Ten companies, picked with no exit plan in mind. Not a portfolio recommendation — a discipline exercise: what actually justifies never selling something, and what happens when a name that looked like a permanent holding two years ago gets out-executed in real time, in front of everyone, this year.
This letter is a thought experiment we were asked to run properly rather than wave at: pick ten companies from anywhere in the world, assume you can never sell them, and explain why each one earns that permanence at today's price — not at some imagined future price. The honest constraint that makes this hard is the same constraint that makes it worth doing: "forever" is not a genre of stock, it is a claim about durability that markets test constantly, and at least one name on this list has already had that claim tested and partially fail within the past two years. We kept it in anyway, because the failure is more instructive than a clean list of ten winners would be.
Market cap approximately $2.0–2.2 trillion, trailing P/E near 29–33x, return on invested capital above 50 percent. TSMC is not merely a market leader in advanced chip fabrication — it is the sole commercially viable source for the most advanced nodes that Apple, Nvidia, and AMD design against. That is not a moat in the usual sense; it is a chokepoint the entire modern computing industry routes through. The multiple looks expensive only until you price what happens to global technology supply chains if this single company stumbled.
P/E near 47x as of mid-2026, against a European semiconductor-equipment peer average closer to 64x — genuinely cheap relative to its own sector despite the headline multiple looking rich in isolation. ASML is the only company on Earth manufacturing the EUV lithography machines that make leading-edge chips possible at all; TSMC cannot produce its most advanced chips without ASML's tools. Few businesses anywhere occupy a position this close to a literal monopoly, verified by regulators and competitors alike simply failing to produce a substitute.
Trailing P/E near 23x, forward P/E near 21x, return on equity 34 percent, $125 billion in trailing annual profit on $318 billion in revenue. The least exciting name on this list, deliberately: Azure, Office, and enterprise software lock-in compound quietly regardless of which AI narrative currently dominates headlines. Boring compounding, at a reasonable multiple for the quality on offer, is the entire thesis.
Visa and Mastercard operate a duopoly over the rails nearly all global card-based commerce runs through, without ever taking on the credit risk of the transactions themselves — a toll-booth business model with structurally higher returns on capital than almost any lending institution it sits alongside in a typical portfolio. The moat here is regulatory, historical, and network-effect-driven simultaneously, which is a difficult combination for any new entrant to replicate.
Increasingly a bet on permanent capital structure rather than on any individual at the helm: decades of insurance float deployed patiently across railroads, utilities, and public equities, with a demonstrated willingness to hold cash for years waiting for genuine mispricing rather than forcing capital into mediocre opportunities. The thesis survives management transition better than almost any other name on this list, which is precisely the point of a "forever" holding.
This is the name that requires the most honesty, and the reason we said above that a failure is more instructive than ten clean winners. Two years ago, this slot would likely have gone to Novo Nordisk, the GLP-1 and obesity-drug category leader. As of mid-2026, Novo Nordisk's share price has fallen approximately 36 percent over the trailing year, its price-to-earnings ratio has compressed to roughly 12x from a much richer multiple, and its own recent quarters show real pressure from Eli Lilly's competing drugs and from Medicare pricing dynamics tightening across the category. We are not including Novo Nordisk on this list. We are including Lilly instead, and naming directly why: even a category-defining healthcare moat, built over decades, can be genuinely out-executed by a specific competitor within a couple of years. That is not a reason to avoid this sector — obesity and metabolic disease treatment remains one of the largest addressable healthcare markets in the world — but it is a direct, current demonstration that "forever" requires re-underwriting who is actually winning, not just which category is growing.
"Novo Nordisk looked like a permanent holding as recently as 2024. The lesson is not that healthcare moats are fake — it's that even a real one needs the same discipline as any other holding: check who is actually winning, not just which market is growing."
Search cash flow, still the core engine, funds genuine optionality in cloud infrastructure, AI research, and Waymo's autonomous-vehicle effort — priced, persistently, at something close to a conglomerate discount relative to Microsoft despite broadly comparable underlying business quality. The thesis here is that the market has not yet decided how much credit to give the optionality, which is exactly the kind of mispricing a permanent holding can afford to simply wait out.
