A Audi deu início a um projeto de “prova de conceito” para testar o uso de robôs humanoides móveis na área de logística de sua fábrica em Ingolstadt, na Alemanha.
A iniciativa, desenvolvida em parceria com a empresa de robótica Agile Robots e as companhias de tecnologia Nvidia e Dassault Systèmes, busca automatizar o processo de sequenciamento de peças, etapa que exige cerca de um milhão de seleções diárias (picks) para abastecer a linha de montagem dos veículos premium da marca no ritmo atual.
O projeto utiliza robôs dotados de dois braços mecânicos e inteligência artificial, instalados em uma base móvel. A tecnologia foi projetada para atuar na mesma altura de operadores humanos e se integrar às instalações existentes da unidade sem requerer mudanças na infraestrutura fabril.
Simulação em ambiente digital acelera o processo
Audi testa robôs humanoides e IA para otimizar linhas de montagem em parceria com Agile robots, NVIDIA e Dassault Systèmes • Divulgação Audi
Antes de entrarem em operação prática no laboratório de produção da Audi (Production Lab), as tarefas dos robôs são validadas em um ambiente inteiramente virtual. Por meio das plataformas digitais da Dassault Systèmes e das bibliotecas da Nvidia, a equipe de engenharia criou um “gêmeo digital” da fábrica.
O modelo virtual permite simular ergonomia, tempo de execução, viabilidade operacional e protocolos de segurança. A utilização de inteligência artificial aplicada ao ecossistema digital reduz o tempo de homologação desses testes de semanas para poucos dias.
Manipulação de componentes flexíveis e caixas mistas
Audi testa robôs humanoides e IA para otimizar linhas de montagem em parceria com Agile robots, NVIDIA e Dassault Systèmes • Divulgação Audi
A automação do sequenciamento representa um dos maiores desafios logísticos da indústria automotiva, por envolver peças semirrígidas, itens flexíveis e componentes dispostos de forma desordenada em caixas mistas.
Os robôs móveis desenvolvidos para a operação utilizam sistemas de percepção por inteligência artificial que identificam, alinham e movimentam peças variadas com precisão.
O objetivo do programa é transferir tarefas repetitivas de triagem para os sistemas automatizados, permitindo que os operadores humanos foquem em etapas de maior complexidade do processo produtivo.
Nvidia vice president Laura Fay has retired after a decade at the company.
Fay grew Nvidia's enterprise marketing team from roughly 60 to 600 employees.
Nvidia brought in Google Cloud's Alison Wagonfeld as its first chief marketing officer this year.
Laura Fay, a longtime Nvidia marketing executive who reported to CEO Jensen Huang, retired in July after roughly a decade at the AI chip giant.
Fay formerly served as Nvidia's vice president of global enterprise marketing, overseeing campaigns, events, analyst relations, and partner marketing. She said her team grew from roughly 60 to 600 people by the time she left.
Fay was among 36 employees on an internal list obtained by Business Insider last year of company leaders reporting directly to Huang. Prior to joining Nvidia, she spent 22 years at Cisco, concluding her tenure there as VP of global marketing.
Fay said Huang recruited her to build Nvidia's enterprise marketing organization as the company looked to expand beyond gaming.
"I want to change the world, and I want you to change it with me," she recalled him telling her on their first call.
Fay said she began considering retirement around her 10th anniversary in January, though a call from Huang congratulating her and expressing hope that she'd stick around for another decade persuaded her to stay a little longer.
Nvidia declined to comment.
Fay's retirement came as Nvidia also brought in new marketing leadership. In February, the company hired Google Cloud executive Alison Wagonfeld as its first chief marketing officer.
The Wall Street Journal reported at the time that Wagonfeld would oversee all marketing and communications teams and assume responsibilities previously divided among several leaders.
Those changes follow another longtime marketing departure. Greg Estes, Nvidia's former vice president of corporate marketing and developer programs, retired in 2025 after roughly 15 years at the company.
Several other longtime Nvidians have retired this year, including sales chief Jay Puri, chief accounting officer Donald Robertson, and VP of hardware engineering Robert Huang. Some retirees told Business Insider that Nvidia's exponential stock run helped precipitate their decisions.
Nvidia's communications organization has also undergone a shake-up as scrutiny around the company has intensified, with roughly 15 employees departing since 2025 amid a reorganization under new leadership.
Have a tip? Contact this reporter via email at gweiss@businessinsider.com or Signal at @geoffweiss.25. Use a personal email address, a non-work WiFi network, and a non-work device; here's our guide to sharing information securely.
Germany-based robotics firm Agile Robots reportedly plans to produce up to 2,000 humanoid robots next year through a fresh partnership with Nvidia, targeting universities and corporate research labs.
The collaboration was announced at Nvidia’s GTC conference in Berlin. Agile CEO Zhaopeng Chen told Reuters the pair are set to begin production of the Agile ONE S and ONE SN humanoid robot models by the end of 2026, expecting to manufacture between 1,000 and 2,000 units. Stanford University and ETH Zurich are among potential customers.
The Agile ONE SN, with the N standing for Nvidia to denote the chipmaker-specific configuration, combines Agile’s humanoid hardware with Nvidia’s Jetson Thor onboard computing platform and Isaac GR00T robotics AI model. It features five-fingered hands, cameras and other sensors to support dexterous manipulation and interaction with the physical environment.
