The proximate trigger, as reported at the time, was the conflict involving Iran, Israel and the United States, which pushed crude oil prices sharply higher and sent India’s VIX volatility gauge — the market’s so-called fear index — from below 12 to a peak of 28.9 by 30 March, before it eventually retreated. According to Outlook Business, foreign portfolio investors swung from net inflows of roughly $2.5 billion in February to net outflows of nearly $6 billion by mid-March, and new investor registrations on Indian exchanges fell 24.5% month-on-month — the steepest such drop of the financial year, suggesting that while mutual funds kept buying, the flow of fresh retail money into direct equity slowed sharply during the worst of the panic.

What made the year noteworthy was not the fall itself but who did the buying while foreign investors did the selling. Foreign portfolio investors (FPIs) sold Indian equities worth Rs 1.81 lakh crore over the fiscal. Domestic institutional investors (DIIs), led overwhelmingly by mutual funds, absorbed that pressure and then some, investing a record Rs 8.09 lakh crore — comfortably the highest figure in at least five years, and roughly four times the previous year’s DII total of Rs 5.05 lakh crore.

Mutual funds’ own contribution to that domestic buying was substantial: net equity inflows into mutual funds stood at Rs 3.47 lakh crore for the fiscal. March 2026 marked the 61st consecutive month in which mutual funds recorded positive net equity inflows — a streak that has now run through multiple market corrections and several bouts of foreign capital flight without breaking.

The mechanics of this shift are visible month by month. Foreign investors turned net sellers in July, August, September, October, December and, most emphatically, March 2026, when they pulled out Rs 1,17,775 crore in a single month — by far the largest monthly outflow of the year. Domestic institutions were net buyers in every single month bar none, and in March, as foreign money fled, DII buying actually accelerated to Rs 1,42,960 crore, comfortably its highest monthly figure of the fiscal.

This is not simply a story of resilience; it is a structural change in who sets the marginal price of Indian equities. Five years ago, in fiscal 2022, DIIs invested Rs 2,20,193 crore against FPI outflows of Rs 1,40,008 crore — already domestic-dominated, but at roughly a third of this year’s scale on the DII side. For retail investors, the practical implication is that systematic domestic flows now provide a buffer against foreign-flow volatility that did not exist at anything like this scale even three years ago. Whether that buffer holds in a genuinely sustained downturn, rather than a single bad month, remains untested. But fiscal 2026 was as close to a live-fire drill as Indian markets have had in years, and the buffer held.

A layered view of the cycle

The table below summarises how each layer of the AI value chain behaves during the boom and bust phases, and where long-term winners are most likely to emerge.

A useful way to analyse any AI investment is to separate the value chain into layers, each with its own boom and bust dynamics.

Infrastructure -- chips, data centres, networking and power -- sees heavy spending during the boom and, in a bust, excess capacity and falling pricing power; the eventual winners are the lowest-cost, most efficient providers. Platforms and foundation models attract enormous investment early on, before consolidating around the handful that prove commercially viable, favouring firms with network effects, proprietary data and developer ecosystems. Applications multiply by the thousand in the boom phase and fail in droves once the bust exposes a lack of differentiation; survivors tend to be businesses solving real problems with recurring revenue. Enterprise adoption follows a similar pattern of wide experimentation followed by abandoned pilots, rewarding firms that actually weave AI into existing workflows. End-users, finally, move from sky-high expectations to the sober realisation that AI is a tool rather than a miracle -- benefiting, in the end, from lower costs and higher productivity.

Five questions help sort the durable from the fashionable. Where does a company sit in the value chain -- chipmaker, cloud provider, data-centre operator, software firm or AI-enabled business -- since each faces different economics? Is demand structural or merely cyclical: data generation and computing needs look structural, whereas buying GPUs ahead of demand may not be. Who captures the economics, given that technological adoption does not always reward the technology’s creators -- electricity generated enormous value even as utilities earned only regulated returns, and the internet enriched users more than most telecoms firms? Does AI strengthen or weaken a firm’s competitive position, boosting those with proprietary data, distribution and loyal customers while eroding businesses whose knowledge can be commoditised? And what happens if capital becomes expensive, since booms typically end when financing conditions tighten and cash flow, not funding rounds, decides who survives?

Historically, this cycle unfolds in four phases: an infrastructure boom (benefiting chipmakers, equipment suppliers and data-centre operators, at the risk of overbuilding); an application explosion (thousands of start-ups, high valuations, little differentiation); a shake-out (falling valuations, consolidation, bankruptcies -- the phase in which the strongest companies typically emerge); and finally a productivity era, in which AI becomes invisible infrastructure and the biggest beneficiaries may not even call themselves AI companies. Railways, electricity, the internet and smartphones all followed variants of this pattern; in each case the lasting wealth accrued not to the miners but to the suppliers of picks and shovels, and to the “towns” that grew up around the goldfields using the new technology to work more efficiently.