The closest thing the luxury goods sector has to a permanent moat: multi-decade brand desirability compounding across dozens of maisons, from Louis Vuitton to Dior to Moët. The honest caveat, stated directly rather than glossed over: China's consumption slowdown has weighed on the entire global luxury sector recently, LVMH included, and this is realistically the most cyclical, lumpiest name on this list in any given two-to-three-year window. The "forever" case rests on multi-decade brand equity outlasting any single macro cycle, not on smooth near-term performance.
The deliberately unglamorous pick. A regulated utility with one of the largest renewable energy development pipelines in the United States, generating cash flows that are structurally indifferent to whichever technology theme is dominating markets in any given quarter. Utilities rarely feature on lists like this because they rarely excite anyone — which is precisely the quality a genuine buy-forever list needs at least one example of: an asset whose value proposition does not depend on being interesting.
Market capitalisation near ₹3.9–4.0 lakh crore, P/E near 27–28x, India's third-largest private bank by market capitalisation. The thesis is diversification within a single holding: banking, broking (an 11.8 percent market-share business in its own right), asset management, and insurance all sit under one roof, recently expanded further by the acquisition of Deutsche Bank's India private banking and wealth management business — precisely the high-margin segment that benefits most as India's affluent population grows. The honest complication: the stock is down roughly 11 percent over the trailing year, and CEO Ashok Vaswani's departure has left a leadership transition underway with no permanent successor yet named. The "forever" case here rests on Kotak's multi-decade record of conservative underwriting through India's various credit cycles, not on the smoothness of the next two quarters.
Every name on this list earns its place through one of three mechanisms: a genuine structural chokepoint no competitor can currently replicate (TSMC, ASML, Visa), a diversified and patient capital-allocation structure that survives any single leader's departure (Berkshire, Microsoft, Alphabet), or a multi-decade brand or franchise durability that specifically outlasts near-term cyclicality (LVMH, Kotak, NextEra). Eli Lilly sits partly in the second category and partly as this letter's deliberate reminder that even the first category — a seemingly durable moat — requires active re-underwriting, not passive faith.
This letter's single biggest vulnerability is survivorship bias dressed up as conviction: every "forever" list ever published has, in hindsight, included at least one name whose moat quietly eroded over a decade rather than collapsing dramatically enough to force a rethink. Novo Nordisk's inclusion two years ago and its removal now is presented here as evidence of intellectual honesty, but the same fate could await any of the other nine names on this list without warning — TSMC and ASML carry genuine geopolitical concentration risk specific to Taiwan and EUV export controls; Visa and Mastercard face a slow-moving but real regulatory scrutiny cycle over interchange fees globally; and LVMH's China exposure could persist far longer than a single "cyclical" label implies. "Forever" is a discipline for how you evaluate a holding, not a guarantee about what happens to it.
Ten names, picked for durability rather than momentum, at valuations checked directly rather than assumed: TSMC and ASML as genuine chokepoints in global technology; Microsoft, Visa, and Berkshire as compounding machines that survive leadership change; Alphabet as underpriced optionality; LVMH and Kotak as multi-decade franchises currently priced through real near-term cyclicality; NextEra as the deliberately boring anchor; and Eli Lilly standing in for Novo Nordisk as this letter's direct, current proof that even the best-looking moat requires re-checking who is actually winning. A genuine forever list should be able to name its own most likely failure candidate. This one can.
Founder, NextGen Economics · Bangalore, India · July 2026
Sources: StockAnalysis.com & Multiples.vc (TSMC valuation data, July 2026) · Yahoo Finance (ASML valuation analysis, May 2026) · ASML Holding NV Form 6-K, Q1 2026 results · StockAnalysis.com (Microsoft statistics, July 2026) · Yahoo Finance, CNBC, MacroTrends, CompaniesMarketCap (Novo Nordisk market cap and valuation, June–July 2026) · CompaniesMarketCap, Bajaj Broking, MoneyWorks4Me, Value Research, Groww, INDmoney (Kotak Mahindra Bank valuation and financials, June–July 2026).
Not investment advice. This letter is a thought experiment on evaluating long-term holding discipline, not a recommendation to buy or sell any security. Valuations fluctuate continuously; verify current figures before any decision.