Aug 25, 2025 Generative AI (part 2)– From Hype to Integration
*A follow up to our original commentary on Generative AI published in April 2024
The Shift from Hype to Reality
Artificial intelligence (AI) is no longer just a buzzword. In 2023, most companies were experimenting cautiously, running pilot programs or tasking innovation teams with “exploring possibilities.” Fast forward to today, and AI is increasingly present in daily life and business operations. It is writing marketing copy, analyzing financial statements, helping doctors draft clinical notes, and even serving as a digital assistant for students.
This shift from hype to reality has happened quickly, and it is only the beginning. The investment community is treating AI much like past industrial revolutions: a disruptive force that will create winners and losers, change how businesses operate, and reshape economic growth. For investors, in our opinion, understanding both the opportunities and the risks is essential.
The Economic Impact: Productivity Returns to Center Stage
Productivity growth in developed economies has been sluggish for nearly two decades. AI has the potential to change that. Recent estimates suggest that generative AI could add $2.6–$4.4 trillion annually to the global economy, possibly more as adoption scales. Goldman Sachs projects global GDP could rise by 7% over the next decade thanks to AI.
These gains come primarily from efficiency. AI doesn’t just replace tasks – it frees up employees to focus on higher-value work. A lawyer might spend less time drafting basic contracts and more time on complex negotiation. A banker can automate parts of loan processing and devote more attention to client relationships. A doctor can reduce the hours spent on paperwork and spend more time with patients.
It’s worth noting that AI’s contribution will not appear overnight in official GDP statistics. Companies must invest first – in hardware, in cloud capacity, in software, and in training. Global spending on AI is already approaching $200 billion per year, similar to the heavy infrastructure buildout seen in the early internet era. The payoff will likely come in the second half of the decade, once these investments are fully absorbed and productivity gains become widespread.
Technology Progress: From Tools to Co-Workers
Technology itself has taken major steps forward in the last 18 months.
- Multimodal AI enable models to handle different types of input at once – text, images, audio, and video. That means a single system could summarize a long report, analyze a chart, and describe a photograph all in one workflow.
- On-device AI is bringing models directly onto smartphones and laptops. This reduces reliance on the cloud, cuts costs, and enhances privacy. It also makes AI faster and more widely available.
- AI Agents represent the next frontier. These are systems that don’t just respond to prompts, but can plan, take multi-step actions, and make decisions to accomplish a goal. While still early, they hint at a future where AI feels less like a tool and more like a co-worker.
These advances are reshaping business software and daily life. Already, “copilot” features in office applications that can draft emails, prepare presentations, and summarize meetings. Customer service platforms can resolve issues faster, while programmers use AI to generate code. None of these systems are flawless, but the direction is clear: AI is moving rapidly from experimental to essential.
Regulation & Geopolitics: Guardrails & Competition
No technology this powerful develops without oversight. Governments are working quickly to set the rules, though approaches differ.
- Europe has passed the EU AI Act, the world’s first comprehensive AI law. It bans certain high-risk practices outright, and requires strict oversight for applications in healthcare, finance, and hiring. Enforcement begins in 2026, but companies are already preparing.
- The United States has not passed a single AI law but is using existing agencies and regulations. Guidance on privacy, consumer protection, and discrimination all apply to AI. Washington is also funding domestic semiconductor production and encouraging industry self-regulation.
- China is taking a different path, requiring strict content controls on what AI systems can generate, while at the same time fostering hundreds of homegrown AI models.
What it means for business and investors is that compliance is now a feature, not an afterthought. Providers that build transparency (clear labeling, audit trails), quality controls, and bias testing into products will be preferred vendors, especially in regulated industries. For portfolios, that favors firms with strong governance and clear documentation over quick-and-dirty deployers.
AI is also now a geopolitical tool. The U.S. and China are in a race for leadership, particularly over the semiconductors that power AI. Export restrictions, “friendshoring” of supply chains, and massive subsidies for domestic chip production are reshaping the global tech landscape. For investors, it means that national policy will continue to play a large role in determining which companies thrive.
Near- & Long-Term Outlook
In the near term (2025–2027), AI adoption will remain uneven. Some corporate projects will stall – often due to poor data quality or unclear return on investment – while others will generate clear savings. Think of this as the learning curve: companies are figuring out where AI fits and where it doesn’t.
By the late 2020s, however, adoption should accelerate. As businesses learn, infrastructure matures, and costs come down, we expect to see a broad wave of productivity gains. By 2030, AI will likely be embedded across industries in ways that feel routine – as natural as using email or spreadsheets today.
The long-term prize is significant: higher economic growth, new industries, and entirely new categories of work. Just as the internet created jobs that didn’t exist before – app developers, digital marketers, cybersecurity specialists – AI will create its own ecosystem of roles. The transition will require adaptation, but history suggests that overall employment and living standards will rise.
Sector-by-Sector: Where We See the Most Practical Uses
Healthcare. The near-term win is administrative relief: recording visits, drafting notes, pre-populating forms, and summarizing histories. That reduces burnout and frees clinicians for patient care. Over time, expect AI to support documentation, triage, and imaging workflows—always with clinician oversight. When investing, we look for software and services that fit within regulatory frameworks and prove time savings.
Financial services. Underwriting, fraud detection, compliance screening, claims processing, and client communications all may benefit from AI. The upside is faster decisions with better audit trails. The winners will pair models with clean data pipelines and strong governance. Expect a steady shift from “digital-first” to “AI-first” operations.
Industrials and logistics. This is less about chatbots and more about predictive maintenance, quality inspection, inventory optimization, and scheduling. Visual systems catch defects; forecasting reduces stockouts; routing saves fuel and time. Returns show up as higher asset utilization and better on-time performance.
Education and training. AI helps personalize materials and provides instant feedback, from K-12 to corporate learning. Adoption takes time due to budgets and procurement cycles, but the long-term addressable market is large and global.
Public sector. Permitting, benefits administration, and response centers can all be improved with AI summaries, document extraction, and self-service. Governments move slowly, but when they do, volumes are high and contracts are lengthy.
Portfolio Implications for Investors
What does all this mean for investors? We see opportunities at several levels:
- Infrastructure: Semiconductors, cloud providers, and data centers remain the foundation. Without them, AI cannot function.
- Enterprise Software: Productivity platforms, cybersecurity, and developer tools embedding AI features are seeing strong demand. Companies that combine AI with unique data sets have a defensible advantage.
- Services & Consulting: Every business wants AI, but few know how to implement it effectively. Firms that specialize in guiding adoption will remain in high demand.
The flip side is that not every company will succeed. Some will overpromise and underdeliver, just as many did during the dot-com boom. Others will fall behind by refusing to adapt. The key for investors is diversification across the AI ecosystem and alignment with firms that demonstrate real, measurable productivity gains.
Generative AI is not a passing fad. It is a structural force that will reshape global growth and business over the next decade. Technology is advancing rapidly, governments are setting guardrails, and businesses are investing heavily. The short term will include missteps and volatility, but the long-term opportunity remains compelling.
For investors, in our opinion, the message is clear: maintaining exposure to the AI theme is no longer optional. It should be a core part of long-term strategy, balanced with diversification and risk awareness.
-The Seventy2 Capital Team
Commentary and Research provided by:
Michael Levitsky, CFA®, CAIA® – Managing Director, Investment Strategy