Global Chatter: Micron, SpaceX, Palantir & More
Edition #2
Welcome to the second edition of Global Chatter — a fortnightly newsletter where we dig through what the world’s biggest companies are saying and bring you the most interesting insights, whether about their businesses, industries, global trends, or the broader economy. We read through major global earnings calls, shareholder letters, investor presentations, and management interviews so you don’t have to.
From Big Tech and semiconductors to global banks, consumer giants, energy majors, and industrial leaders, Global Chatter is our attempt to cut through the noise and surface the conversations that actually matter.
We’re always looking to make Global Chatter more useful and insightful—so if you have ideas on how we can improve or innovate the format further, we’d love to hear them.
In this edition, we have covered 6 companies across 5 diversified industries.
Semiconductors & AI Technology
Micron Technology
NVIDIA
Advanced Data Integration & Analytics
Palantir
IT Services & Consulting
IBM
Food Services & Facilities Management
Sodexo SA
Aerospace & Space Technology
SpaceX Corporation
Micron Technology | Semiconductors
Micron Technology is a leading American multinational semiconductor company that designs and manufactures computer memory and data storage products, including DRAM, NAND, and High Bandwidth Memory (HBM).
[Concall]
On demand visibility for High Bandwidth Memory (HBM), management said customer demand remains well ahead of available capacity, even several years out.
“The demand that we have for our HBM products—HBM3E, HBM4 and beyond—is far in excess of our ability to supply. Through these Strategic Customer Agreements, we’re discussing demand over a multi-year horizon, even beyond 2027 and into 2028. We are able to get very high-confidence demand from customers that is well above what we can support. Even when we sign multi-year agreements, the committed volumes are still less than what customers actually want. Our demand for HBM, not just in 2027 but even in 2028, is well above our ability to supply.”
— Sumit Sadana, Chief Business Officer
Mark Murphy updated Micron’s industry outlook following stronger-than-expected AI demand.
“Today we indicated that we expect market tightness to continue beyond 2027. Part of the reason is that we have increased our estimate for the HBM total addressable market. Previously, we expected it to cross $100 billion in 2028. We now see the HBM TAM easily crossing $100 billion in 2027.”
— Mark Murphy, Chief Financial Officer
Asked about long-term DRAM and NAND bit demand growth, management explained why it is no longer providing multi-year CAGR forecasts.
“For the foreseeable future, shipment growth for bits is not really determined by demand anymore. It’s actually determined by supply. Demand is so much above the industry’s ability to supply that supply growth will determine shipment growth far more than demand growth. Our expectation is that supply growth will continue to remain short of what is needed to meet demand, and we don’t really see when supply will be able to catch up.”
— Sumit Sadana, Chief Business Officer
Management explained the structure of the new Strategic Customer Agreements (SCAs), highlighting their long-term nature and financial protection.
“These Strategic Customer Agreements cannot be cancelled. There is no provision that allows a customer to walk away. Outside of automotive, these are generally five-year take-or-pay agreements. Customers commit to annual volumes and are obligated to pay whether or not they take delivery. Prices are negotiated within agreed floor and ceiling bands, and there are premium provisions for higher-performance products. The one provision that does not exist is any customer’s ability to walk away from these agreements.”
— Sumit Sadana, Chief Business Officer
Management described the financial commitments supporting the new contract model.
“The agreements we have already signed aggregate to more than $22 billion in total cash and related financial commitments, of which the cash alone is almost $18 billion. As we expand these agreements toward our target of covering roughly half of the company’s revenue, the cash associated with them will increase significantly.”
— Sumit Sadana, Chief Business Officer
On memory architecture trends, management expects LPDRAM to gain share as hyperscalers focus on power efficiency.
“We continue to expect LPDRAM usage to grow over time as a percentage of DRAM consumption in data centers. Micron has been a pioneer in LPDRAM for the data center, first with the technology and first with SO-DIMM products. We have strong customer engagement and expect to remain the recognized leader in this space because LPDRAM helps reduce power consumption, improve performance and reduce memory footprint.”
— Sumit Sadana, Chief Business Officer
Management discussed the cost implications of new fabs and the industry’s transition toward higher-performance memory.
