landed late last month, when the company announced a massive expansion of its agreement with OpenAI rival Anthropic valued in the tens of billions of dollars. With the partnership, Google is expected to bring well over a gigawatt of AI compute capacity online in 2026.

“Anthropic’s choice to significantly expand its usage of TPUs reflects the strong price-performance and efficiency its teams have seen with TPUs for several years,” Google Cloud CEO Thomas Kurian said at the time of the announcement.

Google has invested $3 billion in Anthropic. And while Amazon remains Anthropic’s most deeply embedded cloud partner, Google is now providing the core infrastructure to support the next generation of Claude models.

“There is such demand for our models that I think the only way we would have been able to serve as much as we’ve been able to this year is this multi-chip strategy,” Anthropic Chief Product Officer Mike Krieger told CNBC.

That strategy spans TPUs, Amazon Trainium and Nvidia GPUs, allowing the company to optimize for cost, performance and redundancy. Krieger said Anthropic did a lot of up-front work to make sure its models can run equally well across the silicon providers.

“I’ve seen that investment pay off now that we’re able to come online with these massive data centers and meet customers where they are,” Krieger said.

Hefty spending is coming

Two months before the Anthropic deal, Google forged a six-year cloud agreement with Meta worth more than $10 billion, though it’s not clear how much of the arrangement includes use of TPUs. And while OpenAI said it will start using Google’s cloud as it diversifies away from Microsoft, the company told Reuters it’s not deploying GPUs.

Alphabet CFO Anat Ashkenazi attributed Google’s cloud momentum in the latest quarter to rising enterprise demand for Google’s full AI stack. The company said it signed more billion-dollar cloud deals in the first nine months of 2025 than in the previous two years combined.

“In GCP, we see strong demand for enterprise AI infrastructure, including TPUs and GPUs,” Ashkenazi said, adding that users are also flocking to the company’s latest Gemini offerings as well as services “such as cybersecurity and data analytics.”

Amazon, which reported 20% growth in its market-leading cloud infrastructure business last quarter, is expressing similar sentiment.

AWS CEO Matt Garman told CNBC in a recent interview that the company’s Trainium chip series is gaining momentum. He said “every Trainium 2 chip we land in our data centers today is getting sold and used,” and he promised further performance gains and efficiency improvements with Trainium 3.

Shareholders have shown a willingness to stomach hefty investments.

Google just raised the high end of its capital expenditures forecast for the year to $93 billion, up from prior guidance of $85 billion, with an even steeper ramp expected in 2026. The stock price soared 38% in the third quarter, its best performance for any period in 20 years, and is up another 17% in the fourth quarter.

Mizuho recently pointed to Google’s distinct cost and performance advantage with TPUs, noting that while the chips were originally built for internal use, Google is now winning external customers and bigger workloads.

Morgan Stanley analysts wrote in a report in June that while Nvidia’s GPUs will likely remain the dominant chip provider in AI, growing developer familiarity with TPUs could become a meaningful driver of Google Cloud growth.

And analysts at D.A. Davidson said in September that they see so much demand for TPUs that Google should consider selling the systems “externally to customers,” including frontier AI labs.

“We continue to believe that Google’s TPUs remain the best alternative to Nvidia, with the gap between the two closing significantly over the past 9-12 months,” they wrote. “During this time, we’ve seen growing positive sentiment around TPUs.”

WATCH: Amazon’s $11B data center goes live: Here’s an inside look

The post Google’s decade-long bet on TPUs company’s secret weapon in AI race appeared first on Our Story Insight.

Inside a secretive set of buildings in Santa Barbara, California, scientists at Alphabet are working on one of the company’s most ambitious bets yet. They’re attempting to develop the world’s most advanced quantum computers.

“In the future, quantum and AI, they could really complement each other back and forth,” said Julian Kelly, director of hardware at Google Quantum AI.

Google has been viewed by many as late to the generative AI boom, because OpenAI broke into the mainstream first with ChatGPT in late 2022.

