Nvidia Earnings This Week; Biggest Power Deal in History | Bloomberg Tech 5/18/2026

NVIDIA's Earnings and Market Position

• NVIDIA is expected to release its earnings results this week, with all eyes on the company as it continues to grow and capitalize on the demand for its chips, despite some concerns about rising memory prices and supply chain issues 10s.
• The company's growth is accelerating, with a topline growth rate above 80%, and it is well-positioned to take advantage of the increasing demand for compute capacity, particularly in the era of more agentic AI and enterprise investments 2m6s.
• NVIDIA's CEO, Jensen Huang, will be speaking at an event in Las Vegas, where he is expected to address some of the supply chain questions and provide more information about the company's plans and partnerships, including its recent acquisition and partnership with Anthropic 4m20s.
• The company's earnings call on Wednesday is highly anticipated, with analysts expecting to learn more about the company's margins, the impact of rising memory prices, and its position in the market, particularly with regards to its chips and CPUs 6m30s.

The AI Data Center Boom and Energy Deals

• Elon Musk has announced that he is working on plans for the SpaceX IPO, which is expected to be the largest listing ever, and he made these comments virtually while in Texas 10m10s.
• The AI data center boom has triggered the largest power deal in history, with companies like NVIDIA and Dell playing key roles in the state of play, and the timing of these developments is curious, particularly in light of the current geopolitical tensions and market trends 8m40s.
• SpaceX has notified investors that it is executing a five-for-one stock split, with shareholders being told that the fair market value of each share was adjusted following the split, and the company is also teaming up to build a futures market for computing power 10s.
• The largest power deal in history has been triggered by the AI data center boom, with a value of $67 billion, and it involves acquiring Dominion Energy to create a giant energy company, driven by the need for scale and the increasing demand for power 1m42s.
• The deal is expected to be approved within a year to 18 months, as it only requires regulatory approval from three states, including Virginia's Data Center Alley, and analysts believe it is not an uphill battle, despite other regulatory bodies also regulating power 2m6s.
• The timing of the deal is key, with political pressure and focus on the cost of energy due to the AI buildout, and the companies involved are aware of the need to focus on cost and price to ensure consumers do not think their utility bills will go up 3m10s.

Futures Market for AI Computing Power

• A futures market for AI computing power is being launched, with the goal of making compute an asset class, and it is expected to be incredibly important for the industry, allowing companies to hedge their exposure to compute costs 5m30s.
• The futures market will involve players such as banks, data centers, and companies that use GPU resources, who will be able to hedge their costs and manage their exposure to compute prices, with the market providing transparency and the ability to put assets on their balance sheet 7m20s.
• The idea for the futures market was sparked by the lack of transparency around access to compute, with the founder noting that unless you were part of the ecosystem, you probably didn't know the reserve price that people were paying for long-term contracts 10m40s.
• The discussion revolves around the concept of term structure risk and the lack of visibility in pricing, particularly for non-homogeneous products like GPUs, which require indices to normalize underlying assets and produce trackable data, 10s.
• The cost of computer memory is going up, with prices increasing since December of last year, due to a shifting supply and demand curve, and new formats emerging, such as memory being a key focus this year, 2m6s.
• There is pressure to solidify a deal with the CME, with regulatory bodies being worked with to create a traditional data futures product, similar to those done for oil and gas, with no major concerns expected, 2m6s.

Apple's Privacy-Focused Siri Upgrades

• The offer is global, with the first GPU indices launched with Bloomberg last year, covering all indices except China, and the goal is to capture the global market, 4m6s.
• Apple is set to unveil its iOS 27, featuring a revamped Siri experience with new privacy features, such as auto-deletion of chatbot conversations, which will help differentiate Siri from other chatbots, 6m8s.
• Apple's focus on privacy has been a key differentiator for the company, with an emphasis on better encryption and privacy than the competition, and this will extend to the new Siri features, 8m10s.
• The new Siri features will include the ability to set chat conversations to auto-delete after a certain period, such as 30 days or a year, which will be a unique feature compared to other chatbots, 10m0s.

MIAX and the Bloomberg 100 Equity Index Futures

• The conversation starts with a discussion about technologically advanced chatbots and their memory, where it is mentioned that some chatbots can retain memory forever, while others, like Apple, retain memory for a shorter amount of time, with a technique called differential privacy to retain some memory without including exact user information 10s.
• Miami International Holdings, MIAX, has seen tremendous growth since it started operating a decade ago, and it is launching the Bloomberg 100 Equity Index Futures, a product geared towards retail traders, with commission-free trading, and the company is leveraging the favorable tax environment in states like Florida 2m6s.
• The launch of the Bloomberg 100 Equity Index Futures is part of a series of products, and MIAX is looking to further growth and development in the southern states, with its trading floor located in Miami, while its headquarters are in Princeton, New Jersey 2m6s.

Semiconductor Market Volatility and Samsung Labor Dispute

• The semiconductor market is experiencing losses, with a 2.4 percent drop, and NVIDIA's earnings are expected on Wednesday, which is highly anticipated, especially since AI is not being seen as a savior to the markets 5m42s.
• There is a risk aversion creeping into the market, largely focused on geopolitics, with oil prices rising, and chip stocks under pressure, while Samsung shares in South Korea are up about 3 percent after a local court move to limit the threat of a potential strike 6m15s.
• The labor dispute between Samsung and its union is a significant issue, as the company has never had a major conflict with its union before, and the threat of a strike could disrupt the memory chip market, which is experiencing a boom due to the demand from the AI industry 8m30s.
• The union and the company have agreed to go back to the table after setting a possible strike date, but a resolution has not been reached, and there is still a possibility of a strike later this week that could affect the AI industry more broadly 10s.