“Higher-performance solutions such as HBM require more silicon per bit, and greenfield fabs do not benefit from the same leverage as traditional technology transitions. Both the move to higher-performance products and greenfield investments will increase DRAM bit costs in the near term.”
— Manish Bhatia, EVP, Global Operations
On NAND strategy, management highlighted strong momentum in enterprise storage.
“Our enterprise SSD momentum is exceptionally strong. We delivered a $5 billion quarter in enterprise SSDs within our $25 billion data center business. We continue to gain record market share because of the strength of our NAND portfolio.”
— Sumit Sadana, Chief Business Officer
Summarizing the company’s long-term outlook, management said AI and structural industry changes have reshaped the memory market.
“The combination of demand, the structural supply challenges in the industry, AI creating newfound relevance and strategic importance for memory, and now these Strategic Customer Agreements are completely transformative for our business.”
— Sumit Sadana, Chief Business Officer
Asked whether the company would increase DRAM exposure given tight market conditions, Sumit Sadana reiterated Micron’s commitment to maintaining a diversified memory portfolio across end markets.
“Our mix in the business, DRAM versus NAND, tends to oscillate between roughly 80% DRAM and 20% NAND, or about 75% DRAM and 25% NAND. We are pretty comfortable with that mix. As it relates to HBM, we have made a strategic decision that our goal is to have our HBM share consistent over time with our DRAM share. We intend to support customers on HBM as well as the non-HBM portion of the DRAM business across all market segments. We definitely believe in the strength of diversity.”
— Sumit Sadana, Chief Business Officer
Management emphasized that despite AI momentum, Micron continues to value diversification across end markets.
“If you look at the AEBU business and the MCBU business, both of which are non-data-center businesses in our business unit structure, that’s almost 40% of our company revenue. We like that diversity and will continue to focus on supporting customers across all market segments, including non-HBM DRAM, HBM and NAND.”
— Sumit Sadana, Chief Business Officer
On NAND competitiveness, management highlighted the strength of its enterprise storage portfolio.
“We have hit record share after record share in data-center SSDs because of the strength of that portfolio. We are the QLC leader in the world, the first company to introduce Gen 6 drives in volume, and we are also leading with the highest-capacity 245-terabyte SSDs.”
— Sumit Sadana, Chief Business Officer
Responding to a question on CXMT and YMTC, management downplayed the competitive threat outside China.
“Those companies have certainly grown over the years in terms of their capabilities and market share. However, the overwhelming majority of their output is sold within China. We have not really seen meaningful competition from them outside China.”
— Sumit Sadana, Chief Business Officer
Management explained how it plans to defend its market position despite increasing competition.
“Our focus is on the highest-performing and most complex products in the portfolio. We look for products that are difficult to get right, engage deeply with customers over multiple years, consistently meet or beat time-to-market targets, and maintain leadership in innovation. We also have one of the strongest intellectual property portfolios in the world, with almost 65,000 patents, and we have a long track record of defending that IP.”
— Sumit Sadana, Chief Business Officer
While both memory categories remain supply constrained, management highlighted that DRAM shortages are more acute.
“Customers are definitely interested in securing NAND supply, but DRAM is far more constrained and much more difficult to supply in the quantities they need. NAND is very constrained too, but the sense of concern and urgency around DRAM is significantly higher.”
— Sumit Sadana, Chief Business Officer
NVIDIA Corporation | Semiconductors & AI Infrastructure
NVIDIA Corporation is a technology company that designs GPUs, AI computing platforms, networking solutions, and software for applications across data centers, cloud computing, gaming, robotics, autonomous vehicles, and high-performance computing. In this shareholder address, CEO Jensen Huang discusses the evolution of AI from a research tool to commercial infrastructure, the growing adoption of Agentic AI, the role of Blackwell and Vera Rubin in the company’s roadmap, and why NVIDIA believes AI infrastructure investment is still in its early stages.
Management explains that computing has shifted from simple execution to reasoning, which fundamentally changes how hardware is utilized. This transition signals a massive expansion in the utility and market size of computing infrastructure.
“For 60 years, humans wrote software and computers executed instructions. That paradigm has changed. With AI, computers can understand, reason, plan, use tools, and do useful work. AI becomes useful computer is no longer just a tool.”
