iShares S&P GSCI Commodity-Indexed Trust (GSG)

The Goldman Sachs Group provides investment banking, securities, and investment management services to corporations, financial institutions, governments, and high-net-worth individuals worldwide. Goldman has three operating segments: Asset & Wealth Management, Global Banking & Markets, and Platform Solutions, with the latter including transaction banking and consumer partnerships.

10 Questions About Artificial Intelligence: Part 2 Artificial intelligence (AI) has been and remains one of the fastest-growing technology developments in history. Just a few years after becoming broadly known, AI represents a significant portion of the global economy and is now the preeminent focus among investors. Developments in AI have progressed and continue to unfold rapidly, and new revelations on one day - say, the release of Meta Platforms Inc.'s Muse model - can make preconceptions about AI seem out of date. Based on input from clients, investors, industry experts, and our own analytic team, Argus has put together what we see as 10 current key and enduring questions about AI, along with our responses. We've broken these 10 questions into Parts 1, 2, and 3. Part 2, presented below, addresses three key questions including risks that white-hot growth at the GPU makers might slow, key adjacencies benefiting from AI infrastructure build-outs, and a discussion of companies best and worst-positioned amid AI disruption. Following publication of Part 2, we will be publishing Parts 3. We welcome feedback from clients and will continue to refine our 10 questions about AI in real time in response to this exciting and ever-changing opportunity. QUESTION 4: At the heart of the artificial intelligence (AI) revolution are the graphics processing unit (GPU) developers, led by Nvidia Corp. and Advanced Micro Devices Inc. and followed by a host of others. What is the risk that these companies' white-hot growth slows or stops? ANSWER 4: AI infrastructure investment has broadened out from an initial focus on hyperscaler data centers to a broader set of customers, including enterprises, sovereigns, and neoclouds. Wider adoption is being driven by the much more diverse range of use cases driven by agentic AI. This broadening of the market has benefited the makers of GPUs, which represent the foundational piece in massive clusters required for model training, deep learning, and inference. Nvidia is the tentpole company of the AI revolution, initially repurposing its GPU gaming cards for early AI tasks. The company was also subject to fluctuations in the crypto market and consumer cycles in gaming demand. Nvidia's revenue declined 7% in fiscal 2020, rebounded to mid-50% annual growth in FY21 and FY22, and was flat in FY23 on softening in crypto and consumer end markets. After two years of triple-digit growth in FY24 and FY25, growth moderated to 6% in FY26 before reaccelerating to 87% in the January 2027 fiscal year. Although growth is on a path to 'moderate' to the 70% range in FY28, note that growth in dollars remains robust: rising from $216 billion in FY26 to an estimated $410 billion in FY27 and to $680 billion in FY28, according to the consensus. If not for supply constraints, according to CEO Jensen Huang, FY28 top-line growth would be meaningfully higher. Nvidia has teamed up with Wall Street giants (Apollo Global Management Inc., Blackrock Inc., Goldman Sachs Group Inc., and others) to raise $500 billion in capital for AI infrastructure. Nvidia anticipates multiple years of strong growth ahead. AMD is No. 2 in GPUs and is also benefiting as inference drives demand for industry standard servers (ISSs) using AMD's central processing units (CPUs). AMD generated $6.6 billion in revenue in 2010 but did not get back above that number until 2019. After three years of mid-double-digit growth in 2020-22, revenue declined in 2023 and grew 14% in 2024. Fueled by both its CPU and GPU franchises, growth accelerated to 34% in 2025 and is on a path to accelerate from 47% in 2026 to 73% in 2027. In dollar terms, AMD's revenue is forecast to rise from $35 billion in 2025 to $51 billion in 2026 and $88 billion in 2027. Companies including Broadcom Inc. and Marvell Technology Inc., which make application accelerators and XPUs (both of which are task-specific GPUs), are reporting similar growth trajectories as customer and market opportunities broaden. AI infrastructure spending grew about 60% annually to $320 billion in 2025, according to industry analysis firm IDC. AI infrastructure spending is on a path to exceed $1 trillion in 2029, the firm estimates. IDC noted that AI infrastructure investment is accelerating at a pace that signals a clear shift away from experimentation to a