Late last year, Google made clear that it wouldn’t be caught on the backfoot again. The company unveiled a breakthrough quantum computing chip called Willow, which it says can solve a benchmark problem unimaginably faster than what’s possible with a classical computer, and demonstrated that adding more quantum bits to the chip reduced errors exponentially. 

“That’s a milestone for the field,” said John Preskill, director of the Caltech Institute for Quantum Information and Matter. “We’ve been wanting to see that for quite a while.”

Willow may now give Google a chance to take the lead in the next technological era. It also could be a way to turn research into a commercial opportunity, especially as AI hits a data wall. Leading AI models are running out of high-quality data to train on after already scraping much of the data on the internet.

“One of the potential applications that you can think of for a quantum computer is generating new and novel data,” said Kelly. 

He uses the example of AlphaFold, an AI model developed by Google DeepMind that helps scientists study protein structures. Its creators won the 2024 Nobel Prize in Chemistry. 

“[AlphaFold] trains on data that’s informed by quantum mechanics, but that’s actually not that common,” said Kelly. “So a thing that a quantum computer could do is generate data that AI could then be trained on in order to give it a little more information about how quantum mechanics works.” 

Kelly has said that he believes Google is only about five years away from a breakout, practical application that can only be solved on a quantum computer. But for Google to win the next big platform shift, it would have to turn a breakthrough into a business. 

Watch the video to learn more.

The post How quantum computing could supercharge Google’s AI ambitions appeared first on Our Story Insight.

Sundar Pichai, CEO of Alphabet Inc., during Stanford’s 2024 Business, Government, and Society forum in Stanford, California, April 3, 2024.

Justin Sullivan | Getty Images

Google’s blowout earnings report in April, which sparked the biggest rally in Alphabet shares since 2015 and pushed its market cap past $2 trillion for the first time, tempered fear that the company was falling behind in artificial intelligence.

As executives enthusiastically talked about the results with Google’s employees at an all-hands meeting the following week, it was clear that Wall Street viewed things differently than the company’s workforce.

“We’ve noticed a significant decline in morale, increased distrust and a disconnect between leadership and the workforce,” one employee wrote in a comment that was read by executives at the meeting. “How does leadership plan to address these concerns and regain the trust, morale and cohesion that have been foundational to our company’s success?”

The comment was highly rated on an internal forum.

“Despite the company’s stellar performance and record earnings, many Googlers have not received meaningful compensation increases” another top-rated employee question read.

That meeting set the stage for what would be a year of contrasting takes from the company’s vocal workforce. As Google faced some of the most intense pressure its experienced since going public two decades ago, so too did CEO Sundar Pichai, who took the helm in 2015.

Pichai oversaw a steady stream of revenue growth this year in key areas like search ads and cloud. The company rolled out groundbreaking technologies, rounded out its AI strategy despite a slew of embarrassing product incidents and saw its stock price rise more than 40% as of Thursday’s close, ahead of the S&P 500 but trailing rivals Meta and Amazon.

Over the course of 2024, many staffers questioned Pichai’s vision following product mishaps in the first half of the year as well as internal shake-ups and layoffs, according to conversations with more than a dozen employees, audio recordings and internal correspondence. 

As the second half of the year progressed and Google rolled out a number of eye-catching AI products, Pichai’s standing improved, though some skepticism remains, sources told CNBC.

Google DeepMind chief Demis Hassabis (L) and Google chief executive Sundar Pichai open the tech titan’s annual I/O developers conference focusing on how artificial intelligence is being woven into search, email, virtual meetings and more. 

Glenn Chapman | AFP | Getty Images

The AI race pressure cooker

After the introduction of ChatGPT in late 2022, the tech industry saw an influx of AI products from Microsoft, with its Copilot AI assistant, and Meta, which placed its Meta AI chatbot in the search functions of its apps, as well as from hot startups like OpenAI and Perplexity.

The popularity of those tools has eaten into Google’s grip on U.S. search. The company’s share of the search advertising market is expected to dip below 50% in 2025, which would be the first time falling below that mark in more than a decade, according to research firm eMarketer.

