It’s the stock every investment style can justify buying, while also justifying avoiding. Nvidia’s rise from a niche gaming chipmaker to a $5-trillion artificial intelligence (AI) bellwether is more than a stock call. How an investment manager answers the Nvidia question, and why, says as much about their discipline as it does about the company itself. At the bottom of the Covid-19 pandemic Nvidia traded as low as $5 a share (split-adjusted); it now trades above $200, with a market capitalisation of about $5-trillion. Long before AI went mainstream Nvidia built its proprietary Cuca software ecosystem, which let its GPUs handle far more than graphics rendering. Researchers soon found the chips uniquely suited to training AI models, and in 2016 CEO Jensen Huang repositioned the business from gaming to AI computing. That repositioning is now the investment case. So why don’t all active managers buy Nvidia? Competitors such as AMD and Intel can match its raw hardware speed but struggle to replicate a software ecosystem built over two decades. Yet valuation, capital spending and uncertainty over how the AI theme plays out keep managers divided. Nvidia’s trailing price-earnings ratio has compressed from an average of 60-70 times in recent years to about 30 times now, its lowest in about seven years, though still at a premium to the market. TSMC, another key AI supplier, averaged closer to 20 times over the same period. The scale of hyperscaler capital expenditure is a related concern: Alphabet, Microsoft, Meta, Amazon and Apple are estimated to spend upwards of $700bn this year, raising questions about when these investments turn profitable. The narrative is also shifting from training to inference, and Nvidia’s next-generation Rubin chip, weighted towards inference, has entered full production. Meanwhile, several hyperscalers are developing their own custom chips. Nvidia’s second-quarter fiscal 2027 results, released on August 26, offered a live test. Revenue came in at $96.2bn, up 106% year on year and ahead of the about $92bn analysts expected, with non-Gaap earnings of $2.22 a share beating the consensus of $2.09. Management guided third-quarter revenue to $108bn, again above expectations, and Huang described AI demand as still accelerating. Yet shares fell more than 2% on the day, extending a pattern of Nvidia dropping despite beating consensus forecasts — a reminder of the gap between fundamentals and price action that momentum managers watch closely. A new cost pressure also emerged: gross margin held at 75% for the quarter, but management guided margins down to 71%-72% by the fourth quarter, citing memory scarcity linked to the broader AI buildout. Supply commitments more than doubled quarter on quarter to $279bn, largely tied to memory procurement. On the competition question, Nvidia and Amazon Web Services announced that AWS will purchase 2-million Nvidia GPUs and deploy its new Vera CPU, some units integrating with the forthcoming Rubin chip — a sign the largest cloud providers remain heavily committed to Nvidia’s ecosystem even as they build their own silicon. Mapping this onto investment styles is instructive. Growth managers are typically less deterred by price, though some still apply valuation discipline and may see Nvidia as expensive regardless of its prospects. A growth at a reasonable price (GARP) manager may instead look to the PEG ratio, which weighs price against expected earnings growth, rather than the price-earnings ratio alone. Value managers are more straightforward: price-sensitive by nature, they have historically avoided Nvidia, though a case can now be made given how cheap the stock is relative to its own history. Quality managers favour wide moats, and Nvidia’s Cuda ecosystem and high switching costs qualify — though moats are not permanent, and technology, regulation and other unknowns can erode them. Momentum managers, who invest on technicals over a three- to 12-month look-back, have found Nvidia a strong play, despite pullbacks such as the postresults dip in August. There are reasons to like or dislike Nvidia, and style alone doesn’t determine the stance — each style has nuances, and not every manager fits neatly into one bucket. For fund allocators, the task is to understand why a manager holds their position and whether it aligns with their stated philosophy and history. A manager known for avoiding expensive stocks suddenly buying Nvidia at peak valuation is a red flag for style drift, particularly after a period of underperformance. Portfolio construction and risk management matter too: an overweight position far exceeding the index may have paid off, but it also introduces sizeable idiosyncratic risk. Distinguish luck from skill — performance attribution shows whether returns came from diversified sources or a handful of concentrated bets. Any investment manager could have bought into the Nvidia hype. If it aligned with sound risk management, that’s a good outcome. If not, the celebration may be premature, and questions will follow. • Munnik is a specialist in research management at INN8 Invest, a division of Stanlib Wealth Management.Business Day