Standing 1.4 metres tall and weighing around 40 kilograms, the robot is designed as a smaller alternative to full-sized humanoids. Nvidia said the robot intends to simplify the integration of hardware, simulation, training and deployment tools on a single platform, “cutting setup time from weeks to days”.
Spencer Huang, Nvidia’s product director for robotics, told Reuters the standardised configuration could save researchers two or three months of work procuring separately purchased components. “If we can bring researchers something that’s already 80% of the way there, it accelerates their development,” he said.
Nvidia said Agile’s ONE SN is scheduled to become available in early 2027, as competition intensifies among European humanoid developers.
ChatGPT moment Alongside the partnership, the chipmaker also revealed its expanded robotics development platform, Nvidia Isaac, equipped with new tools for training and deploying AI-powered machines. The expanded platform includes Isaac OS, an open-source runtime framework, and Isaac Lab 3.0, a robot learning framework designed to train and refine robot skills. The company also introduced new reference workflows and simulation resources to speed up development.
“General-purpose robotics is advancing at extraordinary speed — the ChatGPT moment for robotics is fast approaching,” noted Jensen Huang, CEO of Nvidia, adding the platform would give developers “the models, data, simulation tools and computing systems to bring physical AI to the world’s largest industries”.
The chipmaker added European robotics players including UK-based Humanoid, French start-up Hugging Face and Swedish firm ABB Robotics are applying its technologies including Nvidia Isaac to industrial autonomous machines.
O Banco Santander está a reforçar a parceria com a Nvidia para desenvolver modelos de inteligência artificial especializados em tarefas bancárias, com o objetivo de melhorar a qualidade do atendimento ao cliente e reduzir custos operacionais, anunciou o banco numa nota de imprensa esta quarta-feira.
No âmbito do Santander AI Lab, a instituição está a testar e a aperfeiçoar modelos e agentes de IA adaptados a necessidades específicas da banca, passando progressivamente de grandes modelos de uso geral para versões mais pequenas e especializadas.
Em casos de uso experimentais, esta abordagem permitiu reduzir os custos de processamento de cerca de 5 euros por interação para aproximadamente 0,03 euros, mantendo elevado desempenho. Num projeto-piloto de atendimento ao cliente, o banco melhorou a qualidade das conversas e cortou os custos de serviço em mais de 50%.
O Santander criou a plataforma Model Arena para avaliar modelos de pesos abertos (open-weight) em tarefas bancárias e identificar o mais adequado a cada caso. O banco está também a especializar modelos com o Nvidia NeMo Customizer e a usar a família de modelos NVIDIA Kumo para dados estruturados, além de industrializar a inferência com a solução Nvidia Dynamo-Triton.
Para dar continuidade ao trabalho, as duas empresas firmaram um acordo-quadro de colaboração de três anos. O entendimento abrange modelos fundacionais, agentes de IA, prevenção de fraude e combate ao branqueamento de capitais, tomada de decisões em tempo real e tecnologias de simulação, incluindo o conceito de “digital twin”.
“A Nvidia é um parceiro fundamental nas infraestruturas, no software e na investigação, impulsionando de forma significativa os mais recentes avanços em inteligência artificial”, afirmou Iñaki Bernal, responsável global de Inteligência Artificial do Santander.
“Ao conjugarmos estas capacidades com a escala do Santander em áreas de negócio, clientes e geografias, acreditamos estar numa posição única para transformar, com sucesso, esses avanços em capacidades bancárias concretas”, acrescenta-
Já Aser Blanco, responsável global pelo setor bancário na Nvidia , destacou que a banca é um dos setores mais intensivos em dados e que “o custo por resultado de inteligência é uma métrica fundamental”. Segundo ele, o Santander está a demonstrar como a abordagem com modelos mais pequenos e especializados com dados próprios funciona em produção e à escala de um banco global.
O Santander AI Lab continua a desenvolver processos rigorosos de experimentação, otimização de desempenho e testes adversariais a sistemas de IA conversacional, antes de implementar à escala as soluções que demonstrem valor em todo o Grupo.
INTERVIEW: Nokia chief technology and AI officer Pallavi Mahajan and Nvidia SVP of telco Ronnie Vasishta told the Mobile World Live Podcast their AI-RAN partnership is showing tangible progress, as they announced close to 20 operators are lined up for lab and commercial trials of the technology.
In an exclusive interview to outline progress of their efforts, Vasishta and Mahajan highlighted the fact it had only been 12 months since the two companies announced a joint AI-RAN partnership, which included a $1 billion investment from Nvidia into the Finnish vendor.
A year on, today (21 October) at Nvidia’s GTC 2026 event in Berlin the duo announced almost 20 operators globally have committed to AI-RAN collaboration, including 10 from EMEA. Mahajan said these operators are at different stages, lined up for both lab and commercial trials.
In 2027 trials are expected from BT Group, Deutsche Telekom and Elisa, with the latter preparing for commercial deployment.
In the US, T-Mobile has completed what was claimed as the world’s first AI-RAN call on a live network, while in France Orange is introducing new voice and intelligence services for enterprises based on the technology.
Speed of light Vasishta told MWL it was hard to overhype development based on the progress it had seen over the past year, and the number of trials and deployment is only set to grow.
“A year ago, or two years ago for sure, I think people would have scoffed at the idea. Nokia talking about operator commercialisation with AI-RAN, I think, that would have been perhaps seen as somewhat sceptical.”
