Address
804, 805, 8th Floor, Tower B4, Spaze I Tech Park Sec-49, Sohna Road, Gurgaon-122018 , Haryana, India.
Download Alpha Portfolio App
Download Consolidated Portfolio App
Social Media
By Alpha Research
Jul 04, 2026
Three articles appeared in the global financial press in the last 10 days. Read together, they highlight where we are in the AI investment cycle, and where the danger and opportunity lie.
The first was an interview with renowned valuation expert Aswath Damodaran, NYU’s Professor of Finance, who questioned whether the massive AI infrastructure build-out is being financed sustainably. The second was Ruchir Sharma's argument that the real beneficiaries of AI may not be Silicon Valley's software giants at all, but rather the countries and companies building the physical infrastructure that powers the revolution. The third was perhaps the market's clearest signal yet. Nearly $2.2 trillion of market value disappeared from the Magnificent Seven in a single trading session, while semiconductor companies continued their remarkable ascent.
Damodaran argues that today's AI boom differs from the dot-com era in one crucial way: it is increasingly being financed with debt rather than equity. During the dot-com crash, shareholders bore most of the losses. Banks and the broader financial system remained largely insulated. Today's AI buildout, however, is being funded through private credit, leveraged infrastructure funds, and project finance for data centers. Goldman Sachs estimates hyperscalers could spend $5.3 trillion on AI infrastructure between 2025 and 2030, while firms like Apollo and Blackstone have already committed billions to finance AI expansion.
Damodaran's concern is not that AI will fail, but that if expected revenues do not materialize, debt-funded projects could create risks extending well beyond the technology sector. Unlike equity-funded bubbles, debt can transmit losses across the broader financial system. As he bluntly asked: "Who are these lunatics who are lending money to the data centers?"
While Damodaran focuses on where the risks lie, Ruchir Sharma asks where the real profits will be made. His answer is surprising. The biggest winners from the AI revolution may not be the American software giants, but the companies and countries that build the physical infrastructure powering AI. Taiwan, South Korea, and Japan are benefiting from their leadership in semiconductors, memory chips, and AI hardware. Countries such as Vietnam, Thailand, and Mexico are also gaining through electronics manufacturing, while China remains an important part of the global AI supply chain despite geopolitical tensions.
Sharma's key insight is simple: the most visible AI companies are not necessarily the ones capturing the greatest long-term value. Much of the economic gain is flowing to the "picks-and-shovels" providers that make the AI revolution possible. As Sharma puts it: "The boom is now so powerful and widespread that it is overwhelming all other drivers of returns—but rests on the narrowest base in decades."
The third article provides the market evidence supporting the earlier arguments. On 1 July 2026, the Magnificent Seven lost nearly $2.2 trillion in market value in a single day, while semiconductor stocks continued to surge. The Philadelphia Semiconductor Index has almost doubled in 2026, marking its strongest performance since the dot-com era. Companies supplying AI infrastructure have significantly outperformed: TSMC is up around 50%, ASML about 80%, while memory and storage companies such as Micron, Intel, Western Digital, and Seagate have posted exceptional gains.
Persistent memory shortages, expected to continue until 2028, are also pushing up hardware prices, with companies like Apple and Microsoft already passing higher costs on to consumers.
This is more than a temporary market rotation. It reflects a shift in investor focus, from companies promising future AI applications to those already generating profits by building the infrastructure that powers AI. In other words, the market is beginning to reward cash flows over narratives and physical enablers over AI hype.
History has always celebrated transformative technologies, be it railways, electricity, automobiles, or the internet. Artificial Intelligence now joins that list. But investors often confuse two entirely different questions.
“Will the technology change the world?”
“Will today's shareholders become wealthy?”
The answers are not always the same. Railways transformed commerce, yet thousands of railway investors still lost money. The internet reshaped every aspect of modern life. Yet Pets.com disappeared, while Amazon became one of history's greatest investments. Technology succeeds. Individual investments do not always follow. That distinction matters enormously today.
Artificial Intelligence is not simply another software cycle. Economists classify it as a General Purpose Technology (GPT), a rare class of innovation capable of reshaping almost every sector of the economy. Steam power, electricity, the internal combustion engine and the internet all belong to this category. Unlike ordinary innovations, GPTs do not generate productivity gains immediately. They require complementary investments in infrastructure, organisational redesign, skills, regulation and entirely new business models before their full economic potential is realised.
Economic historian Carlota Perez argues that every technological revolution unfolds in two distinct phases. The first is an installation period, characterised by heavy capital investment, speculative enthusiasm and rapid infrastructure build-out. Financial capital races ahead of productive capital as investors compete to fund the technologies that may dominate the next era. This phase is often accompanied by asset bubbles, because expectations rise faster than realized cash flows. Only later comes the deployment period, when the infrastructure is largely in place, adoption broadens, business models mature, and productivity gains diffuse across the wider economy.
