Generative AI: Shaping the Future of Economy, Markets, and Investments

November 2024

I recently attended the AI in Media conference at Oxford, where experts and thought leaders gathered to explore artificial intelligence's transformative potential. Each seminar provided unique insights, ranging from the technical foundations of generative AI to its far-reaching implications across various industries, particularly in media. Watching live demonstrations and engaging in discussions about real-time applications of AI made it abundantly clear how significant its impact has become.

I frequently encounter the hype and momentum building around AI. The excitement is undeniable, but beneath it lies a profound transformation. What follows is a high-level perspective—a 10,000-foot view—of the generative AI landscape and how it is reshaping our world, along with ways investors can participate in this change.

We are on the brink of the generative AI revolution, reminiscent of the internet revolution that began in 1993. While its potential to transform industries, streamline processes, and democratize innovation is tremendous, it also has its own share of nuances.

The Rise of Generative AI

GenAI is a nascent technology that will shortly upend the course of technology and industries with its singular capacity for content creation across multiple media types. Whereas other forms of AI are geared predominantly towards prediction or classification, GenAI can generate text, images, audio, video, and even code, matching or sometimes exceeding human ingenuity. This is achieved through more sophisticated machine learning models, such as neural networks, LLMs, and GANs. These technologies allow the output from a GenAI model to be similarly styled, in tone and even in complexity, to what it has seen in its training data.

Ahead now lie promising changes in Generative AI. Examples of such furtherance include enhanced NLP, which allows for more sophisticated and natural interaction between human beings and computers and, therefore, makes the AI system more conversational. Advanced image and video generation would help industries like entertainment, advertising, and education create very realistic and imaginative visual content. Additionally, better context understanding and reasoning will enable AI systems to make more accurate predictions and decisions by effectively interpreting context. Innovations such as neural architecture search and reinforcement learning-based models are expected to push the boundaries of how AI is designed and trained, while improvements in hardware, especially quantum computing, promise to accelerate its computational capabilities.

Generative AI already influences how businesses operate and compete, shifting key elements of business strategies. The growth of subscription-based and on-demand services is transforming traditional revenue streams. In addition, AI-driven insights allow companies to offer personalized products and services that cater closely to individual customers' preferences, enriching customer experiences. Predictive analytics adds a lot to data-driven decision-making by availing actionable insights that drive strategy, mitigate risks, and unlock innovation. This frees businesses from mundane tasks, easing loads in operations to give them a more strategic way of applying resources to achieve strategic goals.

As Generative AI slowly permeates business functions across organizations, its applications in information technology stand out. Here, GenAI automates software testing, debugging, and code generation, significantly minimizing development timelines and boosting overall efficiency. In marketing, it personalizes content for customers, analyzes market trends, and interprets consumer sentiment, enabling companies to launch highly targeted campaigns.

Generative AI will help sales teams gain profound insights into customer data, locate potential leads, and make fairly accurate sales forecasts that enhance decision-making and resource allocation. It can process a huge amount of data to unlock patterns and trends, automate repetitive tasks like data entry and report generation, and offer actionable insights to mitigate risks in financial services. It also optimizes supply chain management by anticipating demand, locating bottlenecks, smoothing logistics and enhancing inventory management.

In procurement, Generative AI can evaluate supplier performance and perform contract management, hence bringing in the proficiency factor and compliance. Similarly, in talent management, it automates resume screening processes, recognizes top candidates, and provides personalized career development recommendations, subsequently enhancing overall workforce productivity. These diverse capabilities are enhanced through prompt engineering and fine-tuning, instructing machine learning models to do specific tasks in particular ways, especially in natural language processing.

Adoption of Generative AI is accelerating in more ways than one across industries. Companies use the technology to offer personalized product recommendations, analyze social media interactions, provide AI-powered network monitoring, and even design innovative vehicle concepts. These are just some examples of applications that indicate the rising dependence on AI for innovation and efficiency.

Generative AI in 2024

In 2024, generative AI spending surged to $13.8 billion, up sixfold from the previous year. This rapid growth has created a high degree of optimism among decision-makers, with 72% expecting wider adoption of generative AI tools in the near future. Departmental spending also increased significantly, with a nine-fold increase in departmental AI investment and a twelvefold growth in vertical-specific AI spending.