Mahajan concurred, describing the pace of development as occurring “at the speed of light” with the game changer centred on being able to outline the true “value of what AI brings in terms of making RAN efficient” such as significant spectrum efficiency gains.
“In the next few years, the distinction between the networks and computing infrastructure will just blur,” she said, inevitably opening up a host of operator opportunities across applications beyond 5G and 6G.
Vasishta added Nvidia had started on the journey of software-defined RAN eight years ago and the inclusion of Nokia in 2025 had accelerated this dramatically.
“We see all operators around the world are created differently. They service different markets, they have different DNA, but overall, all of them have gravitated towards this concept now of AI RAN.”
Imagine if Sam and Dario and Elon and Mark, along with their boards of directors, bore individual liability for the damage their companies may inflict on society.
Imagine if Sam and Dario and Elon and Mark, along with their boards of directors, bore individual liability for the damage their companies may inflict on society.
Jensen Huang interacted with Microduck robots onstage.
Sean Gallup/Getty Images
Nvidia CEO Jensen Huang's Berlin keynote was interrupted by Greenpeace protesters onstage.
Huang was making the case for Europe to build an AI economy with factories and agents.
His finale featured a wobbly Microduck robot demo and an AI-generated Eurovision-style video.
Even for a seasoned keynote pro like Jensen Huang, there's always the risk that things don't go fully to plan.
The Nvidia CEO opened the company's Berlin conferenceon Wednesday to make the case that Europe should build its own AI economy — complete with AI factories, autonomous agents, and robots.
But the pitch was interrupted about 20 minutes into the keynote by protesters holding a Greenpeace banner reading, "Nvidia, clean your dirty business!"
One protestershouted: "Clean up your act! What about the technology that was found in Russian drones? Those were Nvidia chips. What are you doing about that? What are you doing about the environmental destruction?"
Greenpeace protesters held a banner, reading "Nvidia, clean up your dirty business!"
John MACDOUGALL / AFP via Getty Images
The protester appeared to be referring to an August statement from Ukraine's Defence Intelligence, which said it had identified an Nvidia Jetson Orin computer module in a Russian S-71 Monochrome cruise missile. The agency did not allege that Nvidia knowingly supplied the component to Russia.
Greenpeace said in a press release that it's calling on Nvidia to extend its 100% renewable energy target across its wider supply chain and to conduct audits to ensure its technology doesn't end up in military equipment. Nvidia did not immediately respond to Business Insider's request for comments.
As security escorted the protester away,Huang said, "Give her a chance," and that "everybody has a chance to say something." He then told the applauding crowd: "All important issues … but let me get through my keynote."
Huang, after joking that he had to "torture" the audience with "some education" before "the entertainment,"then pressed on with Nvidia's pitch for Europe's AI future. He unveiled DSX POD, a modular system for AI factories, and promoted OpenShell, a platform designed to help companies run autonomous AI agents securely.
'Pay attention!'
The entertainment came near the end of the keynote, when Huang brought four small robots, known as Microducks, from French startup Pollen Robotics onstage.
The demo was charming, but not entirely obedient.
When Huang told the robots to "take a nap," one kept walking before eventually falling alongside the others. Later, only two appeared to perform a somersault when he asked them to.
Another robot lingered farther away from the group and tripped before getting back up. "Pay attention! Stay focused. Come here, come here. Stay focused," Huang told it. It eventually approached and sat down, drawing laughter from the audience.
With Nvidia’s GTC Berlin event in full swing, the company’s telecoms boss Ronnie Vasishta and Nokia chief technology and AI officer Pallavi Mahajan sit down with the Mobile World Live Podcast in an exclusive interview to reveal the next stage of the pair’s AI-RAN journey, as they team up with European heavyweights to turn networks into distributed intelligence platforms.
AWS expected to get $5 billion of Nvidia GB300 racks, followed by $3 billion more each month.
Some of these GPU systems are being stored elsewhere due to lack of data center power.
Amazon CEO Andy Jassy sees AI as a major opportunity, but getting capacity online is a challenge.
Amazon is racing to secure AI hardware earlier than before, even as data center power shortages leave some GPU systems waiting in warehouses.
Under a new Early Rack Receipt process, AWS expected to receive roughly $5 billion in Nvidia's GB300 server racks at US facilities before final deployment, according to August internal emails reviewed by Business Insider. The company projected roughly $3 billion more in GB300 racks each month "for the foreseeable future," one of the emails said.
Some AWS sites had room for more racks but not enough available power to turn them on, people familiar with the effort said, leaving finished equipment to be staged at intermediary warehouses until the facilities were ready.
At least one AWS team did not have "a clear forecasted path" for getting the incoming racks into data centers at the time, according to one of the emails.
Dark GPUs
A lack of new electricity capacity is one of several bottlenecks slowing the buildout of some AI data centers. This has led to a phenomenon known as dark GPUs, where AI chips and servers are purchased but can sit idle due to power supply issues or other challenges. (Dark GPUs can also refer to system design flaws that mean this gear isn't fully used after being plugged in).
This is partly why financial pressure is mounting on big data center projects. Money must be spent upfront to secure supplies of Nvidia GPUs and other valuable AI components. But if some data centers aren't ready to run, those early investments can tie up huge amounts of cash. Debt financing can fill in the gaps, but the market has become less willing to lend against some of these projects lately.