sustained, multiyear capital commitment cycle by end customers. Based on these and other forecasts and guidance from the major GPU providers, we expect GPU and related semiconductor growth to sustain and in many cases accelerate from high levels in coming years. QUESTION 5: What are some the key adjacencies that are benefiting from AI development? And which of these companies are best positioned to thrive? ANSWER 5: The initial gold-rush phase of AI was characterized by intense activity among hyperscalers to build massive GPU clusters and support infrastructure as they sought to develop and train their own large language models (LLMs). But this focus on AI data centers also represented one of the key bottlenecks in AI infrastructure. AI content, in the words of rock singer Meat Loaf, was 'all revved up with no place to go.' Late in 2024 and early in 2025, Argus noted that the AI revolution had primarily been focused within hyperscalers (major cloud service providers) and was concentrated in GPU compute at massive cloud data centers. Argus forecast that in coming years, the AI revolution would spread out from this initial concentration at hyperscalers to include customers in other categories including major enterprises, sovereigns, and neoclouds. We further forecast that the need to accommodate AI-generated data would trigger massive business growth at data storage and memory companies. And we forecast that the need to disseminate AI data to these new classes of customers would trigger massive business growth at data networking companies. Those forecasts began to play out during 2025, as data storage and memory companies had huge business growth and were among the best stocks in the S&P 500. During 2026, data networking companies have reported strong double-digit growth rates, and their stocks are leading the Information Technology sector and broad market. The first phase of generative AI, dominated by the training of LLMs, is giving way to the age of agentic and physical AI, which requires much more compute-intensive inference in multimodal and frontier models. This is further contributing to AI-related data growth. Technology segments benefiting from the need to deliver AI-as-a-system include data storage and memory, networking, and traditional enterprise infrastructure. In these segments, Argus has identified multiple companies best positioned to benefit from the broadening in AI demand. In the networking space, Argus favors optical-networking leader Ciena Corp., 'cloud titan' network provider Arista Networks Inc., and enterprise industry leader Cisco Systems Inc. as three companies well positioned to deliver AI services to customers. In the data storage and memory space, Argus favors Seagate Technology Holdings plc and Western Digital Corp., which form a duopoly in nearline hard disk drives for AI data centers; Sandisk Corp. in the nonvolatile memory space; and Micron Technology Inc., the leading U.S.-based provider of DRAM and NAND memory. Data storage systems providers including NetApp Inc. are subject to higher input costs for their solutions, but NetApp has driven revenue acceleration with successful cost pass-alongs and higher prices. Dell Technologies Inc. is a good example of the fast-developing market opportunity in the legacy enterprise infrastructure space. As of mid-2026, Dell had thousands of customers for its AI servers and related IT infrastructure including storage arrays, industry standard servers, and networking gear. None of these customers included the major hyperscalers. As of August 2026, Dell expected at least $74 billion in AI-optimized server revenue in FY27, almost triple the $25 billion in FY26. Dell's traditional x86-based ISS revenue grew more than 120% annually in the most recent quarter. In 1H27, Dell generated as much revenue from ISS and networking than it generated in any full prior year in company history. Outside the Information Technology sector, industrial companies providing power, cooling, electrical, and related solutions are also experiencing extraordinary growth. Top beneficiaries include Caterpillar Inc., Emerson Electric Co., Honeywell International Inc., Rockwell Automation Inc., and others. As AI demand continues to broaden out, we expect strong and even accelerated growth for providers of the adjacent infrastructure enabling the AI revolution. QUESTION 6: Which incumbents are most vulnerable to AI disruption? And which are best positioned to integrate AI defensively? ANSWER 6: The AI revolution has been so consistently strong and inclusive that it initially swept along much of the Information Technology sector. But as the AI revolution moves beyond year three, several significant developments have revealed risks to industry incumbents