— Jensen Huang, President and CEO, NVIDIA
Management is reframing their products as production assets rather than traditional IT expenses. This shift in customer mindset supports higher pricing power as buyers focus on revenue generation potential rather than just hardware cost.
“Customers are not buying computers. They’re building revenue generating AI factories. The architecture of the factory matters. The question is how much revenue the factory can produce and at what cost.”
— Jensen Huang, President and CEO, NVIDIA
NVIDIA is entering the CPU market with a product designed specifically for AI agents rather than human users. This opens a significant new revenue stream and challenges traditional CPU manufacturers in the data center space.
“Agents don’t rent cores but demand ultra fast responses. There will be billions of agents and they will need a CPU built for them. We believe Vera will be one of the most significant product launches in our company’s history and the orders are already coming in.”
— Jensen Huang, President and CEO, NVIDIA
Management identifies robotics and autonomous machines as the next major growth driver following the current digital AI boom. This long-term roadmap provides visibility into growth opportunities beyond the current language model trend.
“Physical AI is the next wave of growth for Nvidia. Physical AI is agentic AI in the real world. Robots, cars, and factories will perceive, reason, plan, and operate in dynamic environments. Nvidia pioneers this field and builds the full loop.”
— Jensen Huang, President and CEO, NVIDIA
In response to a question about sustainability, management asserted that the shift to AI is a multi-decade structural change rather than a short-term bubble. This long-term perspective is meant to reassure investors about the durability of current growth rates.
“For the first time in 60 years, computing is being reinvented from largely retrieving, storing, and sending information to generating intelligence with AI. ... This buildout will be measured in decades.”
— Jensen Huang, President and CEO, NVIDIA
Management highlighted that demand is broadening from a few large cloud providers to entire nations and specific industries. This expansion suggests the total addressable market is significantly larger than just the existing tech giants.
“As organizations seek to manufacture intelligence at scale, demand for AI factories will extend far beyond today’s clouds to enterprises, sovereign nations, and regional AI clouds. NVIDIA’s uniquely enable and uniquely enables AI factory buildouts.”
— Jensen Huang, President and CEO, NVIDIA
Addressing the shift from training to inference, the CEO shared data proving their hardware remains the most cost-effective for running AI models. Maintaining leadership in inference is critical as it becomes the dominant portion of the AI market.
“In semi analysis inference X benchmarks, Blackwell was declared the inference king, delivering the best performance per watt, the lowest cost token and 30 times higher token throughput. ... NVIDIA AI infrastructure delivers the best performance and therefore the best inference economics.”
— Jensen Huang, President and CEO, NVIDIA
Management confirmed that most of their hardware in the field is already doing inference work, contrary to some market fears. This validates the company’s competitive position against specialized inference chip competitors.
“Today, the vast majority of NVIDIA’s compute footprint is used for inference, and we are favorably positioned to expand our share, as we demonstrate with our recent announcements with Anthropic and and Apple.”
— Jensen Huang, President and CEO, NVIDIA
The CEO discussed how AI investments are driving broader economic activity and infrastructure upgrades in the US. This deep integration into national industrial strategy provides a layer of political and structural support for the business.
“AI is a once- in a generation opportunity to reindustrialize America with chips and [clears throat] systems manufacturing, bring back advanced manufacturing jobs, upgrade our aging power grid, and invest in sustainable energy. Because of AI, we have the profits and scale to invest in partners.”
— Jensen Huang, President and CEO, NVIDIA
The company reaffirmed its commitment to returning half of its free cash flow to investors. This consistent policy provides a predictable floor for shareholder returns as the company continues to scale.
“Supported by our conviction in sustainable market growth and free [clears throat] cash flow generation, we plan to return 50% or more of our free cash flow this year, next year, and beyond, growing both our share repurchases and dividend overtime.”
— Jensen Huang, President and CEO, NVIDIA
Palantir | Advanced Data Integration & Analytics
Palantir Technologies is a specialized software company that develops advanced data integration, analytics, and AI platforms for governments and large commercial enterprises. Its core systems unify siloed datasets into a single operational “digital twin,” enabling organizations to visualize information, make AI-driven decisions securely, and deploy operational software.