Companies spent about $4.6 billion on generative AI applications in 2023, eight times the amount paid in 2023. Organizations have reported an average of ten use cases, with 24% already prioritized for near-term action. The leading use cases are code copilots, support chatbots, and enterprise search functionalities.

Departmentally, the technical departments accounted for 55% of AI spending, with IT, product engineering, and data science leading the investment areas. To a lesser extent, customer-facing functions such as support and sales also contribute.

Industry vertical applications also continue to progress well. Healthcare leads with $500 million, followed by the legal verticals with $350 million. The financial services and media and entertainment sectors each invested $100 million in different forms of automation and optimization.

Foundational model investments reached $6.5 billion, and the trend was primarily multi-model for enterprises, with typically three or more foundation models. Market share is 81% for closed-source models against 19% for open-source models. This leads to the interesting fact that while Open AI's market share declined from 50% to 34%, Anthropic increased from 12% to 24%.

Of these, the AI design patterns show that Retrieval-Augmented Generation has reached 51% company adoption, while fine-tuning of production models is still uncommon. A final milestone is the advent of agentic architectures, powering 12% of implementations.

Businesses are now making workflows more effective with AI, and autonomous solutions start to set in. Innovations in finance and market strategies evidence this. Companies are almost evenly split between developing AI tools and buying from vendors, reflecting a growing confidence in their internal capabilities. Enterprises prefer tools with quantifiable ROI, while most failed implementations result from underestimating integration costs and addressing data privacy concerns.

The generative AI landscape is evolving rapidly, showcasing many use cases that support creative and practical tasks in many domains. The most common use case is the introduction of autocomplete features that help users create drafts that they can refine. Tools like Jasper.ai revolutionize writing, GitHub Copilot aids programmers, and Stable Diffusion accelerates rapid prototyping, showing the versatility of generative AI in fostering creativity.

In addition, AI-powered answer engines enhance the search experience by providing direct answers instead of merely linking to information. Tools like Context, which helps users find YouTube playlists, and Elicit, which helps summarize research, are good examples of how these engines can deliver immediate, relevant information to users.

Engagement with generative AI tools will also vary significantly, ranging from an occasional use by some to explore options for specific tasks to consistent use by others. While organizations are increasingly adapting AI tools to different user profiles, factors that drive such varying levels of engagement remain unclear.

As generative AI adoption matures, the focus is shifting away from broad adoption metrics to the measurable impacts on productivity and quality. There is evidence that high-impact, low-frequency use cases executed by small, dedicated teams can drive substantial benefits, but there are also trade-offs with these tools. For example, while GitHub Copilot can improve output quality, it can also introduce errors or decrease job satisfaction by undermining creativity and collaborative effort.

Building trust in AI is another complex challenge. The balance of trust needs to be appropriate; too much trust could lead to abuse, while too little trust will prevent proper usage. Responsible AI teams must sort through these complexities to facilitate the appropriate use.

There are constant debates on how generative AI influences different levels of proficiency. For instance, early indications suggested that AI would be a great leveler and might help novices more than experts. However, newer studies are showing mixed results, raising the question as to under what circumstances these tools will be helpful for either party.

The effectiveness of training related to generative AI tools also has a big question mark. Most current training programs are not able to consistently prevent user errors. This situation creates an immediate need for more customized training methods that assure responsible tool usage effectively.

Rapid improvements in AI models will pose different challenges in the future, especially concerning stability and reliability. We are yet to understand whether we approach the regime of diminishing returns in AI development or whether truly transformative gains are still realistic, and the debate is ongoing.

Generative AI interface design is still uncertain, and experimentation is needed to see whether chatbots, task-oriented agents, or embedded applications will emerge as the dominant models for business and industry use. Beyond technical capabilities, the skills required to integrate AI effectively encompass domain knowledge and problem-solving skills. Identifying the most relevant skills for nontechnical workers will be necessary for success in an increasingly AI-driven work environment.

Other current research points to the role of AI in research and development, improving productivity, especially senior researchers, yet simultaneously leading to decreased job satisfaction, potentially stifling creativity and skill application. One study on GitHub Copilot showed that generative AI can improve the task completion rate by as much as 26%, particularly to the benefit of less experienced developers. However, challenges remain, including juniors trying to upskill their seniors in emerging technologies so that custom training strategies will be required.