Google responded to the pressures from new AI tools with offerings of its own. The company in 2024 rebranded its family of AI models as Gemini and released a number of products that were well received. But in its scramble to play catch-up, the company also released a pair of AI products that initially proved embarrassing. 

In February, Google launched Imagen 2, which turned user prompts into AI-generated images. Immediately after it was introduced, the product came under scrutiny for historical inaccuracies discovered by users. Notably, when one user asked it to show a German soldier in 1943, the tool depicted a racially diverse set of soldiers wearing German military uniforms of the era. 

The company pulled the feature, and Pichai told employees the company had “offended our users and shown bias,” according to a memo. Google said it would take a few weeks to relaunch Imagen 2, but it ended up being six months before it was revived as Imagen 3 in August. 

“We definitely messed up on the image generation,” Google co-founder Sergey Brin told a small crowd at a hacker house in March, in a video posted to YouTube. “It was mostly due to just not thorough testing.” 

The launch of AI Overview in May caused a similar reaction. 

That product showed users AI summaries atop Google’s traditional search results. Pichai hyped the product, calling it the biggest change to search in 25 years. Once again, users were quick to find problems.

When asked “How many rocks should I eat each day,” the tool said, “According to UC Berkeley geologists, people should eat at least one small rock a day.” AI Overview also listed the vitamins and digestive benefits of rocks.

Google responded by saying it would add more guardrails to AI Overview for health-related queries but said the mistakes weren’t hallucinations, and were rather just rare edge cases. Search Vice President Liz Reid told employees at an all-hands meeting in June that AI Overview’s launch shouldn’t discourage them from taking risks. 

“We should act with urgency,” Reid said. “When we find new problems, we should do the extensive testing but we won’t always find everything and that just means that we respond.”

Jaque Silva | Nurphoto | Getty Images

Beyond its AI blunders, Google also saw its greatest regulatory challenges to date in 2024.

In August, a federal judge ruled that the company illegally holds a monopoly in the search market. The Justice Department in November asked that Google be forced to divest its Chrome internet browser unit as a remedy for the ruling

The DOJ’s request represents the agency’s most aggressive attempt to break up a tech company since its antitrust case against Microsoft, which reached a settlement in 2001.

The remedies are expected to be decided next summer, and Google has said it will appeal, likely dragging out the situation a couple more years, but the company faces more antitrust hurdles. 

In a separate case, the DOJ accused the company of illegally dominating online ad technology. That trial closed in September and awaits a judge ruling. In October, a U.S. judge issued a permanent injunction that will force Google to offer alternatives to its Google Play app store for Android phones. After the ruling in October, Google won a temporary pause on the ruling, meaning it won’t have to open up Android to more app stores yet.

A search for vision

Amid the external pressure, Google notched some notable victories particularly toward the end of 2024, leading to a more positive sentiment from people within and outside the company.

Google successfully launched its most powerful suite of new Gemini models that underpin all of the company’s AI products, including its lightweight model Gemini Flash, which has been popular among developers. YouTube’s combined ad and subscription revenue over the past four quarters surpassed $50 billion. 

In the third quarter, Google saw the fastest-growing cloud business across the big tech players, up 35% over last year, with operating margins of 17%. The company has also seen double-digit revenue growth for each of the past four quarters and launched Trillium, its powerful sixth generation Tensor Processing Units, or TPUs, which were also found to have powered Apple’s AI models. 

Despite the blunders, AI Overview reached nearly 1 billion monthly users by the end of October. Demand for AI software has also driven consistent growth for the company’s cloud infrastructure. And Google launched an impressive video generation product, Veo 2, this month as well as an updated AI note-taking product, NotebookLM.

Beyond AI, Google in December announced Willow, a chip the company calls its biggest step in the march toward commercially viable quantum computing. The Waymo self-driving car unit was also a bright spot, expanding its robotaxi service to three cities and laying the groundwork for even more expansion in 2025. The company has delivered 4 million fully autonomous rides this year, with plans to commercially launch in Austin, Texas, and Atlanta next year.  