Explaining why Palantir partnered with Nvidia, Karp said customers increasingly want ownership and control over every layer of their AI stack.
“If you want to know how this came together from my perspective, there were a lot of technical issues—who controls the models, who controls the weights, who controls the value of your business. We’re sitting on critical infrastructure across America, Ukraine and Israel. Everyone who uses LLMs on the battlefield runs on top of our ontology. Our clients are unhappy with the Frontier Labs. There is a level of discomfort and loss of trust. What aligns me with Nvidia, and I think what technical customers want, is control over their compute, their models, their data stack and their alpha. They want to know they own the means of production. It’s not being transferred to someone else.”
— Alex Karp, CEO, Palantir
Karp explained why Palantir’s ontology layer is critical for deploying AI in regulated environments.
“When you’re using large language models, everyone technical realizes they’re a critical resource. To make them valuable in an enterprise, battlefield, regulated or manufacturing context, you have to have what’s called an application layer. We have this thing called ontology that now everyone’s copying. It takes a large language model and makes it safe, useful and precise. Safe because it doesn’t touch your underlying data. Safe because it prevents the large language model from caching your data and replicating your business. Safe because it doesn’t transfer your IP or your classified information.”
— Alex Karp, CEO, Palantir
Karp argued that enterprise customers increasingly fear losing proprietary information when using foundation models directly.
“Something has gone completely wrong. The basic view among enterprises in this country is, ‘I’m going to waste my time with tokens, I’m going to get no value, and they’re going to get my IP.’”
— Alex Karp, CEO, Palantir
Responding to criticism that he was targeting OpenAI and Anthropic, Karp emphasized customer trust.
“This is reporting. I’ve literally called these people against my own interest because I’m profiting from this. The reality is you may not like us at my former schools, Harvard or Berkeley, but enterprises in this country trust and love us, especially the ones involved in critical infrastructure, both public and private.”
— Alex Karp, CEO, Palantir
Karp explained why Palantir’s software has become the standard for mission-critical AI deployments.
“The whole secret here is the forward-deployed model, the products that have been five years ahead. Everyone said forward deployed engineers were services. They said they didn’t even know what an ontology was. That’s the only thing people talk about now. The secret was we delivered the best things for the warfighters.”
— Alex Karp, CEO, Palantir
Karp described the shift in enterprise buying preferences.
“Our products are completely agnostic. We now sell a product that allows customers to switch from model to model. But we need to rebuild trust. That trust is going to happen where everyone gets to ask and answer basic questions: Who owns the data? Where is it cached? Are the prompts secure? Is this being transferred to you?”
— Alex Karp, CEO, Palantir
Karp explained why foundation models alone are insufficient.
“What I am claiming is it’s the model plus an application layer plus compute. It is really all three. The reason why everyone is struggling with bad financials and growth while losing money is that clients refuse to pay the true cost. The two places that actually make money—profit and free cash flow—are our application layer called ontology and compute.”
— Alex Karp, CEO, Palantir
Karp argued that enterprises don’t need proprietary frontier models if they retain ownership.
“We can take an open model and, in either a classified or non-classified context, get it to the point of a frontier model while you control the weights. We can get the frontier application to be exactly the same as a frontier model without the risk of transferring the alpha of your business to somebody else.”
— Alex Karp, CEO, Palantir
Karp summarized what he hears from enterprise customers.
“Every single enterprise I deal with is livid. They’re saying, ‘I’m paying for tokens that create no value. These people are stealing the weights and the alpha of my business.’”
— Alex Karp, CEO, Palantir
Explaining why frustration has grown among enterprise customers.
“The voice of American business is being channeled through me. It is absolutely a problem for this country because we’re on the cutting edge of every AI technology, but people have triply oversold something. The enterprises are tired of it.”
— Alex Karp, CEO, Palantir
Karp challenged investors to verify his claims directly.
“I want everybody watching this, especially investors who think somehow this is working, to pick up the phone and call a CEO in private—not in public—and ask whether they’re livid. I’m telling you, they’re twice as livid as me.”
— Alex Karp, CEO, Palantir
Karp distinguished between real AI progress and excessive marketing.