The near-term future is more speculative, with assumptions on AI's role in complex task handling, including the smooth integration of multimodal input and external knowledge. Challenges remain, especially in developing multimodal capabilities and ensuring extensibility for booking flights.

Significant changes are foreseen in creating multimedia content with the development and application of generative AI. Among a few challenges that lie ahead are capturing the temporal dependencies of long-form contents such as films or video games, ensuring intuitive but editable outputs, moving beyond pure text-to-X interfaces, and increasing personalization so users can feature in customized games, books, etc. These developments only lead to increased possibilities of generative AI, day by day, in reshaping domains.

While mainstream vendors currently dominate this category, growing dissatisfaction among their customer bases has provided opportunities for startups to fill the gaps and challenge traditional vendors to at least meet specific needs. This trend is expected to influence many industries, mainly IT, product engineering, and data science; much progress is already being witnessed in healthcare, legal, financial services, and media.

Companies are stabilizing their tech stacks in support of AI efforts and adopting strategies that exploit multi-model approaches. Other emerging trends include rising retrieval-augmented generation and the use of vector databases. Agents will drive the next wave of transformation, startups will challenge incumbents, and AI talent will be in higher demand and command higher salaries due to increased competition.

The new up-and-coming trends and technologies within Generative AI include agentic AI and multimodal models. Agentic AI works autonomously to build iteratively via planning, researching, and refining a problem. Consequently, this is especially valuable in such applications as legal document processing, where the need for precision is very high, or in healthcare diagnostics. Textual AI integrates images and videos on top of texts, opening new directions for industries with diverse datasets. Applications range from video indexing to metadata generation, opening advanced content analytics and management.

Agentic systems exhibit two fundamental critical design patterns regarding the design of workflow: Reflection lets the AI systems review and edit their outputs to be more accurate, while Tool Use gives them greater functionality by making use of tools or APIs outside; in Planning and Reasoning, complex tasks are broken down into more manageable steps that then get carried out systematically, while Multi-agent Collaboration allows multiple AI agents that work together may have different stages on which other agents would focus to obtain a sound output.

With such development, GenAI will give rise to unprecedented personalization across industries: from adaptive learning paths on educational platforms to custom content delivered upon entertainment platforms. Businesses will embrace GenAI as a way to produce customized content, while the notions of marketing, product design, and customer relationship management will not remain the same.

The rise of GenAI will significantly impact the job market, automating repetitive tasks while creating new opportunities that focus on managing and optimizing AI tools. Reskilling programs will be vital to help workers transition into hybrid roles that combine human creativity with AI capabilities. GenAI's influence will drive new business models and economic shifts, reshaping industries such as healthcare diagnostics, content creation, and legal services.

Generative AI & Impact on Economy

With generative AI, the full transformative power is increasingly evident, but the final result will come piece by piece. The immediate-term outlook, for instance, has clouded uncertainty over ethical considerations, regulatory complexities, and changes within transition dynamics at the workforce. Nevertheless, from what has emerged for long-term AI, substantial ways should emerge in which this emerging technology might enhance productivity, economic growth stimulation, and innovativeness.

Generative AI often resembles historical GPTs, like the steam engine, the internal combustion engine, electrification, and computers. These three groundbreaking technologies were major accelerants of economic growth, transformative across industries, and dramatically altering the shape of our societies. Indeed, generative AI is associated with three core characteristics shared among these historic general-purpose technologies:.

First, it is improving rapidly. The rate of progress for generative AI is nothing short of breathtaking. For example, OpenAI's GPT-3.5 outperformed 10% of human test-takers on the U.S. bar exam, while its successor, GPT-4, released shortly thereafter, surpassed 90% of test-takers. This rapid trajectory illustrates generative AI's capacity to drive substantial economic and societal change in a remarkably short timeframe.

Second, there is its pervasiveness. A key trait of general-purpose technologies is their widespread adoption across multiple sectors and integration into various aspects of economic activity. Evidence suggests that many workers are either using or planning to adopt generative AI tools. This swift uptake across different industries indicates that generative AI is on a path to becoming deeply embedded within economic systems, thus enabling transformational change.