Why India will not simply mirror America

This framework travels well. But applying it to India requires adjusting for three features of the Indian economy that have no close American equivalent.

Retail investors could cushion market volatility. India’s equity markets have changed dramatically over the past decade. Millions of households now invest regularly through Systematic Investment Plans, making them an increasingly important source of market liquidity. Annual SIP contributions have risen from under ₹44,000 crore in FY2016–17 to nearly ₹3.5 lakh crore in FY2025–26, and these inflows held up through both the pandemic and subsequent corrections (AMFI, 2026) -- evidence that retail investors have become long-term participants rather than short-term traders. At the same time, direct retail trading now accounts for a smaller share of daily turnover than it once did, as more investors favour mutual funds over individual shares and institutional and algorithmic trading grows (Whalesbook, 2026). The implication is nuanced: an AI-driven correction could still make day-to-day trading choppier, but steady SIP inflows may continue to underpin equity markets, limiting the depth of any prolonged sell-off -- a shock absorber American markets largely lack.

The Reserve Bank has less room to manoeuvre. Where the Federal Reserve can often cut rates aggressively in a slowdown, the Reserve Bank of India must also defend exchange-rate stability and guard against imported inflation. A global AI-driven slowdown could trigger capital outflows from emerging markets and weaken the rupee; a weaker currency would raise the cost of imported oil, electronics and industrial inputs, pushing inflation up even as growth slows. In that scenario the RBI might struggle to cut rates at all, meaning an AI bust could tighten financial conditions in India precisely when the economy needs support -- a far more constrained position than that of advanced-economy central banks.

India’s biggest exposure is as an exporter of talent. For America, AI is chiefly about lifting domestic productivity. For India, it also threatens one of the country’s largest export industries. The IT services and business-process-management sectors employ millions of professionals doing precisely the kind of routine knowledge work -- documentation, coding support, customer service, research assistance, report preparation -- that generative AI is designed to automate. The evidence so far is mixed: NASSCOM reckons AI could add nearly $621bn to India’s economy over time and expects the technology to transform jobs rather than eliminate them, and surveys suggest most employers are redesigning roles rather than cutting headcount (NASSCOM, 2026b; Enterprise Times, 2026), while the RBI has noted that IT services exports remain resilient despite growing concern about AI (New Indian Express, 2026). The deeper risk may be slower to show. If companies keep hiring fewer graduates while leaning more on AI, India’s talent pipeline could thin out gradually, since junior staff typically learn by doing routine work before taking on more complex responsibilities. Shrinking entry-level opportunities today could mean a shortage of experienced professionals a decade from now -- a lag effect that would not show up in this year’s export figures.

What each scenario would mean for India

A boom would lift productivity across manufacturing and services, expand Global Capability Centres, draw fresh investment into the digital economy and push technology firms further up the value chain, boosting demand for AI specialists, cybersecurity professionals and cloud engineers. Consumers would gain from better AI-powered healthcare, education, financial services and public administration. But the gains would not be shared evenly: workers with advanced digital skills would capture a disproportionate share, while routine knowledge jobs would come under mounting pressure.

A bust would look different from a Wall Street correction. Investment in data centres and AI infrastructure could slow sharply, straining the banks that finance such projects. States with already-stretched finances would find it harder to sustain employment if tax revenues weakened. The RBI would have limited room to help if capital outflows hit the rupee and stoked imported inflation. Most consequentially, India’s IT services industry could see slower growth if global clients cut spending on AI and digital transformation. Where American markets might absorb much of the adjustment through asset prices, India’s risks would be spread across banks, state governments and the labour market itself.

The bottom line

The boom-or-bust framework -- infrastructure, platforms, applications, enterprise adoption, end-users; structural versus cyclical demand; who captures the economics -- remains a useful lens for thinking about AI’s future anywhere. But India’s story will not simply mirror America’s. A young workforce, a large informal economy, a bank-led financial system and a globally competitive IT services industry give the country its own distinct set of opportunities and risks. A boom could accelerate productivity, strengthen exports and support long-term growth; a bust would expose structural weaknesses in employment, state finances and bank lending rather than trigger the market dynamics familiar from developed economies.

For investors and policymakers, the lesson is straightforward: analyse AI in India through India’s own economic realities, not assumptions borrowed from elsewhere. The technology may be global. Its consequences will remain deeply local.

References

• Association of Mutual Funds in India. (2026). SIP contribution data.

• Business Today. (2026). IT industry likely to hire 135,000 employees in FY26, projects Nasscom.

• Enterprise Times. (2026). Top skills employers prioritise in 2026 as AI redefines jobs across industries.

• Ministry of Finance, Government of India. (2026). Union Budget 2026–27.

• NASSCOM. (2026a). India’s services sector and the AI opportunity.

• NASSCOM. (2026b). India’s workforce transformation opportunity in the AI era.