Trade-off
Amazon's early-rack program is one piece of the cloud leader's broader AI buildout, which is putting new pressure on its cash flow, as Business Insider exclusively reported this week. Amazon expects roughly $220 billion in capital expenditures this year — more than any Big Tech peer. Its free-cash-flow turned negative last quarter for the first time since 2023, and could stay negative for a while.
For AWS, the trade-off is securing scarce GPU racks before competitors do, even if some data centers are not yet ready to turn them on. But taking ownership earlier can tie up billions in equipment before it starts generating cloud revenue.
Previously, AWS took ownership of a finished rack only upon delivery to a data center, according to internal documents. Under the new approach, it can pay for and record certain racks earlier in the supply chain.
An Amazon spokesperson said the company routinely stages finished hardware near deployment sites so it can install it once capacity becomes available, calling the practice "best business practice." The spokesperson added that Amazon has long-established systems for tracking infrastructure assets and disputed that the broader company lacked a path for getting racks into data centers.
Amazon CEO Andy Jassy previously said the company has a "clear line of sight" to strong returns from its AI investments, calling the technology a "once in a lifetime opportunity."
"We've been through this cycle with the first big AWS growth wave, and liked the results," Jassy wrote in Amazon's annual shareholder letter. "We expect to feel similarly about this next wave."
Nvidia CEO Jensen Huang urged Europe to move faster on AI in a conversation with Germany Chancellor Friedrich Merz on Monday.
Nadja Wohlleben/Getty Images
Nvidia CEO Jensen Huang wants Europe to be less cautious in its approach to AI innovation.
Huang told Germany's chancellor that the country and others in Europe are overreliant on regulation.
"I think you overthink things," he told British journalists on a trip to the UK last week.
Cast off the shackles and be bolder, or risk losing the AI race for good.
That's the pointed message Nvidia CEO Jensen Huang has delivered to European leaders during his trip to the continent over the past week.
Over the last seven days, Huang has visited France, Sweden, the UK, and Germany ahead of Nvidia's GTC event in Berlin, which began Tuesday.
During that time, he's been clear that he believes the continent needs to move faster on artificial intelligence and rely less heavily on regulating the sector.
Huang made the case most directly on Monday evening in Berlin, during a public conversation with German Chancellor Friedrich Merz.
Asked how Europe should regulate AI, Huang said safety had to be paramount, but cautioned that policymakers could not regulate their way to technological leadership.
"There's nothing to regulate if the industry doesn't thrive," Huang told Merz, arguing that Germany needed to foster a domestic AI industry while ensuring it developed safely and responsibly.
He said that policymakers should focus on regulating AI applications, similar to the way they regulate products that use the internet, rather than trying to regulate the underlying technology wholesale.
"We should be regulating the applications right away," Huang said.
Merz replied that Europe was still trying to find the balance between innovation and regulation. The German leader said he did not believe "no regulation" was the right answer, but that overly tight rules were not either.
Huang’s response: "Give Germany a chance. Give this industry a chance."
The exchange came as Huang and Merz pitched AI as a potential answer to the economic malaise that led Germany into recessions in 2023 and 2024.
Huang said the country's manufacturing sector, engineering expertise, and industrial know-how are national resources, and that AI could help its companies turn them into an advantage in the next phase of the technology race.
"The next wave of AI is physical," Huang said, referring to robotics, manufacturing, cars, and heavy industry. "Electrons will never replace atoms."
It was the same urgent message Huang had delivered in London days earlier. Speaking to reporters last week, Huang said there was a "strange general behavior" going on in the UK.
He took a swipe at what he described as Britain’s tendency to dwell on failure.
"You have smart people, you have a great culture, you have young people, and the energy is vibrant. But for some reason, if there was a million reasons to win, you find the one reason not to do it."
"I think you overthink things," he added.
The remarks underscore a familiar tension in Europe’s approach to AI. The continent has moved early to establish sweeping rules for the technology, including the EU AI Act, while the US and China have pulled ahead in developing frontier models and the infrastructure needed to train them.
Huang, whose company dominates the market for the advanced chips powering the AI boom, is hardly a neutral observer. Nvidia stands to benefit enormously if governments and companies move faster to build AI infrastructure and deploy its technology.
Regardless of his stake in Europe's AI success, Huang's message was blunt. If the continent spends too long debating how to regulate AI, it may find that the meaningful innovation is done elsewhere.
Prasad Kalyanaraman, vice president of infrastructure services at Amazon Web Services
BASTIEN INZAURRALDE/AFP via Getty Images
Amazon is on an unprecedented spending spree to build more AI data center capacity than anyone else.
The company has extensive experience, yet even it's struggling to bring capacity online fast enough.
Internal documents show Amazon getting bilions of dollars of Nvidia GPUs before they can be deployed.
At an internal staff meeting in September, AWS CFO John Felton had a joke for Prasad Kalyanaraman, the executive overseeing much of Amazon's AI infrastructure expansion.
"You've taken us free cash flow negative as a company, so congratulations," Felton said, according to a recording of the meeting obtained by Business Insider.
The remark was lighthearted. But it captured a stunning reversal.
Just two years ago, Wall Street was debating what Amazon would do with its growing cash hoard. Now, cash flow has evaporated as Amazon embarks on its biggest bet ever, spending $220 billion this year alone to build the physical backbone for AI.
At the meeting, Kalyanaraman did not dispute Felton's premise. Amazon's machine-learning and AI business is "hugely free cash flow negative," he said, adding that he expects the company's free cash flow to remain negative for "the next couple of years or so."