AI Profits and Policy in South Korea

• In South Korea, there is a debate about how AI profits should be shared with the nation, and the government is considering a citizen dividend to establish a more stable foundation for society, with one of the policy advisers to the president suggesting taking some of the profits from big companies to support people who want to start up companies 2m6s.

Google's Compute Bottleneck and AI Research Challenges

• Google's leadership is prioritizing computing power as a resource, but the bottleneck has frustrated some employees, prompting a number of researchers to leave for startups where they can get more access to compute 4m42s.
• The shortage of computing power has become more acute in the AI industry due to the vast amounts of computing power required by large language models, and researchers often cannot get meaningful research done with their allocations if their work is not a priority at Google 6m15s.
• A female-led company has gotten backing and has articulated the challenges of getting access to compute, citing the popularity of consumer products generated by AI, the growth of the cloud business, and the increasing computer intensity of AI research 8m10s.

Frontier Labs and the Pursuit of Super Intelligence

• Frontier Labs, valued at $4.65 billion, is conducting experiments on how to safely improve itself, and its CEO aims to build super intelligence that automates knowledge discovery, allowing AI to replace human ingenuity and come up with ideas and implement them 12m20s.
• A company has been founded with a team of cofounders from notable labs like Google, DeepMind, OpenAI, and Salesforce Research, focusing on open-ended AI development and having the entire company work on allowing AI to build the next better version of AI, with the goal of creating a Eureka machine that can invent everything else thereafter 10s.
• The company requires a significant amount of money to operate, with compute being one of the largest costs, and has partnered with companies like Nvidia to help reframing the demand for GPU's and other computing resources 2m6s.
• Safety is a huge concern for the company, with some of its cofounders having done research on improving the safety of large language models, and the company takes safety very seriously, believing that their approach is the best and fastest way to achieve super intelligence 4m20s.
• The company's founder works extensively and has incredible teams around them, including AI agents and humans, and the company shares equity with its employees to encourage and excite them 6m15s.
• The founder believes that sharing equity is important to create a sense of ownership and to grow something amazing together, and that this approach will help the company to achieve its goals 8m30s.

Iren's AI Infrastructure and Strategic Growth

• Iren, an AI cloud provider, has partnered with Dell Technologies as its key server partner and uses Nvidia GPU's, and the company's cofounder and co-CEO, Daniel Roberts, has been busy with M&A activities, including acquiring a marketing creative agency to increase brand awareness and grow its customer base 10m50s.
• Iren has taken a strategic approach to building its infrastructure, avoiding metropolitan areas and instead locating in regional communities with an abundance of renewable energy, which has created a structural advantage over time and has been well-received by the local communities 14m10s.
• The company is facing a challenge in meeting the vast demand for compute, with demand vastly outpacing its ability to supply it, and is working to address this issue 16m40s.

Challenges in Scaling AI Infrastructure

• The biggest barrier to getting more capacity online is the dialogue era of AI, where crafting prompts carefully takes about 60 seconds, and the focus is on what the chokepoint is, which includes power, steel, concrete, and building physical instrument structures that take years to complete 10s.
• Building a gigawatt factory can take until 2030 to get the first compute online, and it takes 18 months to get the attention of utilities to work with, which is the starting point, and then it takes six years to get permissions and get the project online 42s.
• Utilities are doing their best to manage complex networks and power demand, but they are naturally risk-averse and deal with an inundation of requests, which is why reconnecting capacity is valuable, as they are underwriting guaranteed power 1m14s.
• Working with Dell and Nvidia involves having great relationships, relying on them for servers and GPUs, and working on reference architecture for a gigawatt factory, with the opportunity to build up the whole ecosystem from land, steel, and concrete to end users 2m6s.
• Texas is considered a good place to build a data center, with the majority of capacity being there, and every geography handles things differently, but the state has been pretty good in terms of handling data center projects 3m30s.
• The whole industry is moving from silicon being the chokepoint to solving steel, concrete, and kilowatts, and everyone is trying to work it out, including the current administration, which has an answer for that 4m40s.

Global Expansion and Regulatory Challenges

• Working across borders and getting projects done in different jurisdictions is challenging, with projects in North America, Europe, and Australia, and the need to travel frequently, with the CEO's answer to where he lives being "on a plane" 5m50s.

Hedge Fund Activity and Market Volatility in Q1

• New regulatory filings, known as 13F filings, are revealing how hedge funds are positioning around big tech and the AI trade, with Bill Ackman taking more than $2 billion worth of Microsoft stock, making it his fourth biggest holding, and believing Microsoft is stronger and more resilient than investors think 7m10s.
• Other hedge funds, such as Tiger Global, are reducing their stake in Microsoft by more than half, or getting out entirely, showing divergent views on the company's value and potential 8m30s.
• The stock of a particular company fell 23% in the first quarter but has since rebounded somewhat 10s.
• In the first quarter, there was a decent amount of selling from the Tiger Cubs, which are firms that tend to focus on tech and are directional and intentional in their investments 42s.
• The first quarter was incredibly volatile, with many funds losing a lot of money, which could be a reason why some firms tried to sell their holdings 1m6s.
• Some popular stocks, such as Amazon, experienced a decline in the first quarter, with Amazon's stock sinking about 10% 1m23s.
• Certain firms, including Tibor Global and Code-2, trimmed their stakes in popular companies, while Berkshire Hathaway exited some companies altogether 1m42s.
• Despite some firms selling their holdings, others have written up higher on the back of the news, indicating differing views on the market 2m6s.