• New Indian Express. (2026). Reserve Bank of India says IT services exports resilient despite growing AI concerns.

• Press Information Bureau. (2025). Periodic Labour Force Survey Annual Report.

• PRS Legislative Research. (2025). State of State Finances.

• Rees, D. (2025). AI economic scenarios: Revolutionary growth, or recessionary bubble?

Read more

Read more

That is the deeper implication emerging from J.P. Morgan’s recent thematic report, In the Garden of Good and Evil, which argues that frontier AI models are reshaping cybersecurity from a persistent operational challenge into a systemic infrastructure risk. The report’s core thesis is not merely that cyberattacks will increase. It is that AI is compressing the time between vulnerability discovery and exploitation so rapidly that much of the world’s digital and industrial infrastructure may no longer be defensible using existing methods.

Read more

That is the uncomfortable thesis advanced by Jessica Riedl, a long-time budget expert who has spent more than two decades moving through conservative think tanks, Capitol Hill and policy circles in Washington. In a wide-ranging conversation on Freakonomics Radio with Stephen Dubner, Riedl argued that America’s fiscal crisis is no longer a distant risk. It is already embedded in the country’s politics, tax system and entitlement structure.

Her central claim is blunt: almost everything Americans believe about taxes is wrong.

The politics of the “free lunch”

Riedl’s critique spares neither Republicans nor Democrats. Conservatives, she says, exaggerate the economic power of tax cuts. Liberals exaggerate how much money can be extracted from corporations and the wealthy.

Republicans have long argued that lower taxes stimulate enough economic growth to offset lost revenue. This is the famous “tax cuts pay for themselves” argument. Riedl dismisses it as fantasy. Tax cuts may generate some additional growth, she says, but historically they do not come close to fully financing themselves.

Another conservative belief is what economists call “starve the beast”: cut taxes, deprive government of revenue and politicians will be forced to cut spending. Yet Riedl notes that the opposite has often happened in America. Taxes fall, spending rises and deficits expand.

Democrats, meanwhile, tell a different but equally convenient story. According to this narrative, America’s deficits exist because the rich and corporations are undertaxed. Raise taxes on billionaires, and the fiscal problem disappears.

Riedl argues that the numbers do not support this either.

Even confiscatory taxes on the ultra-rich would barely dent America’s long-term deficit problem. Seizing all billionaire wealth, she argues, would fund only a fraction of federal spending, and only once. Meanwhile, America already operates one of the most progressive tax systems in the developed world, particularly in income taxation.

Her criticism is less ideological than arithmetic. The scale of future spending obligations is simply too large.

The debt machine

The figures are staggering. America’s national debt now exceeds $39trn, roughly 124% of GDP. Interest payments alone are approaching $1trn annually and are projected to climb sharply over the coming decades.

Riedl believes the fundamental driver is not taxation but spending.

Since 2000, federal spending has expanded dramatically, fuelled by wars, stimulus programmes, healthcare obligations and demographic ageing. Politicians from both parties, she argues, abandoned the fiscal restraint that briefly emerged in the 1990s when balanced budgets became politically fashionable.

Today, neither side campaigns seriously on discipline. Republicans favour tax cuts without corresponding spending reductions. Democrats support spending expansions without matching tax increases. Both postpone the consequences.

Riedl describes Washington as trapped in a dangerous equilibrium: politicians know the system is unsustainable but believe voters will punish anyone who speaks honestly about sacrifice.

The result is a bipartisan addiction to deficits.

The third rail: Social Security and Medicare

The interview becomes most politically explosive when the discussion turns to entitlement reform.

In American politics, few issues are more sensitive than Social Security Administration and Centers for Medicare & Medicaid Services. Yet Riedl argues that meaningful fiscal reform is impossible without changes to both.

Contrary to popular belief, Social Security and Medicare are not fully self-financing. Current workers fund current retirees. As populations age and birth rates decline, the worker-to-retiree ratio deteriorates.

Riedl estimates that Social Security and Medicare together face a long-term financing shortfall of roughly $124trn over 30 years.

Her proposed remedies are politically hazardous but economically conventional: gradually raise eligibility ages, increase taxes and reduce benefits for wealthier retirees through means-testing.

That final proposal is especially controversial. Millionaires, she argues, continue receiving generous Social Security benefits despite not needing them financially. Yet even discussing reductions for affluent retirees can trigger fierce political backlash.

The paradox is striking. Privately, many lawmakers acknowledge the problem. Publicly, few dare address it.

America versus Europe

One of Riedl’s more provocative claims concerns middle-class taxation.

She argues that Americans, especially middle-income earners, are taxed relatively lightly compared with Europeans. In much of Europe, governments finance extensive welfare states through broad-based consumption taxes such as VATs, along with heavier payroll taxes.

In the United States, by contrast, the federal income-tax burden is concentrated heavily among top earners. According to the data she cites, the top 20% of earners account for the overwhelming majority of federal income-tax revenue.