Amazon has made big bets before. It spent heavily on cloud computing, fulfillment centers, and its delivery network long before they became some of the company's most valuable assets. Investors questioned those investments at the time, but Amazon ultimately built enormous, profitable businesses from them.
AI is testing this playbook at a far larger scale. Amazon is committing hundreds of billions of dollars to data centers, servers, and power before investors know how quickly those investments will generate revenue.
The company is the best, or among the best, at building data centers, with decades of experience. But the size and complexity of these projects, and the bottlenecks holding progress back, are challenging — even for such a respected cloud pioneer.
Internal documents reviewed by Business Insider and people familiar with the matter show the company trying to coordinate a sprawling industrial system whose parts move on different schedules. A delay in power, equipment, or construction can leave billions of dollars in hardware waiting — and Amazon's cash tied up — while forecasts and costs keep shifting.
Why Amazon thinks it will work
Amazon says it has strong evidence that the capacity will be needed. AWS revenue grew 37% in the second quarter, its fastest pace in 18 quarters, while its backlog reached $496 billion.
CEO Andy Jassy thinks the company has a "clear line of sight" to strong financial returns from this record investment cycle. In his latest annual shareholder letter, he called AI a "once in a lifetime opportunity" and said Amazon would not be conservative in its investments.
"We've been through this cycle with the first big AWS growth wave, and liked the results," Jassy wrote. "We expect to feel similarly about this next wave."
Investors are concerned by the scale of the industry's expansion, especially as AI data centers are costing more to build and taking longer to complete.
The market is "repricing anything that funds multi-year builds against revenue that arrives later," Dan Ives, an analyst at Yorkville Ives, wrote in a recent report. Growing uncertainty around data center permitting and the political fight over electricity prices have given investors another reason to question how quickly these investments will pay off, he added.
Racks, but not enough power
One major industry challenge is securing enough power to run the equipment that goes inside data centers. This includes expensive racks of servers that contain GPUs and other chips that train and run AI models and services such as ChatGPT.
Internal emails from August show Amazon preparing a new "Early Rack Receipt" process for an expected wave of server racks carrying Nvidia's GB300 chips at US facilities before final deployment.
One Amazon email projected the company would take delivery of about $5 billion of Nvidia GB300 racks in August, followed by another roughly $3 billion each month "for the foreseeable future."
Some AWS data center sites had room for more racks but not enough available power to turn them on, people familiar with the effort said, leaving the equipment to be staged at intermediary warehouses until the facilities were ready. These people asked not to be identified discussing sensitive matters.
Timing and coordination
Usually, Amazon doesn't pay for racks until the gear is put into a data center, leaving suppliers to carry the cost of the equipment in transit, according to the internal documents obtained by Business Insider. This helps the company conserve cash.
However, the intensity of the AI data center building boom has made this type of coordination harder. Amazon has to compete more for scarce resources as it jockeys with rivals for access to Nvidia GPUs and other AI components.
Under the new "Early Rack Receipt" process, Amazon now takes ownership of these racks earlier in the supply chain, rather than waiting until the equipment reaches a data center.
The means spending cash sooner. And any delays getting the racks installed and plugged into a reliable power source potentially pushes back the moment when revenue starts flowing from these assets.
At least one AWS supply chain team did not have "a clear forecasted path" for getting racks into data centers, according to one of the emails reviewed by Business Insider.
An Amazon spokesperson said the company routinely stages finished hardware near deployment sites so it can be installed once power capacity is available, calling that approach "best business practice."
The spokesperson added that AWS has "long-established systems" to track infrastructure assets and disputed that the broader company lacks a forecasted path for getting racks into data centers.
This problem is not unique to Amazon.
Bernstein analysts wrote in a recent note that power permits, cooling systems, electrical equipment, construction, and hardware all move on separate schedules.
"Timing is imperfect," the analysts said.
"Inherently uncertain"
John Felton, Chief Financial Officer of Amazon Web Services
Chris Jung/NurPhoto via Getty Images
During the internal staff meeting in September, Felton, the AWS CFO, described the broader problem eloquently: "How do you work in the future when the future is inherently uncertain?"
Kalyanaraman, the AI infrastructure leader, said Amazon's internal outlook for the return on its AI investments had improved. But the assumptions behind those returns keep shifting.
"The forecast is changing on a daily basis," Kalyanaraman said. "The world that we live in is extremely dynamic."
Those shifting assumptions include when data centers will be ready and how much equipment AWS needs to keep on hand, he explained.
An eight-week supply buffer might be routine for a smaller company. But at Amazon's scale, Kalyanaraman said, this could mean holding "billions and billions of dollars" worth of additional equipment inventory — tying up precious cash.
Frugality meets the AI boom
Such tensions are especially striking for Kalyanaraman, a 21-year Amazon veteran who earlier this year was promoted to the elite S-team that reports directly to Jassy.
Amazon's founding story is defined by frugality. It famously used cheap doors as desks when founder Jeff Bezos started in Seattle in the mid-90s.
Kalyanaraman has absorbed this lore into bones. In late 2023, as AWS faced slower growth and layoffs, he urged employees to do more with less. "Frugality breeds innovation," he said back then.
That instinct still shapes his approach. In the September meeting this year, Kalyanaraman told Felton that AWS tries to make major investments at the "last responsible moment," when it has the clearest possible view of demand.
"If we can wait to spend the capital as late as possible, it's better for us," he said.