Finally, generative AI brings complementary innovations in various fields, including science, engineering, and business processes. It enhances both the individual tasks and has the potential to automate and redesign the entire workflows in every direction, pushing innovation forward. This ability to integrate and better the present processes reinforces the classification of generative AI as a general-purpose technology.

Unlike earlier general-purpose technologies, which had to be embodied in new physical infrastructure, for example, factories for the steam engine or power grids for electrification-generative AI are built on top of existing infrastructure, such as internet-connected devices and cloud computing. This greatly lowers barriers to adoption and enables the technology to permeate economic systems more rapidly.

Moreover, generative AI is based on natural language processing, so it has been made available for many users, contributing much to its fast diffusion across different industries and groups of people. The intuitive interface it offers allows people and organizations without vast technical experience to use the technology, which increases the effects on economies and societies. Already, the influence of Generative AI can be seen in stock markets. The most vivid example is Nvidia, whose chips underlie many different types of generative AI. Fueled by a demand surge, Nvidia's market capitalization has skyrocketed, making it the world's most significant post-war stock and one of this year's most prominent contributors to equity markets, earnings growth, and industrial production returns. This evolution has been no different from other transformative technological periods, like the mainframe era of the 1960s and the fiber-optic boom of the 1990s. Yet these gains have been occurring alongside record levels of corporate concentration in the U.S.

Research shows that Generative AI is set to increase global GDP by some 7%--about $7 trillion--and raise productivity growth by 1.5 percentage points over the next decade. With these projections, generative AI can be a key driver of future economic growth and equity market expansion. Among the many predictions about the future implications of Generative AI, there is one critical inconsistency- understandably so-when it involves its potential contribution towards global GDP. For example, the consultancy Accenture estimates that, on average, less than $1 trillion additional annual value will be created over the next decade. In contrast, McKinsey provides a more upbeat forecast of extra global GDP gains ranging from $2.6 to $4.4 trillion annually.

Taking the upside of McKinsey estimates, the uplift to global GDP could be more than the total size of the UK economy, currently ranked the sixth-largest globally. If we take conservative estimates, $1 trillion additional yearly would translate into adding one economy the size of the Netherlands or Saudi Arabia every year into global GDP.

With industrialized OECD countries greying, GenAI is changing the focus from GDP to productivity. Estimates by Goldman Sachs project that this growth may be between 0.1% and 2.9% per year, while McKinsey gives a wider range, starting from 0.2% and going as high as 3.6%. These estimates suggest that Generative AI has the potential to actually double the historical annual productivity growth rates in OECD nations, which used to range from 0.5% to 1.6% before COVID-19 started.

Comparing projections for Generative AI with earlier technologies, studies find that forecasts from the 2010s about the broader impact of AI are dramatically more ambitious than today's projections of Generative AI. This is particularly the case about productivity, suggesting that today's assessments of Generative AI are more closely oriented to its potential for cost savings rather than gains in productivity.

Generative AI will disrupt enterprise software, healthcare, financial services, and other industries while making workflows more seamless, automating mundane tasks, and catalyzing next-generation business app development. Goldman Sachs Research estimates that against a global software industry of $685 billion, the total addressable market for generative AI software would be $150 billion. Huge potential abounds in a raft of applications, improving office productivity and adding greater relevance to health diagnostics.

Hence, it could bring spectacular gains in productivity. According to estimates by the McKinsey Global Institute, generative AI could add $2.6 trillion to $4.4 trillion annually to corporate profits globally, across 63 use cases scrutinized-from customer interaction and marketing content creation to the drafting of software code.

The productivity gains that generative AI enables are likely to add 15% to 40% value to AI and analytics relative to prior technologies. As generative tools become common, that could double the economic impact of AI. Realizing these gains will depend on workforce adaptation, including retraining and upskilling employees for evolving job requirements.

This makes generative AI well on track to give equity markets with exposure to AI a material lift. For example, if this technology boosts annual productivity growth by 1.5 percentage points over the next ten years, that would translate to a compound annual growth rate in S&P 500 EPS of 5.4% over the next 20 years and imply a gain of 9% in the S&P 500 fair value.