This distinction matters because many Americans admire European-style public services while opposing the tax structure that funds them.

Riedl’s argument is not necessarily that America should adopt Europe’s system. Rather, she believes fiscal debates should at least acknowledge the trade-offs honestly.

Large welfare states require large tax bases. Taxing only the rich is insufficient.

Trump, Biden and fiscal continuity

Although American politics often frames elections as radical breaks from the past, Riedl sees more continuity than difference.

She criticises Joe Biden for large spending increases, expansive stimulus measures and policies that contributed to inflationary pressures. But she is equally critical of Donald Trump, particularly his combination of tax cuts, tariff policies and reluctance to touch entitlement spending.

Her criticism of Trump’s second administration is especially severe. Programmes such as DOGE, the now-defunct Department of Government Efficiency initiative associated with Elon Musk, focused heavily on symbolic culture-war targets while leaving the largest spending categories untouched.

For Riedl, this reflects a broader truth about modern fiscal politics: headline-grabbing gestures are easier than structural reform.

The deeper problem

What makes Riedl unusual in Washington is not merely her fiscal conservatism. It is her insistence on criticising both tribes equally.

That has come at a cost. She jokes in the interview that she has few friends in Washington because her work involves “sharing uncomfortable truths”. Yet that independence also explains why her analysis attracts attention across ideological lines.

Her broader warning extends beyond taxes. America’s political system increasingly rewards emotional narratives over numerical reality. Tax myths persist because they tell voters what they want to hear.

Conservatives are told tax cuts are painless. Progressives are told billionaires can finance everything. Both stories avoid sacrifice.

But debt compounds regardless of ideology.

Riedl ultimately believes the mathematics will prevail. The only uncertainty is whether reform happens gradually through political compromise, or suddenly through crisis.

For now, Washington appears content to delay the reckoning.

The concept of capital cycles is well documented. According to Boston Consulting Group, industries experiencing rapid investment frequently suffer from overcapacity, which compresses returns over time.

Historical episodes reinforce this pattern. The railway boom of the 19th century and the telecom expansion of the late 1990s both involved heavy infrastructure investment followed by periods of financial distress. A report by International Monetary Fund highlights how credit expansion and capital surges often precede overinvestment cycles.

The dynamics are consistent. When a technology is widely expected to reshape the economy, firms rush to build capacity. Capital becomes abundant, competition intensifies and supply eventually exceeds demand. Prices fall, and returns on invested capital decline.

Yet the long-term beneficiaries are often different from the initial investors. The infrastructure created during these booms, whether rail networks, fibre-optic cables or cloud computing capacity, enables subsequent waves of innovation at lower cost.

The current investment cycle in advanced technologies shows similar characteristics: rapid capital expenditure, intense competition and uncertain end-demand visibility. While the eventual economic impact may be substantial, the distribution of financial gains remains unclear.

Investors therefore face a familiar dilemma: distinguishing between transformative technologies and attractive investments. History suggests the two do not always coincide.

Source: BCG; IMF

https://www.bcg.com/publications

https://www.imf.org/en/Publications

Read more

The Case for Humility in Investing

Markets often reward confidence during stable periods. However, when structural changes reshape the economy, excessive conviction can become a liability. The current environment illustrates this clearly. Artificial intelligence is advancing rapidly, geopolitical conflicts are reshaping global supply chains, and economic policy responses remain uncertain. These forces make forecasting difficult.

Rather than attempting to predict every outcome, investors can manage uncertainty through diversification and process discipline. For example, if an investor previously held a concentrated portfolio of 15–20 stocks based on high conviction, expanding the portfolio to 30 holdings can reduce the impact of unforeseen disruptions. This approach recognises that the future is harder to forecast and spreads risk across more opportunities.

Read more

For decades, investors have relied on valuation ratios such as the price-to-earnings (P/E) ratio or earnings yield to judge whether stock markets are overvalued. Yet these metrics have remained persistently elevated relative to historical norms since the late 1990s. A recent study titled A Macroeconomic Perspective on Stock Market Valuation Ratios examines this puzzle through a macroeconomic lens and suggests that structural changes in the economy may explain why these ratios have stayed high for so long.

Read more

The rise of artificial intelligence has been the dominant investment narrative since late 2022. The emergence of generative AI triggered a wave of capital spending, corporate experimentation and investor excitement. Equity markets followed suit. A small group of technology giants, often labelled the “Magnificent Seven”, generated a disproportionate share of global market gains. By the end of 2025 these companies accounted for roughly 35% of the S&P 500’s market capitalisation while explaining more than half of its cumulative returns over the previous three years.

Such concentration naturally invites comparisons with past bubbles. In earlier episodes; from Japan’s asset boom in the late 1980s to the dot-com mania of the late 1990s, a narrow set of stocks captured investors’ imagination and dominated market performance. When sentiment turned, the results were painful. But the current situation differs in several important respects.