“The reality of compute plus ontology plus models is changing the course of history. Ask the Ukrainians. Ask the Israelis. Ask our Department of Defense. Ask the enterprises where this is working. We do not have to oversell what we have. It’s all being built in this country. We do not have to overhype it.”
— Alex Karp, CEO, Palantir
Karp explained why Palantir remains model-agnostic.
“We’re completely agnostic. We prefer a world where there are more hyperscalers and more choices. What is happening among the most technical players is that they’re saying, ‘I want something I own. This is my business. I want to own the GPUs. I want to own my data. I want to own the model. I want to control the alpha.’”
— Alex Karp, CEO, Palantir
Karp argued that many enterprises believe foundation model providers are extracting value without delivering commensurate business outcomes.
“Every enterprise I deal with is livid. They’re saying, ‘I’m paying for tokens that create no value. These people are stealing the weights and the alpha of my business.’ These models have been irresponsibly oversold. The sales pitch is that they’re dangerous for everyone, yet somehow they can be given to all of our adversaries, but they can’t safely be deployed for the Department of Defense or an enterprise without risking the transfer of that enterprise’s alpha. That’s not a sustainable model.”
— Alex Karp, CEO, Palantir
Asked whether dissatisfaction with Frontier Labs is benefiting Palantir commercially.
“I’m not here to talk about what happens to other businesses. What I can tell you is that in our business we have more demand than we can supply. We have more business than we can execute. If you look at our financials, you can see exactly where we’re headed.”
— Alex Karp, CEO, Palantir
Discussing AI regulation and cybersecurity, Karp argued against restricting advanced AI for Western governments while allowing adversaries access.
“One of the things we’re really trying to get the West to do is move away from slogans and look at the technical realities. It makes no sense to restrict powerful models from your own government because you disagree with how the government fights wars, while effectively opening those same capabilities to the rest of the world, including your adversaries.”
— Alex Karp, CEO, Palantir
Asked about China’s AI progress, Karp outlined how he views the global technology landscape.
“There are really two-and-a-half relevant technology centers in the world today: America, China and Israel. I spent half my life in Europe and I want Europe to be relevant, but China is a true peer competitor. China has many advantages, and it would be a mistake to underestimate them.”
— Alex Karp, CEO, Palantir
Karp warned against complacency in the United States.
“China doesn’t have some of the creativity, ingenuity and deep-tech culture that we have. But it is not a foregone conclusion that we win. The biggest problem in this country is that we debate these issues as though we don’t have adversaries. That’s simply not true. This is binary. Either they win or we win.”
— Alex Karp, CEO, Palantir
Karp argued that long-term public support for AI depends on broader economic benefits.
“We have to find ways to make these models raise the standard of living for every American. People need to feel that it isn’t just those sitting around this table getting richer. Those are real issues, and if we ignore them, we create unnecessary political backlash.”
— Alex Karp, CEO, Palantir
Summarizing his argument throughout the interview, Karp said future winners will be those who retain ownership of their AI infrastructure.
“The most technical customers are all asking the same questions: Who owns the GPUs? Who owns the model? Who owns the data? Who owns the alpha? Whoever owns the means of production will own the future.”
— Alex Karp, CEO, Palantir
IBM | IT Services & Consulting
IBM is a global technology leader at the forefront of hybrid cloud, artificial intelligence, and quantum computing. In this conversation, Chairman & CEO Arvind Krishna shares why IBM believes quantum computing is approaching a commercial inflexion point, the strategic importance of world-class scientific talent, and how the company is building the foundation for the next era of computing.
Management is setting realistic expectations regarding the maturity of quantum technology while emphasising that it is no longer a distant concept. This suggests that while near-term commercialisation is still evolving, the structural shift toward quantum computing is much closer than many expect.
“We are still in the early days of quantum. Let me acknowledge that it’s not far out into the future, but it is still early.”
— Arvind Krishna, Chairman & CEO, IBM
The CEO has provided a specific two-to-three-year window for quantum computing to deliver massive commercial advantages. This accelerated timeline points to a faster path toward real-world enterprise adoption.
“Two to three years. Ambassador Kwatra mentioned it. We have one quantum computer that is going into Amaravati through an agreement with the state government and the national government.”
— Arvind Krishna, Chairman & CEO, IBM
IBM is expanding its quantum infrastructure through strategic government partnerships while continuing to evaluate additional deployments. These investments are aimed at building the ecosystem required for long-term quantum computing adoption.