A recent study by NBER investigated the effect of generative AI on firm valuations and found that the greater the firm's exposure to generative AI, the more significant the valuation increases after the release of ChatGPT. Such a rapid uptick in interest suggests that investors believe generative AI has the potential to raise productivity and profitability, thus leading them to update the future earnings prospects for such firms.

The study also showed that the valuation gains because of generative AI are heterogeneous across industries: the sectors in which more of their jobs are considered susceptible to being automated with this kind of new AI witnessed relatively more significant gains. Parallel to these findings, the test of corporate communications showed that a firm was much more likely to discuss the business strategies relating to AI and further benefits during the earning calls that had taken place following ChatGPT. This proactive communication has changed investor sentiment and influenced the valuation changes.

Historical comparisons complete this discussion. Previous productivity booms, such as the advent of electricity in the early 1900s and the rise of personal computing and the internet in the late 20th century, have often resulted in significant impacts on equity markets, though mainly after the productivity benefits become apparent. The dot-com bubble was one example of the risks involved, where high-growth tech firms faced significant collapses when investor expectations became unsustainable.

Considering current valuations, the S&P 500's risk premium and growth expectations align with historical averages, reflecting a moderate optimism regarding AI's potential. However, some of the more high-profile AI stocks have valuations reminiscent of the dot-com era, raising concerns about the potential risks for some companies.

In this context, AI "agents" are a class of advanced generative AI systems positioned as "digital labor" able to execute tasks well beyond traditional bots. However, these technologies are still largely unproven, and they function more like sophisticated autopilots rather than fully autonomous agents capable of making decisions on their own. Salesforce's Agentforce 2.0 is a case in point where the potential of AI has been demonstrated: CEO Marc Benioff has hinted that it can cut costs by as much as half by reducing human intervention, especially in customer service. Of course, such prospects notwithstanding, several challenges lie ahead: policy responses like increased tax rates, rising interest rates, and recession might turn some of these gains around.

Historical analysis suggests that growth in productivity in and of itself does not guarantee equity market returns-the only 1% variation in five-year S&P 500 performance is explained by five-year productivity growth. Moreover, the parallel with the boom in dot-com during the 1990s brings out the risks from overly inflated investor expectations as AI gains traction.

Generative AI Investment Landscape 

The generative AI landscape has witnessed phenomenal progress nearly two years since the launch of ChatGPT. As this revolutionary technology continues to evolve, investment opportunities have three distinctive phases.

Currently, we find ourselves in the first phase of capital investment. This phase underscores the need for infrastructure building that meets the computational demands required by generative AI. Key developments in this domain include significant upgrades to data centers, benefiting companies in semiconductors and those involved in electrification, cooling, and engineering services. Hyperscalers and well-funded startups fiercely compete, pouring significant investments into accelerated computing. As a result, we can expect continued capital expenditure to enhance computational power shortly.

Looking ahead, the new paradigm for software will be to create applications that catalyze productivity and revenue growth. Early applications are already gaining traction in coding efficiency, customer service automation, IT support, digital advertising, and industry-specific vertical software solutions. This next wave represents significant investment opportunity, especially for companies leading these innovative areas.

The new wave belongs to a phase of broad-based deployment: AI tools would be much in supply and more employed, totally changing revenues, earnings, and free cash flows of businesses. This shift has been, and will be, a very slow process but will most surely materialise over the next years and probably even two decades.

Considering the historical context, the interesting thing is how the current landscape of AI parallels the early days of the internet. Once a novelty, the internet would disrupt communication and connectivity over the next three decades since the Netscape browser launch in 1994. Today, technology stocks represent 45% of S&P 500 free cash flow, up from 15% pre-internet, with almost 30% profit margins. This history can teach us that market participants commonly underestimate industry-changing innovation. Breakthroughs like the personal computer, internet, and cloud computing were frequently over their estimate, reflecting the challenges of forecasting disruptive technologies.

Generally, "shovel providers in a gold rush" are companies that make their money not by directly participating in the trend or boom but by providing the enabling tools, infrastructure, or resources that allow others to. This model is taken from historical gold rushes, wherein merchants selling shovels, picks, and supplies often profit more consistently than miners.