For one thing, the companies leading today’s rally are not speculative start-ups. They are profitable firms with formidable cash flows, dominant market positions and large customer bases. Their valuations, while elevated, remain below the extremes seen during earlier bubbles. Forward price-to-earnings multiples for the technology leaders are significantly lower than those reached during the dot-com era, suggesting that expectations may be optimistic but not entirely detached from reality.

Moreover, the broader market has begun to participate in the rally. During 2025 the S&P 500’s valuation expanded not only because technology stocks rose but also because other sectors started catching up. This broadening participation hints at a possible shift in market leadership rather than an imminent collapse. Investors who once felt compelled to own only the dominant technology names are gradually rediscovering opportunities elsewhere.

Nevertheless, history offers a cautionary reminder that even high-quality companies can become poor investments if purchased at excessive prices. A useful comparison lies in the “Nifty Fifty” episode of the late 1960s and early 1970s. Back then a group of blue-chip American companies, including Coca-Cola, IBM and Procter & Gamble, was widely regarded as so reliable that investors believed they could be bought at almost any price.

The businesses themselves largely justified their reputations. Many continued to grow and dominate their industries for decades. Yet investors who bought these stocks at peak valuations often endured years of disappointing returns as prices eventually adjusted. The lesson is simple but often forgotten: exceptional companies do not guarantee exceptional investment outcomes when expectations become too exuberant.

If the present market resembles the Nifty Fifty period, the next phase may involve rotation rather than collapse. Leadership in financial markets rarely remains fixed. After the technology bubble burst in the early 2000s, commodities, emerging markets and value stocks enjoyed a long period of outperformance. In the inflationary environment of the 1970s, real assets and cyclical industries proved more resilient than the growth stocks that had previously dominated.

A similar shift could occur in the years ahead. Cyclical sectors such as industrials or energy might benefit from economic reacceleration, while healthcare and other defensive industries could attract investors seeking stability. Emerging markets and non-American equities may also draw attention if the dominance of US technology companies begins to fade.

Yet none of this implies that the artificial-intelligence story is finished. On the contrary, the technology is likely to reshape many industries over the coming decade. The question is not whether AI will transform the economy but which companies will ultimately capture the profits. History suggests that the initial pioneers are not always the biggest long-term beneficiaries. Often the greatest gains accrue to firms that adopt the technology effectively rather than those that first popularise it.

Several risks could complicate the outlook. A slowdown in global growth would threaten earnings expectations, particularly for companies priced for rapid expansion. Inflation could also re-emerge as a destabilising force, especially if geopolitical tensions disrupt trade or energy markets. And the enormous investment currently flowing into AI infrastructure; data centres, chips and cloud systems, raises the possibility that capacity could outpace demand.

Another overlooked constraint lies in the real economy. The energy required to power vast networks of AI data centres is growing rapidly. If electricity supply fails to keep pace with demand, it could reshape the economics of the industry and shift investor attention toward utilities, energy producers and infrastructure providers.

Faced with such uncertainties, the most sensible strategy may be neither to chase the dominant technology stocks nor to abandon equities altogether. Instead investors may find greater resilience in diversification. Holding exposure to AI leaders while gradually expanding into undervalued sectors and regions can create portfolios that are better prepared for a shift in market leadership.

Markets at all-time highs often tempt investors to predict dramatic turning points. Yet history suggests that the more common outcome is gradual change rather than sudden collapse. The current environment may indeed display some characteristics of a bubble, lofty expectations, crowded trades and intense narrative enthusiasm. But it also contains the seeds of a broader market.

In other words, the defining feature of the coming years may not be the bursting of a bubble but the spreading of opportunity. For investors willing to look beyond the most celebrated technology names, global equities may offer what the Schroders report calls “something for everyone.”

Understanding What AI Actually Is

One of the first insights highlighted in the memo is a common misconception about AI. Many people assume AI functions like a search engine that retrieves information. In reality, modern AI systems synthesize information and reason from patterns within large datasets.

AI models go through two broad phases. The first is training, where the model processes vast volumes of text and learns reasoning patterns, argument structures, and conceptual relationships. This phase is less about memorizing information and more about learning how ideas connect and how reasoning works. The second phase is inference, when the trained model applies those capabilities to solve tasks presented through user prompts.

This distinction highlights an important lesson: the effectiveness of AI often depends on the quality of prompts provided by users. Poor instructions lead to mediocre outputs, while carefully designed prompts can unlock sophisticated capabilities.

The Debate: Can AI Actually Think?

A major intellectual question surrounding AI is whether it truly “thinks” or simply rearranges patterns from its training data. Critics argue that AI lacks genuine understanding and only performs advanced pattern recognition. In this view, it recombines existing ideas but cannot generate truly original thought.