“We have one quantum computer that is going into Amaravati through an agreement with the state government and the national government. We are also discussing another one. I’ll leave that location unstated until everything is finalized.”
— Arvind Krishna, Chairman & CEO, IBM
The decision to deploy quantum hardware is driven by the scarcity of specialized mathematical talent globally. IBM believes access to world-class scientific expertise will be a defining competitive advantage in quantum computing.
“The reason we’re putting it there is because of the talent. Where are you going to find enough people with deep knowledge of mathematics and physics who want to work on translating problems into the unique mathematics that quantum computers can solve?”
— Arvind Krishna, Chairman & CEO, IBM
IBM is prioritizing specialized, high-impact problem solving over workforce scale. This focus on quality and depth of expertise reflects the company’s long-term strategy for quantum computing.
“It is not about volume; it is about quality, depth, and the ability to think through extremely difficult problems and deploy them on these systems in unique ways over the next decade.”
— Arvind Krishna, Chairman & CEO, IBM
Management believes quantum computing has the potential to solve classes of mathematical problems that have challenged researchers for decades, unlocking entirely new possibilities.
“Let me give you an example. There are problems involving prime number factorization. For nearly a hundred years, people believed these problems were impossible. Three people from one of the IITs solved one such problem about 15 years ago. We can tap into talent like that to leverage these computers. I think it unlocks all kinds of possibilities.”
— Arvind Krishna, Chairman & CEO, IBM
Sodexo SA | Food Services & Facilities Management
Sodexo SA is a French multinational corporation and global leader in food services and facilities management. Headquartered near Paris, it employs over 426,000 people, serves 80 million consumers daily across 43 countries, and acts as a major driver of corporate operations worldwide.
[Concall]
Responding to a question on sales momentum, management highlighted improving conversion rates and a strengthening pipeline.
“At this stage, we are seeing early encouraging signs on net commercial growth. We have a good pipeline and an improved conversion rate. Overall, when we look at our last 12 months forward-looking net new KPI, it has improved compared with the end of Q2. This improvement is expected to come through progressively in the reported in-year net new revenue. The improvement is really coming from development. Q2 was better than Q1, and Q3 is better than Q2.”
— Sébastien de Tramasure, Chief Financial Officer
Investors questioned why the upgraded FY26 guidance still implied a muted Q4. Management explained the rationale.
“Q4 will face tougher comparables. Last year we had a very strong Q4, especially in Energy & Resources in North America, creating a less favorable year-on-year comparison. We also have the normal phasing of net new contract mobilizations and annualization. Overall, we see some uncertainty in the macroeconomic and geopolitical environment. At this stage, we believe it is appropriate to remain prudent for the remaining part of the year. However, we are not seeing any specific deterioration in the business and, based on what we see today, we would currently expect Q4 to be modestly positive.”
— Sébastien de Tramasure, Chief Financial Officer
Management attributed the strong performance to higher attendance and increased customer spending across venues.
“Overall, we had a very good performance across all the activities of Sodexo Live!—convention centers, airline lounges and sports venues. We had very good sporting events such as the BNP Paribas Open at Indian Wells, the Miami Open and a very strong beginning of the baseball season with the Seattle Mariners. Attendance was very good across all events. On top of attendance, we were able to capture more revenue through higher spend per capita, driven by our offerings, innovative concepts and brand partnerships. Overall, Sodexo Live! delivered more than 15% organic growth in Q3, definitely above our expectations.”
— Sébastien de Tramasure, Chief Financial Officer
On pricing and cost inflation, management said there has been no meaningful change in the inflation environment.
“We are not seeing any meaningful change in food inflation overall. There is some pressure on energy prices and transportation and logistics costs, but overall this remains under control. Managing inflation is part of what we do. We continue to manage it through product substitution, menu engineering, client discussions and supplier negotiations. Overall, we are monitoring input inflation very closely.”
— Sébastien de Tramasure, Chief Financial Officer
Responding to a question on next year’s pricing outlook, management expects only modest changes.