AI today appears to be firmly in this installation phase. The hundreds of billions invested annually in data centers, advanced semiconductors, networking equipment, and electricity generation are not merely technology expenditures; they are the physical foundations on which future AI applications will depend. Much as factories had to be redesigned before electricity transformed manufacturing, businesses will need years of complementary investment before AI delivers its full economic impact. Today's spending may therefore appear excessive, but history suggests that infrastructure almost always precedes widespread productivity.
History also suggests that technological revolutions rarely create value evenly across industries or over time. Joseph Schumpeter's theory of creative destruction reminds us that new technologies continuously reallocate profits from incumbent firms to new business models. Yet identifying the winning technology is only part of the investment challenge. Investors must also identify where along the value chain economic profits are likely to accumulate as adoption evolves.
That evolution can be understood through the Bass Diffusion Model, one of the most influential frameworks in technology adoption research since 1969. The model explains how new technologies spread through two forces: innovation, driven by external influences such as marketing and early adopters, and imitation, driven by network effects that make adoption self-reinforcing.
Current research suggests that AI exhibits an exceptionally strong imitation effect compared with previous technological revolutions. In practical terms, adoption is accelerating because firms increasingly adopt AI not merely for competitive advantage but because competitors are already doing so. This creates a powerful positive feedback loop that favors rapid infrastructure expansion long before end-user monetization fully develops.
At this stage of the diffusion curve, investors naturally reward the companies removing adoption bottlenecks—semiconductor manufacturers, memory suppliers, cloud infrastructure providers, networking companies, and power infrastructure. As adoption matures, however, the sources of value creation are likely to shift.
Clayton Christensen's work on value migration demonstrates that profits often migrate over time from one layer of a technology stack to another as industries mature. Early leaders in infrastructure frequently earn attractive returns while capacity remains scarce, but competitive forces eventually compress margins.
Later, businesses that control customer relationships, ecosystems, and applications often capture a larger share of long-term economic value. This evolution helps explain why today's market leadership should not automatically be extrapolated indefinitely. Every technological revolution eventually shifts from building the infrastructure to monetizing it.
The biggest lesson for investors is that AI is no longer a single investment theme. The recent sell-off in the Magnificent Seven does not signal the end of AI investing. Instead, it reflects a rotation towards companies building the infrastructure that powers AI, particularly semiconductor and hardware firms. The "picks-and-shovels" strategy is now mainstream.
However, this does not necessarily mean infrastructure will remain the dominant source of returns throughout the AI cycle. History suggests that investment leadership often rotates as technologies mature. Today's hardware leaders benefit from capacity shortages, pricing power, and enormous capital expenditure by hyperscalers. Over time, however, increased competition may compress infrastructure margins as supply catches up with demand.
Equally, falling computing costs could lower barriers for software developers, allowing application-layer companies to capture a greater share of long-term value through network effects, customer relationships, and ecosystem lock-in. Governments may also intervene through regulation of energy usage, semiconductor exports, competition policy, or data governance, potentially altering returns across different parts of the AI value chain.
Similarly, rising memory costs are creating both winners and losers. Higher chip and memory prices are squeezing margins for AI application companies while boosting revenues for hardware manufacturers. The same trend that hurts one part of the AI ecosystem benefits another. The implication for investors is not that one segment will permanently outperform another, but that the location of economic profits is likely to evolve as the technology itself matures.
Recognising where we stand in the technology cycle is only the first step. Investors must still answer the more difficult question: how should AI businesses actually be valued? Traditional valuation models remain essential, but AI introduces new variables that conventional frameworks may not adequately capture. Competitive advantages increasingly arise from data, computing infrastructure, ecosystem effects, proprietary models, and access to scarce talent.
At the same time, enormous upfront investment, uncertain monetisation paths, and rapidly changing competitive dynamics make forecasting future cash flows unusually difficult. The next challenge for investors, therefore, is not determining whether AI will reshape the economy; it almost certainly will, but identifying which firms are likely to convert technological leadership into durable economic profits and whether current market prices already assume that success.
My dear readers,
The biggest winners of the internet era were not simply those who believed in the internet; they were those who identified the right businesses and bought them at the right price. The AI revolution is likely to reward the same discipline.
PS: During a gold rush, selling shovels often proves more profitable than searching for gold.
Best Regards
Dr. Tarunika Jain Agrawal Ph.D.
Alpha Research
The information contained in this communication is intended solely for the use of the addressee and others authorized to receive it. It may contain confidential/legally privileged information. If you are not the intended recipient you are hereby notified that any disclosure, copying, distribution or taking any action in reliance on the contents of this information is strictly prohibited and may be unlawful. If you have received this communication in error, please notify us immediately by phone, fax or email and delete it from your system. Alpha Capital is neither liable for the proper and complete transmission of the information contained in this communication nor for any delay in its receipt.