However, proponents counter that human thinking itself relies on combining previously learned ideas. Investors, scientists, and thinkers often synthesize knowledge from multiple sources to produce new insights. If AI can perform a similar synthesis at scale, the distinction between pattern recognition and genuine reasoning may become less meaningful from an economic standpoint.

The practical question therefore becomes less philosophical and more economic: if AI can produce reliable analytical output comparable to a skilled professional, organizations will adopt it regardless of whether the system “understands” in a human sense.

The Extraordinary Speed of AI Progress

One of the most striking points in the memo is the pace of development. Historically, transformative technologies took decades to become widely adopted. For instance, the first computer appeared in the mid-1940s, but personal computers only became widespread in the 1980s.

AI has followed a dramatically faster trajectory. In less than two years, generative AI tools have reached hundreds of millions of users and are already integrated into the workflows of a majority of companies.

This speed is significant because technological adoption typically occurs gradually, allowing industries and labor markets time to adjust. AI, by contrast, may change economic structures faster than society can respond.

From Productivity Tool to Autonomous Agent

Another key insight is the evolution of AI capabilities across three stages.

1. Chat AI – systems that answer questions and provide information.

2. Tool-using AI – systems that perform tasks such as analysis or data processing when instructed.

3. Autonomous agents – systems that can take a goal, plan the steps required, execute tasks, and deliver a finished result.

The transition from tool-using AI to autonomous agents is particularly important. At this stage, AI is no longer just assisting workers; it may replace entire categories of structured tasks.

Implications for the Labor Market

The memo raises significant concerns about the future of employment. If AI can perform structured analytical work—from coding to financial analysis — it could substitute for a large share of knowledge workers.

Historically, technological advances eliminated some jobs but created new ones. Optimists believe the same pattern will repeat with AI. However, the speed and breadth of AI adoption may challenge this assumption. Unlike previous technologies that replaced manual labor gradually, AI has the potential to affect a wide range of cognitive professions simultaneously.

Implications for Investing

Marks also examines how AI could transform the investment profession. AI systems can process enormous amounts of data, recognize historical patterns, and remain free from emotional biases such as fear or greed. These qualities align closely with the attributes of successful investors.

However, the memo also notes that investing requires judgment in areas where historical data is limited. Evaluating new industries, management quality, and emerging trends often involves intuition and qualitative reasoning — areas where humans may still hold an advantage.

As a result, AI may raise the bar for investors. Routine analysis will increasingly be automated, while human professionals will need to focus on interpretation, strategy, and qualitative judgment.

Is AI a Bubble?

Despite the excitement around AI, the memo concludes that the technology itself is unquestionably real and transformative. The uncertainty lies not in the technology but in the valuation of AI-related investments.

History shows that major technological revolutions often involve excessive investment and speculative bubbles. Infrastructure spending may overshoot actual demand, and many companies will fail even if the technology ultimately reshapes the economy.

Key Lessons from the Memo

Several practical lessons emerge from the discussion:

1. AI is fundamentally a reasoning system, not just a search tool.

2. The speed of AI development is historically unprecedented.

3. Autonomous AI agents could replace structured knowledge work.

4. Human judgment will remain important in uncertain or novel situations.

5. Investors should balance optimism about AI with caution about valuations.

Conclusion

The overarching message is nuanced. Artificial intelligence is neither a passing fad nor a guaranteed investment boom. It represents a genuine technological transformation whose full implications remain uncertain.

For professionals and investors alike, the most prudent approach is neither blind enthusiasm nor complete skepticism. Instead, a balanced position — recognizing AI’s potential while remaining selective and disciplined — appears to be the most rational strategy as the technology continues to evolve.

Read more

Read more

Markets reward process far more often than prediction. A clear framework helps investors avoid reacting to noise, narratives, and short-term excitement.

This note has three parts.

Part 1 revisits core principles that matter most when markets become less forgiving.
Part 2 outlines our current perspective across asset classes, grounded in long-term data rather than short-term views.
Part 3 focuses on some scenario thinking for 2026, accepting uncertainty as a permanent condition, not an exception.

Read more

Fast forward to 2025. By headlines alone, this should’ve been a terrible year for stocks: tariffs, inflation anxiety, political noise, credit downgrades. Yet U.S. equities pulled off a sharp comeback. By late September, the S&P 500 was up nearly 14%, the Nasdaq over 17%. On the surface, everything looks fine.

Dig underneath, and the story changes. Market performance has been wildly uneven. Technology and communication services—carried largely by Alphabet and Meta—did the heavy lifting. Defensive sectors like healthcare, consumer staples, real estate, and energy went nowhere. This wasn’t a broad-based rally. It was a concentrated one.

Then there’s the elephant in the room: the “Magnificent Seven.” Apple, Nvidia, Microsoft, Alphabet, Amazon, Meta, and Tesla continue to dominate market value. Despite a rough first quarter, they rebounded hard and now account for over half of total market gains in 2025. If you’re betting on a market collapse without these names cracking first, you’re betting against the data.