“We have seen a declining trend in labour inflation. You need to keep in mind that pricing is driven by both food inflation and labour inflation. At this stage, it is fair to expect something quite similar for the beginning of next year, with perhaps some small pressure on food inflation but a declining trend in labour inflation. We do not expect any significant change in the pricing path.”
— Sébastien de Tramasure, Chief Financial Officer
Management shared its expectations for customer retention and commented on the latest education selling season.
“Based on what we see today, we should land broadly in line with last year, around the 94% retention level. Specifically in U.S. Education, the selling season should be slightly better than last year, but still disappointing. This remains a key focus area. We have new leadership in Education and K-12, and they are building clear action plans around retention, development and preparing for the FY27 selling season.”
— Sébastien de Tramasure, Chief Financial Officer
Responding to a question on AI infrastructure, management highlighted the opportunity in data center construction.
“The data center market is clearly a fast-growing opportunity for us. Today we already provide food services at operational data centers, including 24/7 offerings, convenience services and pantry solutions. More importantly, the construction phase is where we believe we can capture the greatest value. We are actively building our pipeline, and we believe we are well positioned because the capabilities we have developed in Energy & Resources—operating large camps and remote sites with thousands of workers—are directly applicable to large data center construction projects.”
— Sébastien de Tramasure, Chief Financial Officer
Management highlighted continued resilience in the Healthcare & Seniors segment despite a more cautious outlook elsewhere.
“Healthcare & Seniors continues to trend very well. Organic growth was 7.8% in Q3, broadly in line with Q2. The underlying trend remains strong, although we do expect some annualization effects in Q4.”
— Sébastien de Tramasure, Chief Financial Officer
On a follow-up question, management confirmed there had been no change to the assumptions supporting margin guidance.
“If you look at the bridge we presented at the half-year results, it is exactly the same. There has been no significant change in the different levers supporting the margin outlook.”
— Sébastien de Tramasure, Chief Financial Officer
Management explained that the improvement in volumes was driven not only by Sodexo Live! but also by project activity in Energy & Resources.
“Volume improved significantly when you compare the year-to-date performance with the first half. The two main reasons were the very strong performance of Sodexo Live!, which was driven by volume, and the additional projects in the Rest of the World, especially in Energy & Resources, which also contributed positively to volumes.”
— Sébastien de Tramasure, Chief Financial Officer
Management reiterated that near-term caution will not come at the expense of long-term investments.
“Our focus remains on strengthening the business over the medium term. We continue to invest in commercial capabilities, competitiveness, supply and technology while maintaining a close watch over the external environment.”
— Sébastien de Tramasure, Chief Financial Officer
Management explained why Sodexo believes it has an edge in winning data centre contracts.
“When you look at what we do in Energy & Resources, we are talking about large camps and remote sites with thousands of workers. It is basically the same offer that we need to address these opportunities in data centre construction.”
— Sébastien de Tramasure, Chief Financial Officer
Responding to a question on foreign exchange, management said the company remains comfortable with its existing assumptions.
“We are comfortable maintaining our overall guidance of around a 3% negative currency impact. With the recent evolution of the euro-dollar, we expect a more favourable impact in Q4 compared with Q3, and we should land around this 3% impact for the full year.”
— Sébastien de Tramasure, Chief Financial Officer
Management highlighted a favourable trend in one of its largest cost components.
“We have seen a declining trend in labour inflation. Pricing reflects both food inflation and labour inflation, and at this stage we do not expect any significant change in the pricing path.”
— Sébastien de Tramasure, Chief Financial Officer
SpaceX Corporation | Aerospace & Space Technology
SpaceX develops reusable launch vehicles, satellite communications infrastructure, and space transportation systems. In this interview, President and COO Gwynne Shotwell discusses why the company chose to go public, how it plans to scale Starship and Starlink, its long-term AI infrastructure strategy, the role of vertical integration, and how management is balancing aggressive innovation with investor expectations in the next phase of the company’s growth.
The company views mission failures during development as valuable data points rather than purely negative events. Investors should expect occasional public setbacks as a necessary part of the company’s rapid learning and innovation process.
“I think it’s actually really important to have failure. If you don’t have failure, like if a launch goes perfectly, all you’ve learned is that that launch vehicle on that day worked. You didn’t get any more data than that, right? So when you have failure, you actually get this treasure trove of data.”