In this new world of generative AI, a cadre of companies are fast becoming the early beneficiaries-those that might be called "shovelers." These are mainly the firms providing the necessary infrastructure: semiconductor developers, cloud computing service providers, and firms offering AI services.

Leading the charge is Nvidia, which has emerged as the dominant supplier of GPUs, the essential building blocks for training and operating large language models. The surging demand for AI hardware places Nvidia in a singular position, with significant revenue growth and high operating margins distinguishing it from the businesses of the dot-com era that saw similar valuations with equally low profits. Competition is heating up, though. Nvidia's long-term profitability depends on how well hyperscalers and other users of AI infrastructure will be able to generate adequate returns on their investments in AI.

These hyperscalers include Microsoft, Amazon, and Google. The big tech companies provide essential infrastructure, platforms, and services for creating and deploying Generative AI. They're dominating this evolving market with heavy investments in data centres and computing power. However, analysts question whether they can generate the revenues required to justify the enormous capital spent on AI infrastructure. Estimates suggest a possible $500 billion annual revenue shortfall. Others have now claimed that "AI is a bubble," mostly influenced by the high investment from hyperscalers, raising questions over the sustainability of those investments. This is not an unusual phenomenon for emerging technologies, where one sees some sort of disconnect between infrastructure buildout and application value. All told, in the past year, Amazon, Google, Microsoft, and Meta have spent a staggering $177 billion on capital expenditures. These AI investments are, to a large degree, demand-led, but there is a distinct lack of clarity on the return on investment. The hyperscalers are making massive bets on future demand for compute power, including heavy investments in the construction of data centers and the expansion of GPU capacity. They are investing around half of their capital expenditures in securing the key natural resources they need-real estate and power-that are required for future data center growth. In the words of Sundar Pichai, CEO of Google, "the risk of under-investing is higher than over-investing," aptly pointing to a balance that hyperscalers must tread in this landscape.

The majority of the value has been constrained in the infrastructure, most notably with Nvidia, and its lead in semiconductor production for AI applications. While revenue from AI applications is reportedly reaching an estimated value of $20 billion, it still represents less than a small fraction of investment in infrastructure. If AI were to be valued, it must be in terms of its ability to confront huge challenges in consumers and businesses.

In the semiconductor market, Nvidia leads with a high run rate of $105 billion AI revenue; TSMC accounts for approximately $10.4 billion AI revenue in 2024, and AMD garners roughly $4.5 billion in revenue within the same time period. Meanwhile, the estimate by memory providers SK Hynix, Samsung, and Micron is roughly generating approximately $16 billion in revenues at high-bandwidth memory. Deal activities in custom AI chips and networking semiconductor sales will see companies like Broadcom and Marvell as beneficiaries.

Data center growth is driven principally by capital spending from hyperscalers, and capacity is expected to double over the next four years. This creates great opportunities for data center developers like QTS and Vantage, and server manufacturers such as Dell and SMCI. Power demand, however, is a bottleneck to growth, with increasing costs in power auctions highlighting substantial challenges facing the infrastructure.

With hyperscalers continuing to ramp up their AI integration, significant cloud AI revenue is seen across the board. Microsoft Azure AI is reportedly at an annual run rate of roughly $5 billion, while Google and Amazon confirm "billions" in AI revenue, though neither company has released exact figures. Oracle has brought in about $1.3 billion in quarterly AI revenue from contracts, and GPU-focused cloud companies like Coreweave, Lambda Labs, and Crusoe are adding billions in revenue in their own right.

However, revenues from the application of AI are still relatively limited, between $5 billion and $10 billion. Besides, cost-saving impacts of AI are more considerable, especially due to labor replacement. Other notable players in the space include OpenAI, which is rumored to be hauling in $1.5 billion in API revenue, and Anthropic, which is projected to reach $600 million in revenue by 2024. Growth may well depend on new breakthroughs in agentic AI-large language models with memory and the ability to plan and utilize tools.

The current landscape shows that infrastructure buildout has outpaced application value, which presents short-term uncertainty. But this positions the industry for long-term growth, provided it can address high-impact problems and unlock unforeseen applications. History has repeatedly proven that one should not bet against technological progress; true value will be created with time, just as the internet was able to catalyze disruptive innovations – big tech trends tend to be overvalued in the short term and undervalued in the long term.