Interestingly, smaller-cap stocks quietly outperformed large caps this year, mostly in the third quarter. Value stocks showed brief life early on, then faded again. Momentum— stocks that were already winning — regained control. Same movie, new act.

What didn’t move much? Interest rates. Despite tariff shocks and a U.S. credit downgrade, Treasury yields barely budged. Corporate credit spreads spiked briefly, then settled. The bond market, usually the nervous one, stayed calm. That alone explains why stocks didn’t panic.

Subscribe now

Globally, non-U.S. markets actually outperformed the S&P 500 in 2025, helped by a weaker dollar. Europe, China, Latin America—all did better. India, notably, lagged badly, likely correcting after entering the year as one of the most expensive markets worldwide.

So, are U.S. stocks overpriced? By almost every metric, yes. Prices are high relative to history. Price-to-earnings ratios are stretched. Earnings yields barely beat Treasury yields. Even when you account for growth using intrinsic value models, stocks still look about 10–15% above fair value. That doesn’t scream “bubble,” but it does whisper “fragile.”

Here’s the catch: overvaluation is not a timing signal.

Every sensible market-timing strategy—whether based on CAPE ratios, earnings yields, or valuation bands—fails once you test it honestly. Not sometimes. Consistently. After taxes and trading costs, doing nothing beats clever timing more often than not. Even during the dot-com era, selling early hurt more than staying invested through the crash.

So what should investors actually do?

First, don’t panic-sell. Valuation alone won’t save you from bad timing.

Second, if you’re uneasy, slow down new investments instead of blowing up your portfolio. Let cash build naturally.

Third, rebalance with discipline, not emotion. Adjust exposure based on risk tolerance, not headlines.

Fourth, forget protection trades unless you know exactly what they cost and when they pay off—they’re expensive and unforgiving.

Finally, if you insist on timing the market, write your rules down, test them brutally, and accept the results without excuses.

Markets can stay expensive longer than your patience—and longer than your confidence.

Takeaway: Being right about valuation is easy; making money from it is the hard part.

The conversation around Bitcoin has matured. The easy gains are gone, and so is the illusion that Bitcoin is a shortcut to riches. What remains is something quieter, slower, and arguably more important: a durable form of savings in a world that is becoming steadily more unstable.

This shift marks a clear dividing line between Bitcoin’s past and its future. The age of explosive 40 per cent annual returns is likely over. Investors today face a “get-rich-slower” landscape. That may disappoint speculators, but it should not discourage long-term thinkers.

Bitcoin still possesses rare qualities. Its supply is fixed. It cannot be printed, diluted, or altered by political pressure. It is portable across borders and durable across regimes. These features matter far more in a chaotic world than they did in a stable one.

And the world is anything but stable.

A Faster, More Fragile Era

We are living through one of the fastest periods of disruption in modern history. Artificial intelligence, automation, and digital infrastructure are reshaping how people work, how capital moves, and how governments respond to economic stress. Jobs are becoming less secure. Skills expire faster. Entire industries are being rebuilt in real time.

This pace of change has consequences. When labour markets are disrupted, governments step in. They spend more, subsidise more, and borrow more. When growth falters or inequality widens, the usual response is monetary expansion. History suggests this pattern will repeat.

For savers, that is bad news. Cash loses value quietly. Fixed-income assets struggle when inflation persists. Traditional savings vehicles are often designed for a slower, more predictable world.

This is where Bitcoin’s relevance sharpens.

The more uncertain the system becomes, the more valuable a dependable store of value becomes. Bitcoin does not promise stability in price. It promises stability in rules. Its supply schedule does not change in response to elections, crises, or popular demand. That consistency is rare.

From Speculation to Accumulation

Bitcoin’s role is changing. It is moving away from being a speculative trade and towards being an accumulation asset.

In its early years, price mattered more than purpose. Today, purpose matters more than price. The sensible approach is no longer to chase waves or time cycles. It is to build exposure steadily, over time, with clear expectations.

This does not mean blind optimism. Lower returns do not mean lower relevance. If anything, they force discipline. Investors must plan for steady appreciation rather than dramatic spikes. Strategies should work even if Bitcoin grows at single-digit or low double-digit rates.

That mindset shift is healthy. It aligns Bitcoin with long-term financial planning rather than short-term excitement.

The Macro Case Remains Intact

Every major technological leap creates winners and losers. Productivity rises, but so does inequality. Capital concentrates faster than labour can adjust. Governments are pressured to intervene, often at the cost of currency stability.

Macro uncertainty is not a side story. It is central to Bitcoin’s case.

Bitcoin stands apart because its rules do not bend. It does not care about growth targets, employment numbers, or fiscal deficits. In a world where almost everything is negotiable, that rigidity becomes a feature, not a flaw.

This does not make Bitcoin a cure-all. It is not immune to volatility, regulation, or sentiment. But its core promise remains unchanged: a form of money that cannot be quietly altered.