Amazon CEO Andy Jassy
Bloomberg/Getty Images
But the frenetic pace of the AI boom makes this principle difficult to follow. By the time demand is fully visible, the servers, power, and construction capabilities that AWS needs may no longer be available. That puts pressure on the company to commit more money sooner to secure supply and meet its data center capacity goals.
Kalyanaraman recalled telling Jassy in 2023 that AWS would "surely" be out of its capacity crunch by 2027. Instead, Amazon still expects demand to exceed its available capacity in 2027, the AI infrastructure chief said during the meeting this September.
Secured Power at Risk
Getting enough electricity may be the biggest challenge for an industry leader like Amazon.
The company has 44 gigawatts of data center capacity under development across 16 states, more than double Google, Meta, and Microsoft, according to Cleanview Research estimates.
An internal AWS dashboard called "Secured Power at Risk" tracks utility and regulatory constraints affecting electricity capacity AWS has lined up for its data centers.
The risks include transmission bottlenecks, requirements to reduce use during grid stress, and "use it or lose it" provisions that can jeopardize capacity if a data center project misses usage or construction milestones.
In other words, AWS is not simply waiting for electricity to arrive. It is managing a portfolio of commitments that can become more costly if a data center project is delayed.
Maryland proposals
A submission to Maryland's utility regulator last year, obtained through a public records request, illustrates the kind of exposure that can accompany massive data center buildouts.
Amazon told the Maryland Public Service Commission that large data center customers could help pay for grid upgrades or bring new power capacity to market without burdening other electricity ratepayers.
The company also proposed setting aside collateral to help utilities get more comfortable with the large investments needed to generate more electricity for data centers. Amazon proposed three years of collateral so that if a large electricity customer abandoned unused capacity, the utility would have time to re-sell the power.
The proposal is not evidence that AWS has made such a payment or posted collateral. It does, however, show that large data centers can face extra financial obligations, beyond their own electricity bills, if projects are delayed or cancelled.
The Amazon spokesperson said such utility rules are not unique to AWS. The company has not disclosed how much cash it has paid upfront to secure power.
New constraints 'keep popping up'
The AI buildout also incurs expenses that are harder to model than a conventional capex plan.
In the September meeting this year, Kalyanaraman said that while reviewing plans for gigawatts of capacity in West Texas, AWS confronted a labor problem: There weren't enough nearby pipefitters, electricians, and other skilled workers to build the facilities.
The company's response, he said, was to pay to construct housing complexes near its data centers so workers could live there for the two or three years needed to complete the projects.
Even the crates used to ship servers have become a planning issue. It may seem absurd to worry about the price of a container when servers cost hundreds of thousands of dollars, Kalyanaraman said. But at AWS's scale, where the company needs so many crates, even trivial items can become a material expense.
"We have new constraints that keep popping up," Kalyanaraman said.
Mercedes-Benz expanded its partnership with Nvidia to develop driver assistance and automated driving systems based on the carmaker’s MB.OS platform and the chip company’s computing technology and software.
The agreement combines Mercedes-Benz engineering and its MB.OS operating system with Nvidia’s Hyperion computing and sensor architecture, Alpamayo driving software and Halos safety system.
Mercedes CTO Jorg Burzer hailed Nvidia as its “most important strategic partner in automated driving”, describing the urban point-to-point assistance system the two built as the first of its kind from Mercedes for Western markets.
He added the extension is a milestone built on years of mutual trust, while stressing customers will still expect the safety standards the brand is known for.
Ali Kani, VP of automotive at Nvidia, said accelerated computing could help vehicles interpret their surroundings and respond to complex driving situations. He added the companies aim to make autonomous driving suitable for everyday use.
MB.OS will manage driving functions and over-the-air updates alongside Nvidia’s Hyperion in-vehicle computing platform. Alpamayo will provide software for urban point-to-point assistance, active safety and automated parking, while Halos will add a safety layer alongside Mercedes-Benz’s own systems.
The companies noted the technologies will also support future SAE Level 4 applications, including robolimousines.
Mercedes-Benz also uses Nvidia technology for vehicle-engineering simulations and in its MO360 digital-production platform, covering virtual planning and AI applications. The companies plan to expand the work across the carmaker’s global production network.
Nvidia's communications operation has been reshaped as the company became the face of the AI boom.
Roughly 15 staffers have left the lean team since 2025 as Nvidia reorganized its communications org.
Former employees described growing workloads as responsibilities expanded.
Nvidia's communications organization has been reshaped as the company became the face of the AI boom, with multiple teams brought under new leadership and a string of departures.
Business Insider identified roughly 15 communications employees who have left Nvidia since 2025, based on LinkedIn profiles and interviews with former employees.
The organization now numbers roughly 70 to 80 people globally, and has more than doubled in size since Mylene Mangalindan joined as vice president of corporate communications in 2023, according to a person close to the company.
The departures have also included newer hires. Anna Soellner, who was recruited to lead corporate communications, left after roughly four months, Business Insider previously reported.
Nvidia is known for high retention. It reported companywide turnover of 3.7% in fiscal 2026, up from 2.5% the prior year, and well below the tech industry average of 21.9% cited in its latest sustainability report.
Mangalindan succeeded longtime communications chief Bob Sherbin, who retired in 2024. She previously worked at the advisory firm Brunswick Group.
She was brought in to "bring fresh blood" and "upgrade the team" to meet Nvidia's growing profile, one former employee said.