Besides, AI serves as a value enhancer. Companies with a lot of data can apply AI to advanced data analysis and achieve more profound competitive advantages. For example, healthcare will be able to make better diagnoses and accelerate drug discovery, and retailers and consumer packaged goods firms will be able to enhance marketing for personalized customer experiences.

Specific industries are set to reap the rewards of integrating generative AI into their operations. Software companies like Adobe can enhance their creative tools, while healthcare organizations are pioneering AI-driven solutions for drug discovery and personalized medicine. Financial institutions can optimize their operations using AI for risk assessment and fraud detection, while retail firms can personalize customer interactions and streamline supply chains.

Businesses with unique datasets and domain expertise, such as those in healthcare and finance, will build special AI models that outperform general-purpose models. Those with significant user bases can already use their data to train the models, giving them a considerable advantage.

As the demand for skilled labour evolves, skills related to AI development, data science, and more creative and cognitive fields will command an increasing premium. Workers who can adapt and build these in-demand skills will command higher wages and enjoy enhanced job security.

This means countries pioneering AI research and implementation in their administrations or respective fields will have conspicuous economic and geopolitical advantages. Given this fact, government policies that encourage innovation, investment, and labor force development within the AI sector are needed.

There is consensus among experts on specific industries most likely to be considerably affected by generative AI. According to McKinsey, it would be a major drug discovery and development breakthrough where pharmaceutical and life sciences could gain the most due to AI. This capability for generative design can make a difference in creating candidate molecules, which may revolutionize drug development processes to achieve faster discoveries and breakthroughs in medical science. Similarly, the finance & banking sector is consistently recognized as ripe for AI-driven transformation, pointing to a broad shift across many industries as generative AI becomes a fundamental component of future operations.

Generative AI is being seen as a force multiplier of productivity enhancement, reshaper of corporate strategies, and facilitator of completely new business models. This will continue to see accelerated adoption, as businesses can gain from improved workflows, increased customer engagement, and innovation in product offerings.

On economic drivers, generative AI excels in automation and efficiency, hence can automate 60-70% of employee time spent on tasks, thereby driving operational productivity. This also fosters innovation, enabling new product and service development, driving competitive differentiation, and opening new revenue streams. Furthermore, with the stagnation of the global workforce, productivity gains from AI can provide sustenance for economic growth and reinforce its importance in the modern economy.

The Winners and Others

Stocks are increasingly being influenced by the impact of generative AI, which has the potential to enhance a firm's free cash flow significantly. Technologies like ChatGPT reduce labour costs by automating specific tasks and boosting overall workforce efficiency and productivity.

In one such exercise, U.S.-listed companies were sorted into the degree their staff is exposed to generative AI, researchers from NBER found that "Top quintile exposure firms significantly outperforming with daily excess return 0.4% following ChatGPT shock", they added, much more than 100% annualized which hints to longer lived advantages beyond mere speculation over large values at stake for the value created with the help of AI. By March 2023, the cumulative return for these high-exposure firms had topped 9%, underlining the long-term benefits of adopting such technologies.

Typically, companies best positioned to take advantage of generative AI will be in industries that are intensive in information technology or R&D, where labour is usually higher-skilled and wages higher. Companies like Microsoft, Adobe, and Nvidia are exemplary examples of companies that integrate AI into their operations to improve their products and operational efficiency.

On the other hand, low-exposure firms are generally positioned in areas involving intensive labour, such as retail sectors such as Walmart and Target, restaurants and fast-food chains, including McDonald's, and old-line manufacturing, and they will see tiny relative advantages from these advancements in AI.

This will create some strong strategies for investors. One good approach could be long-short portfolio strategies: taking long positions in the stocks of highly exposed firms while selling the stocks of low-exposure counterparts. In addition, a company that actively applies AI to its operational structure might expect miraculous growth in workforce productivity and, subsequently, an increase in revenue.

In other words, as the business world changes with the emergence of generative AI, investors should look for firms that have shown huge exposure to such technologies. There is a large potential for gain with those companies that not only adopt such advancements but also integrate them into their long-term strategies.