What If It’s Still Early? | Erik YWR on $10,000 S&P 500 by 2027 Case, Hyperscaler ROIC, AI CapEx, Semis, Exchanges, and Reverse Crash Risk In “Project Zimbabwe”

Erik YWR argues the S&P 500 will reach 10,000 by end of 2027, a target he now considers potentially conservative and may revise to 11,000 or 12,000. The call rests on earnings growth running at 12 to 15 percent annually against a 25-year historical average of 8 percent, with expected S&P 500 earnings of approximately 340 dollars per share for 2026 representing 25 to 26 percent growth, followed by roughly 20 percent additional growth into 2027. Because earnings came in around 50 dollars per share higher than originally anticipated, he now calculates only approximately 25 times earnings is required to hit the target rather than the 30 times he originally assumed. He uses the Gordon growth model to argue that when nominal earnings growth approaches the cost of capital, mathematically justified fair value multiples become very large, and he challenges the idea that 20 times earnings is a natural ceiling, calling it a rule of thumb rather than a fundamental limit.

The Project Zimbabwe thesis holds that higher inflation carries nominal stock prices to unimaginable levels even when underlying economic conditions appear poor, drawing on Zimbabwe's stock market rising dramatically during economic unraveling as investors anticipated inflation before it became obvious. Erik extends this to every nominal stock market priced in local currency, including Turkey, Argentina, Venezuela, Brazil, and Japan, all of which show charts that go up and to the right over time. He argues the underarticulated crash risk is being unexposed to rising nominal asset prices while the cost of living increases around you, and that workers fighting for 10 to 15 percent wage increases while inflation runs at 25 to 30 percent are the ones hurt while holders of stocks, real estate, and businesses see large nominal gains.

Annual data center capex is currently estimated at roughly 600 to 800 billion dollars and Erik predicts it is headed toward one trillion dollars within one to two years, citing a rapid ramp from approximately 200 billion that caught markets off guard. Full-year 2026 capex numbers from Amazon and Google came in higher than expected, which drove the semiconductor trade sharply higher. Cloud revenue growth at hyperscalers has accelerated from approximately 15 percent to over 25 percent compound annual growth rate as AI adoption increases customer usage, which Erik presents as evidence the capex is generating real returning revenue. He does flag a depreciation bomb hitting hyperscalers over the next five years as current data center spending begins to depreciate, and acknowledges that approximately 9.49 percent of relevant S&P 500 earnings growth came from private company valuation markups such as OpenAI and Anthropic, a source that cannot be assigned a recurring revenue multiple. He describes Meta's AI monetization strategy as unclear to him and says if Meta glasses are a flop he has severe concerns about how Meta captures value from its AI investment.

Erik identifies widespread skepticism among fund managers as a contrarian positive. They are bearish on hyperscalers due to doubts about ROIC on AI capex, bearish on software due to disruption risk, bearish on consumers due to AI job displacement, and cautious on memory due to cycle uncertainty. He contrasts this with dot-com era valuations where Microsoft traded at roughly 70 times earnings, pointing out that memory stocks currently trade at roughly 6 to 8 times earnings and Samsung at approximately 7 times, which he does not characterize as valuation bubbles. His global factor model shows the strongest upward earnings revision momentum in computer hardware, semiconductors, memory, and Taiwan companies in the video supply chain, citing SK Hynix 2027 earnings per share estimates rising from approximately 20,000 Korean won to approximately 300,000 Korean won in roughly two years.

Erik views CME and ICE as attractive following a recent selloff, with ICE around 15 times earnings and CME around 18 times, levels he describes as rare for extremely consistent growers and high cash flow generators that rarely trade below 20 times. He notes roughly 90 percent of CME volumes come from institutions requiring contracts that convert to the underlying asset, which perpetual futures do not provide, making him uncertain whether perpetual futures competition is as large a threat as market pricing implies. He also sees European banks including Barclays trading around 8 times earnings with earnings still growing at roughly 5 percent and potential for re-rating to 12 times, and describes global banks broadly as highly capitalized, highly profitable, and facing low loan losses after a decade in the doghouse.

Erik is constructive on energy-adjacent sectors benefiting from Strait of Hormuz supply disruption, including oil and gas, tankers, US refiners, and chemical companies, and predicts oil could be on a road to 150 dollars. He describes Hong Kong as his worst trade despite the Hang Seng trading at less than 10 times next year's earnings, attributing underperformance to concentration in Alibaba, JD.com, and Baidu, and distinguishes it from mainland China A-shares, which represent the hardware and semiconductor play and have performed significantly better. He argues the current environment does not resemble 1999 bubble conditions because investors are not broadly euphoric or overextended, and frames the bear case plainly as hyperscalers overspending on data centers, never recovering their investment, and the S&P falling 40 percent, while conceding there is not yet enough revenue data to fully validate the capex cycle.