At the end of 2024, Nvidia restructured its communications organization. It brought enterprise public relations under Mangalindan's leadership alongside the corporate communications team she already led, according to two former employees.
Corporate communications focuses on Nvidia's reputation and major media outlets, while enterprise handles product and industry messaging,including for its data center business.
One former employee said the consolidation reflected the growing importance of Nvidia's data center business, which sells chips and systems that power AI and had become its biggest revenue driver.
Analyst relations, social media, and influencer marketing also moved under Mangalindan's remit this year, another former employee said.
Three former employees said Nvidia's communications team had long operated with relatively lean staffing, even as workloads grew alongside the company's public profile, leaving staff juggling multiple responsibilities.
These employees said they experienced long hours, difficulty taking time off, and pressure to respond quickly to emails and Slack messages.
A person close to Nvidia said the company's lean staffing allows employees to move quickly and also requires long hours. The person said Mangalindan encourages staff to take vacation, though the team's small size can make scheduling time off difficult, particularly in the summer.
Nvidia is widely known for its demanding work environment. CEO Jensen Huang has emphasized operating at the "speed of light" and is known for fostering a culture of constant accountability.
Have a tip? Contact this reporter via email at gweiss@businessinsider.com or Signal at @geoffweiss.25. Use a personal email address, a nonwork WiFi network, and a nonwork device; here's our guide to sharing information securely.
Apple, Nvidia and EchoStar chair Charlie Ergen reportedly committed $25 million each for the construction of US President Donald Trump’s White House ballroom project, which has raised $333 million to date in private donations.
The Washington Post reported “the vast majority of the funds raised came from companies and wealthy individuals with interests before the federal government” after obtaining an internal spreadsheet used to track contributions for the fundraising effort.
The Post noted more than three-quarters of the funding was pledged in amounts of at least $5 million, with Amazon, Comcast, Microsoft, and T-Mobile US among the largest corporate donors.
It stated Nvidia and Ergen, among others, requested their donations be kept anonymous
The news site noted companies or executives of firms that have had contracts with the federal government during Trump’s second term accounted for about half of the total funds raised for the project which also includes office space and an underground military facility.
According to internal contractor documents obtained by the Post, the total cost of the project is now $600 million after Trump initially stated he and wealthy donors would pay the estimated $400 million project, although Reuters reported he later unsuccessfully asked Congress for $1 billion. The Post noted Trump’s name did not appear on the list of 67 donors who contributed $10,000 or more to the project.
The funds for the project are administered through a complex, multi-step pipeline involving a private nonprofit, federal agencies and direct oversight by the Trump administration.
The news agency reported Nvidia became “a top contributor to the ballroom in October 2025, nearly three months after it struck a groundbreaking pact with Trump to sell advanced microchips to China despite national security concerns”.
“As an American company, our support for the United States is paramount, and independent of any business interests,” a representative for Nvidia told Mobile World Live. “We were proud to be able to make a donation to a nonpartisan trust that serves all Americans and have said so publicly since the donation was made.”
Apple and EchoStar did not reply to requests for comment on their contributions.
Nvidia CEO Jensen Huang (pictured) reportedly outlined plans to invest $2 billion across Europe to ramp adoption of AI technology and the company’s products.
Bloomberg reported Huang revealed the plan yesterday (13 October) during an event in Paris, which is part of a tour including stops in London and Berlin.
Huang said the fresh investments, to be announced in the next week, are on top of $8 billion the company previously committed to the continent, according to the news agency.
He also announced new funding for French start-ups Doctolib, Genesis AI, Pigment and Entalpic.
Nvidia’s founder also promoted open-source models, arguing their lower costs and greater flexibility could help Europe close the gap with the US on advanced AI.
The chipmaker is a major partner of French company Mistral AI, which develops open-weight AI models. It has invested in Mistral and supplies accelerators for its data centres.
“Open models are important as they can be customised for languages and for your own sovereignty,” he said, according to Bloomberg.
Nvidia’s GTC Berlin conference is schedule for 20-22 October.
Nvidia is reportedly in talks to either acquire or increase its investment in open-source start-up Reflection AI, as it seeks to strengthen its position in open-weight AI models and challenge lower cost Chinese rivals.
Sources told the Financial Times (FT) discussions remain at an early stage and could lead to a full acquisition, further equity investment or a deal covering chips and computing capacity. An acqui-hire, under which Nvidia would recruit staff and license technology, is another option.
A deal could apparently be reached in the coming weeks, though the terms under discussion remain unclear.
According to the report, Nvidia has already invested $800 million in Reflection AI, which was valued at $25 billion in a March funding round. Founded in 2024 by former Google DeepMind researchers, the start-up develops AI models and tools for software development.
Reflection launched its first open-weight model, Beam, last week, targeting coding and agentic tasks as it looks to compete with Chinese models such as DeepSeek and Kimi.
Open-weight models allow businesses and developers to customise AI systems and run them on their own infrastructure, potentially cutting costs compared with closed services such as OpenAI’s ChatGPT and Anthropic’s Claude.
Although Nvidia has its own Nemotron family of open-weight models, FT said it lacks a frontier-scale model of its own. In an essay published in July, CEO Jensen Huang called for a domestic open-weight “ecosystem” across the US, arguing it was essential for “creating opportunities for innovation and prosperity across the country”. “In fact, openness may be one of the most important paths to AI safety and security,” he argued, adding “relying solely on closed models is not inherently safe” as “they can be breached, misused, or fail in ways that outsiders cannot detect”.