across the technology landscape. These developments primarily relate to AI displacement and the impact of soaring component costs in a supply-constrained environment. Fears of AI displacement have been mainly centered in the enterprise application software space. The trajectory of these fears is expressed in the iShares Expanded Tech-Software Sector ETF (IGV). The software ETF hit an all-time closing high around 117 in September 2025. The release of certain highly capable frontier models, most notably Anthropic PBC's Claude Code, showed that many formerly proprietary software tasks could be done quickly and cheaply with AI. IGV declined gradually into year-end 2025 and then its decline accelerated in the new year, reaching bottom at a closing low of 74 in April 2026. IGV has since rebounded to a low-100s level, as software companies counter displacement fears by highlighting their proprietary products, deep customer engagements, and defense against 'weakest link' cybersecurity vulnerabilities. Component prices and particularly memory prices have soared due to voracious hyperscaler appetites for hardware to support infrastructure investment. Companies with high personal computer exposure such as HP Inc. report that memory is now up to 70% of the bill of materials cost of building a PC. In releasing its newest iPhone family in September 2026, Apple Inc. raised iPhone prices by approximately $100 per device and more for high-memory-content phones. These rising costs and the urgency to invest in AI are causing chief information officers to reshape their IT budgets in some cases away from long-standing customers. International Business Machines Corp., which in April 2026 warned of a decline in 2Q26 infrastructure revenue, reported a steeper-than-expected demand decline in July 2026. Customers shifted purchases away from IBM's mainframe computers and its complementary transactional software stack. These customers redirected spending to ISS servers, a market in which IBM no longer competes, and data storage in order to secure supply-constrained IT hardware ahead of further price increases. IBM shares declined sharply on the 2Q26 revenue miss and are down 22% year-to-date. But the stock is up sharply from pre-AI days. IBM Watson playing Jeopardy against a past champion helped popularize the concept of AI years before ChatGPT. IBM, like the software companies and makers of AI PCs, should ultimately benefit from positioning within the AI ecosystem. We expect creative disruption to continue to challenge industry incumbents. As to which companies are best positioned to integrate AI defensively, no company is in a position to sit on the sidelines and watch its competitors bring AI practices into their everyday operations and into customer interactions. Market sectors at the forefront of AI adoption, for both offensive and defensive reasons, include Healthcare, Information Technology, and Financial. Even in regulated sectors such as Utilities, which would seem to have less urgency to integrate AI into their operations, sector companies are adding smart meter technology and using AI to upgrade their internal operations. In short, nearly every company inside and outside of the AI ecosystem will need to integrate AI both offensively and defensively.

Goldman Sachs is a storied financial institution, founded in 1869 and best known for its role as a leading global investment bank. The firm has a significant reach across global financial centers and has been the leading provider of global merger and acquisition advisory services, by revenue, for the past 20 years. Since the global financial crisis, Goldman has expanded its offerings into more predictable, fee-based businesses like asset and wealth management, which contributed roughly 30% of postprovision revenue at the end of 2025. The firm generates revenue from investment banking, global market making and trading, lending, asset management, wealth management, and a small and declining portfolio of consumer credit card loans.

Goldman Sachs is a storied financial institution, founded in 1869 and best known for its role as a leading global investment bank. The firm has a sprawling reach across global financial centers and has been the leading provider of global merger and acquisition advisory services, by revenue, for the past 20 years. Since the global financial crisis, Goldman has expanded its offerings into more predictable, fee-based businesses like asset and wealth management, which contributed roughly 30% of postprovision revenue at the end of 2025. The firm generates revenue from investment banking, global market making and trading, lending, asset management, wealth management, and a small and declining portfolio of consumer credit card loans.