A Google quantum processor “Sycamore” is held up to the camera wearing blue gloves. In 2019, Google made a breakthrough in quantum computing. 

Peter Kneffel | Picture Alliance | Getty Images

But as Pichai approaches a decade running Google and starts his sixth year as CEO of parent Alphabet, questions remain about his ability to guide the company into the future.

Internally, employees routinely criticize leadership on the company’s Memegen messaging board, and some have aired their grievances publicly. 

“Google does not have one single visionary leader,” a Google software engineer wrote in a LinkedIn post earlier this year that received more than 8,500 reactions. “Not a one. From the C-suite to the SVPs to the VPs, they are all profoundly boring and glassy-eyed.”

In October, Google announced it would shake up the leadership of its ads and search division.

The company replaced longtime search boss Prabhakar Raghavan with Nick Fox, a deputy of Raghavan’s and a career Google employee. Raghavan was given the title of “chief scientist,” but internally, he is now listed as an “IC,” or individual contributor. 

Google also shifted the team working on its Gemini AI app to the Google DeepMind division, under AI head Demis Hassabis. Employees praised Pichai’s leadership shuffle, but some complained that the moves should’ve happened sooner.

Notably, some employees were perturbed when Raghavan addressed employees at an all-hands meeting in April, when he urged them to move faster, according to several people who spoke with CNBC. Raghavan noted that the staffers working to fix the failed Imagen 2 tool had increased their workloads from 100 hours a week to 120 hours to correct it in a timely manner.

Pichai has made efforts to get Google back to its nimble startup-like culture. 

When addressing employees, Pichai often name-checked co-founders Sergey Brin and Larry Page to remind them of Google’s scrappy roots. He’s flattened the company, removing 10% of middle management, according to audio of a December all-hands meeting. And in the spring, Pichai greenlit a hackathon, allowing employees to build using Google products that have yet to be announced. Pichai has also personally joined meetings with Google’s Labs team and enabled them to move quickly on products like NotebookLM, one of the company’s hit AI products in 2024.

Google Co-Founder Sergey Brin speaks during a press conference after the third game of the Google DeepMind Challenge Match against Google-developed supercomputer AlphaGo at a hotel in Seoul on March 12, 2016.

Jung Yeon-Je | AFP | Getty Images

After Brin’s hacker house appearance in March, some employees internally joked he should retake the helm, nostalgic for what they perceived as a visionary leader devoid of corporate speak. 

Brin co-founded Google with Page in 1998, but he stepped down as president of Alphabet in 2019. Brin, who remains a board member and a principal shareholder with a stake worth more than $140 billion, began appearing more frequently on campus starting in 2023, as part of an effort to help ramp up Google’s position in the hypercompetitive AI market. Employees, particularly working in AI and DeepMind said they’ve seen Brin walking around the company’s Mountain View, California, headquarters throughout the year and have been able to ask him questions for projects they’re pursuing.

Despite Brin’s reemergence, several employees told CNBC they’re doubtful he could adequately run what has become an increasingly larger and complex corporation. 

Employees said that although Pichai didn’t strike them as particularly visionary or as a wartime leader, it’s hard to find someone better suited for the job, given all the complexities of Alphabet. The key quandary remains: move too early and risk widespread criticism; move too late and risk missing the boat.

Culture Clashes

Through the year, morale inside Google wavered. Efforts to cut costs across the company in order to invest more in AI resulted in some teams feeling bifurcated and created yet another challenge for Pichai.

Within the company’s AI and DeepMind divisions, morale is mostly high, according to employees, boosted by hefty investments. Elsewhere, the vibes have been marred by cost cuts, bureaucracy and declining trust in leadership, employees said. 

DeepMind and AI teams have held off-sites, team-building activities, and have much bigger travel and recruiting budgets, people familiar with the matter said. In the spring, the company moved employees out of an eight-story office on San Francisco’s waterfront Embarcadero street and replaced them with AI and AI adjacent teams. 