— Gwynne Shotwell, President & COO, SpaceX
Starlink’s growth is currently limited by how many satellites are in orbit rather than a lack of customers. This high demand suggests that revenue will grow automatically as soon as more capacity is launched.
“We are actually constrained by capacity in many markets, really important markets. So scaling is actually really quite important to the revenue story for us. Meaning there’s more demand than you can actually supply right now. There’s more demand than we can fulfill right now.”
— Gwynne Shotwell, President & COO, SpaceX
Management believes the mobile phone market is significantly larger than the home internet market for Starlink. Success in this area could dramatically expand the company’s total addressable market globally.
“I think Starlink Mobile will far exceed Starlink broadband in the home. I think more than half the population, the global population has a cell phone, which is a shocking, but it’s somewhere, it’s crazy numbers of people have cell phones. Not everybody is going to need broadband, a Starlink broadband in their home.”
— Gwynne Shotwell, President & COO, SpaceX
SpaceX uses its own satellite launches to refine and prove the reliability of its rockets. This internal demand allows the company to improve its launch technology much faster than competitors who rely only on external customers.
“The more you can fly a vehicle, the more reliable it will be, the safer it will be. Starlink provided that market for Falcon 9. We flew, you know, a hundred Starlink missions in a year, right? The AI satellites will be that same market but for Starship.”
— Gwynne Shotwell, President & COO, SpaceX
The company is deepening its vertical integration by producing its own fuel and solar components. This strategy reduces dependency on third-party suppliers but requires significant upfront capital investment.
“We build launch vehicles, we build our launch sites, we write our own software, we’re producing some of our own propellants for the vehicles right now, too. This seems like a very obvious thing, next step for us. Um, we’ll have a lot to learn. There’s a huge amount of equipment that we have to invest in.”
— Gwynne Shotwell, President & COO, SpaceX
SpaceX is betting that space-based data centers will be more efficient than Earth-based ones due to constant solar energy and natural cooling. If successful, this would provide a structural cost advantage for their AI computing services.
“But the most efficient place to put inference compute is on orbit. It’s always sunny in space. You get sixx the amount of power out of a solar cell in space as you do here on Earth. And cooling is free because space is actually quite cold. So radiative cooling is free.”
— Gwynne Shotwell, President & COO, SpaceX
The company plans to sell excess AI computing power to third parties while keeping enough for its own needs. This creates a flexible revenue stream that helps offset the high cost of maintaining satellite networks.
“We will never sell compute capacity that we actually need which is why we wanted the ability to have these contracts be shortterm if necessary but we’ll keep building. I believe we will continue to provide that capability to others actually.”
— Gwynne Shotwell, President & COO, SpaceX
After years of internal growth, the company is now open to acquiring other businesses, particularly in the AI sector. This indicates a shift toward using the company’s equity to accelerate technological expansion.
“SpaceX wasn’t an M&A company for decades. Um and so it’s kind of a new exciting uh it’s a new exciting world for us. Um I do think M&A is in the future, especially when you look at uh the AI world.”
— Gwynne Shotwell, President & COO, SpaceX
Recommended Reading
DSP shares some of the best content for investors to explore and dive deep into what's happening across the markets. Each edition brings together thought-provoking analysis, key management commentary, and insights into the themes shaping businesses and the economy. From AI and data centers to IT services, cement, and rural demand, the June edition breaks down complex trends into simple, actionable takeaways. If you're looking to understand the bigger picture beyond the headlines, this is a great place to start.
That’s it for now! Your feedback will really help shape how Global Chatter evolves. Drop it down in the comments below!
Quotes in this newsletter were curated by Meher, Shahid, Srusti & Kashish.
Disclaimer: We’ve used AI tools in filtering and cleaning up these quotes, so there may be some mistakes. Now, if you are thinking why we are using AI, please remember that we are just a small team of 5 people running everything you see on Zerodha Markets 😬 So, all the good stuff is human, and mistakes are AI.
https://earningsunwrapped.substack.com/p/beyond-gpus-why-memory-is-becoming?utm_source=share&utm_medium=android&r=rer2b
Today published a deep dive of Micron Earnings call, fabulous learnings on tectonic shifts in Memory industry.