Reflection AI has also secured substantial computing capacity from SpaceX. The deal, struck in June, gives the start-up access to Nvidia’s GB300 chips at SpaceX’s Colossus 2 facility in Memphis, Tennessee, for $150 million per month through 2029.
Nvidia highlighted work to help operators rethink dependence on closed AI systems, promoting open-source models to give carriers direct control over network automation and customer workflows.
In a blog, Nvidia stated operators including SoftBank Corp, AT&T and Indosat Ooredoo Hutchison are increasingly deploying open models alongside proprietary systems, explaining they retain full visibility over sensitive network data and can fine-tune models for their specific operational needs.
The move reflects a broader shift across telecommunications.
Nvidia highlighted the findings of its State of AI in Telecommunications 2026 survey in which 89% of operators deemed open-source software critical to their AI strategy, driven in part by independent benchmarks showing open models closing performance gaps in reasoning, coding and autonomous network triage.
Operators are dividing their workloads rather than relying on a single system. Expensive, proprietary models are reserved for complex tasks where closed platforms still hold an edge, while lower-cost open foundation models handle day-to-day operations.
SoftBank used Nvidia’s Nemotron family as the foundation for its Large Telecom Model, employing years of internal operational data.
Rajeev Koodli, principal fellow at SoftBank and SVP of SB Telecom America, explained open models allow it to capitalise on rapid global AI progress and its accumulated network expertise.
AT&T is also taking a multi-model approach. Chief data and AI officer Andy Markus said the goal is to intelligently match “every workload to the right combination of performance, cost and control.”
Nvidia highlighted Indosat Ooredoo Hutchison tailored its open models to Indonesian language and culture.
Paul Krugman is the greatest econ writer in the world, and also a legendary economist. But I sometimes feel that he has a blind spot when it comes to the value of new technologies. In 1998 he famously wrote: The growth of the Internet will slow drastically, as the flaw in ‘Metcalfe’s law’—which states that […]
Mobile World Live brings you our top three picks of the week as European Union (EU) governments pushed for changes to high-risk supplier rules, Nvidia’s chief weighed in on the AI model distillation debate and Poste Italiane boosted its stake in Telecom Italia.
Governments press EU to change high-risk supplier plan
What happened: European governments pushed for amendments to a European Commission (EC) proposal that would require operators to remove network equipment supplied by vendors deemed high risk, according to a draft document seen by Reuters.
Why it matters: Early this year, the EC published a Cybersecurity Act proposal outlining a 36-month phase out for high-risk third-country supplier equipment in mobile networks. However, EU member states have reportedly called for the phase out period to be determined by several factors rather than a fixed deadline, including level of risk, equipment lifespans, compatibility requirements and the availability of alternatives. The proposed rules are expected to affect Chinese vendors Huawei and ZTE.
Mason Hurlocker of Bloomsbury Intelligence and Security Institute (BISI) stated “the reforms carry a range of political, operational, security and economic risks”, adding it is possible the EC narrows its provisions “given industry cost warnings and resistance from some member states to mandatory rip-and-replace obligations”. AGSMA Intelligence report estimated mobile network replacement would cost the industry between €30 billion and €40 billion.
Nvidia chief says AI model distillation is fair competition
What happened: In an interview with CNBC, Nvidia CEO Jensen Huang said AI model distillation should be treated as competition rather than theft, putting him at odds with US officials over the practice.
Why it matters: Huang said developers are entitled to test competing products, comparing distillation with competitors taking Nvidia chips apart to understand how they work. “That’s called competition,” he apparently said. He further argued the burden is on providers to control access, stating: “if you don’t like people to use your products… disable the service.”
The comments come after US officials accused Chinese AI players of conducting industrial-scale distillation campaigns against US models. Anthropic separately alleged that Chinese companies including Alibaba and DeepSeek used large numbers of accounts and proxy networks to extract capabilities from Claude.
Otis Yi, programme director at digital entertainment platform GashPoint, said “the dispute centres on how those outputs are obtained and whether the model owner has authorised their use”, adding industrial-scale extraction can raise “separate questions about contracts, intellectual property and enforcement”. Mark Seery, founder of consulting and technology advisory firm Bohcay, argued: “Competition matters. So does authorisation.” He noted “the customer does not own the underlying infrastructure”, adding when they exceed agreed rights, “that is not simply “testing a product”; it may be a contractual violation with legal consequences”.
Poste Italiane secures 85% stake in Telecom Italia
What happened: Poste Italiane secured an 85.8% stake in Telecom Italia (TIM) after completing the second and final phase of its public tender offer.
Why it matters: The postal group acquired an additional 65.7% of TIM, building on the 20% stake it acquired last year when it replaced Vivendi as the operator’s largest shareholder. The tender offer, launched in July, valued the transaction at around €13 billion and offered shareholders €1.67 in cash per share.
In March, the company had outlined plans to take TIM private. However, Reuters reported the result leaves Poste well short of the 90% threshold needed to take it private, leaving it to consider other options.
The operator’s board has since acknowledged Poste’s exercise of “direction and coordination” over TIM and its subsidiaries, and approved new rules covering operational governance and information flows between the companies. Alessandro Marchesini, a TIM board member, described the transaction as the closing of an “important financial chapter”, adding, “now the industrial one begins”. He said the focus would shift to “turning the opportunities created by this transaction into projects, collaboration and long-term value”.
Login