Google DeepMind co-founder and Chief Executive Officer Demis Hassabis gives a conference during the Mobile World Congress (MWC), the telecom industry’s biggest annual gathering, in Barcelona on February 26, 2024.

Pau Barrena | Afp | Getty Images

A meme posted internally in November summed it up. 

The meme featured a photo of the cast of “Wicked” actors, where one, labeled “execs” looked longingly at one fellow actor labeled “Gemini” while ignoring the other beside her, which was labeled as “users.”

A Google spokesperson contested the idea that AI workers are receiving favorable treatment and said higher travel and recruiting budgets are not exclusive to AI teams or DeepMind. 

“Most Googlers, regardless of team, continue to feel positively about our mission and the company’s future, and are proud to work here,” the spokesperson said. 

A few employees say they’re no longer incentivized by the prospects of landing a promotion, which have become harder to achieve, and rather by the hope of avoiding layoffs. 

Despite slashing 12,000 jobs, or roughly 6% of its workforce, in 2023, Google has continued eliminating roles this year. In her first public statements as Google’s CFO, Anat Ashkenazi, told Wall Street in October that one of her top priorities would be to drive more “cost efficiencies” across the company in order to invest more in AI.

“I think any organization can always push a little further and I’ll be looking at additional opportunities,” Ashkenazi said.

That month, Google posted a job listing for a “Central Reorg Support Team Partner.” The responsibilities of that fixed-term contract position would include consulting with local HR teams and noted the need for the support staff’s “ability to operate with empathy and diffuse/de-escalate challenging conversations/situations.” 

“Hire the smartest people so they can tell us what to do,” one employee wrote on the internal forum in meme-style font atop the images of Brin and Page. “Hire a reorg consultant so they can tell us how to layoff the smartest people,” another said. 

Google ultimately took the job listing down.

Pro-Palestinian protesters are blocked the Google I/O developer conference entrance to protest Google’s Project Nimbus and Israeli attacks on Gaza and Rafah, at its headquarters in Mountain View, California, United States on May 14, 2024. 

Tayfun Coskun | Anadolu | Getty Images

Touting its AI technology to clients, Pichai’s leadership team has been aggressively pursuing federal government contracts, which has caused a heightened strain in some areas within the outspoken workforce since the beginning of the year.

Google terminated more than 50 employees after a series of protests against Project Nimbus, a $1.2 billion joint contract with Amazon that provides the Israeli government and military with cloud computing and AI services. Executives repeatedly said the contract didn’t violate any of the company’s “AI principles.”

However, documents and reports show the company’s agreement allowed for giving Israel AI tools that included image categorization, object tracking, as well as provisions for state-owned weapons manufacturers. Earlier this month, a New York Times report found that four months prior to signing on to Nimbus, officials at the company worried that signing the deal would harm its reputation and that “Google Cloud services could be used for, or linked to, the facilitation of human rights violations.”

In an all-hands meeting in April, a highly rated question asked why employees who did not participate in the protests were also fired, which was reported and cited in a National Labor Relations Board complaint from affected employees. Chris Rackow, Google’s security chief, took the stage at the all-hands and rebutted those claims.

“This was a very clear case of employees disrupting and occupying work spaces, and making other employees feel unsafe,” a Google spokesperson told CNBC, adding that the company “carefully confirmed” that every person terminated was involved in the protests. “By any standard, their behavior was completely unacceptable.”

That round of job eliminations underscored Google’s clampdown on internal discussions related to hot-button topics, including politics and geopolitical conflicts, which was encouraged by executives several years prior.

One internal meme that got more than 2,000 likes, compared Google to Star Wars’ Anakin Skywalker. The meme shows an image of a smiling childhood Skywalker, framed by one of the company’s original, colorful employee badges. The meme progresses Skywalker’s age in two later versions of the badge. 

The final badge shows Darth Vader working for “Google,” spelled out in the font of IBM’s logo.

Don’t miss these insights from CNBC PRO

The post How Google’s Sundar Pichai navigated a pressure-filled year appeared first on Our Story Insight.