How Investors Should Think About It

The practical response is straightforward.

First, adjust expectations. Bitcoin is no longer a lottery ticket. Treat it as a long-term asset with asymmetric upside, not guaranteed miracles.

Second, use consistent accumulation. Regular investing removes emotion and reduces the need to predict market cycles. It suits an asset whose value plays out over years, not months.

Third, keep balance. Bitcoin works best as part of a diversified portfolio. It can hedge against monetary debasement, but no single asset should carry the entire burden of your financial future.

Fourth, extend your time horizon. Think in five- to ten-year windows. Short-term noise matters far less than whether Bitcoin’s underlying logic is strengthening.

Finally, be honest about your goals. If you want stability, independence, and long-term control, Bitcoin fits the moment. If you want fast riches, you are late to that party.

A Different Kind of Advantage

In a world that is moving faster than anyone can fully understand, the real edge is not speed. It is certainty.

Bitcoin’s true power is not speculation. It is the chance for ordinary people to save in a system where the rules are known in advance and cannot be rewritten. That may not make headlines. But over time, it may matter more than anything else.

“Investing in US stocks is tricky, you never know which companies are benefiting from genocide or actively participating in it.”

It stopped me in my tracks.
It also quietly answered a question I’ve been wrestling with for months:

“How did the world allow genocide in Europe in the 1920s and 1930s?”

The truth is uncomfortable: we’re allowing it to happen in exactly the same way, by whispering to ourselves that there’s nothing we can do, that we’re too small to make a difference, and then quietly slipping back into business as usual. I’ve sat with this for a while, and it’s left me with two realisations. One: We need to rethink how we invest. Two: We need to rethink how we live. This piece is about the first — our money, our markets, our choices.

The second, the harder conversation about our habits and daily actions, I’ll save for a follow‑up on my blog. You can follow it by subscribing it below.

Cerebral Circus

This is just scribbling my thoughts, exploring new ideas, observations, and experiences of life. My modern diary

By shoaib zaman

How Companies Get Involved

When you start looking into it, there are usually three layers of corporate involvement:

1. Direct profiteering: The company is directly earning from the genocide. That might mean selling weapons, technology, or services that enable it (Major defence companies and Tech companies in the US).

2. Indirect profiteering: The company supplies goods or services to another company that is directly involved, and benefits from that relationship (example JCB or many others in the construction space).

3. Active support through bias: Company leaders take sides, publicly or quietly, in ways that enable or excuse atrocities. This could also come from bias for Israel or bias against Palestine — either can be used to justify their actions on the ground.

How to Spot Them

The first group is the easiest to uncover. Media reports, NGO investigations, and sometimes even a company’s own earnings calls make it clear. A simple starting point: check if they have a subsidiary or office in Israel.

The second group requires a bit more effort — consider their top clients, partnerships, and supply chains. Follow the money and you’ll see the links.

The third group is harder. Maybe a 50–50 chance of figuring it out.
Two quick checks:

• Is the company named in the Boycott, Divestment, and Sanctions (BDS) movement?

• Where is the CEO based? If in the US, search their speeches, interviews, or donation records for signs of support for the occupation or the violence.

Why This Matters to Us as Investors

We’ve been here before. Remember the global outrage over Blood Diamonds in the late 1990s? The media exposés, the NGO reports, the celebrity campaigns? It led to the Kimberley Process in 2003, an imperfect but significant attempt to choke off the trade.

That happened because enough people refused to keep their money in companies that profited from bloodshed. It’s time we dusted off that playbook.

I’ve been compiling a list of well-known, publicly listed companies that many investors would consider “safe” or “blue-chip” but that, in my view, belong under one label: Genocide Profiteers.

It’s not a comfortable read. It shouldn’t be.

The Small Things We Can Do

No one person can stop a war. But as investors, we hold levers of influence we don’t always use.

• Research before you invest – Don’t just check the P/E ratio. Follow the supply chain.

• Ask your fund manager – If you own mutual funds or ETFs, find out if they’re holding these companies.

• Use your shareholder rights – Question management. Vote on resolutions.

• Support ethical alternatives – Allocate capital to businesses that show transparency and responsibility.

When our children will ask us, “What did you do?”
I’d like my answer to be something more than, “I looked away.”

Footnote:
(1) Blood diamonds, or conflict diamonds, are gems mined in war zones, often under brutal conditions involving forced labour and violence, and sold to fund insurgencies or warlord activities. They fuelled the civil wars of Sierra Leone, Angola, and the Democratic Republic of Congo in the 1990s. Global outrage led to the Kimberley Process in 2003, designed to certify diamonds as conflict-free. It reduced conflict diamonds to under 1% of global trade, but loopholes remain, proof that vigilance can’t be a one-time effort.

China’s economy is at a turning point. For the third straight year, the population has

Read more