Invest1 publisher3 min readPublished
Capital Economics pencils in 7.7% more upside before a 21% fall to 6,500
James Reilly's path has the S&P 500 up 7.7% by December and down 21% by the end of 2027, a net loss of about 15% for anyone who holds through both legs. Ruchir Sharma's 5% trigger is three basis points away.
The Investor · Invest desk
What happened
- Capital Economics senior markets economist James Reilly reiterated a forecast of 8,250 for the S&P 500 at the end of 2026, 7.7% above Friday's close, followed by a 21% fall to 6,500 by the end of 2027.
- Reilly wrote in a note on Thursday that the data look consistent with a late-stage bubble, with most of the factors he considers at or close to levels that have preceded past stock market peaks.
- Ruchir Sharma, chairman of Rockefeller International, wrote in a Financial Times op-ed that the AI bubble could pop once the 10-year Treasury yield decisively breaches 5%.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- decision The forecast pays only on an exit: capturing the 7.7% means selling in December and standing aside for a year, while holding through both legs turns the same path into a loss of about 15%.
- constraint If hyperscaler free cash flow is negative in 2027, the capital spending has to be raised in the bond and equity markets, the two channels Sharma says tighten above a 5% yield.
- contradiction The two halves of the bear case disagree on timing: the price path still has months of upside in it, while the yield trigger Sharma named sits three basis points from Friday's level, and Reilly's note did not discuss yields.
- exposure On Sharma's reading, a debt burden above 100% of GDP means public borrowing costs reach ordinary corporate and household borrowers before they reach the AI balance sheets.
Back out the 7.7% and the target implies a Friday close near 7,660 [1]. From there, 6,500 is a fall of about 15% [2]. So the upside leg pays only an investor who sells in December and stays out through the following year; a holder who does nothing owns the whole path and takes the 15% [1].
The funding piece is the part with a date on it. Reilly expects the combined free cash flow of the top AI hyperscalers to turn negative in 2027 [6], the same year his forecast puts the index at 6,500 [1]. Negative free cash flow means the capital spending gets paid for from outside the business, in the bond market or the equity market. Ruchir Sharma, chairman of Rockefeller International, argued in a Financial Times op-ed that both get harder above 5%: hyperscalers would issue fewer bonds, and equity would be a tougher sell because yields at that level have historically been a headwind for stocks [13]. "This breach would signal the start of a new era of tighter money, in which AI mega projects will be harder to fund," he wrote [12].
The 10-year hit 4.97% on Friday, three basis points below Sharma's line [c9, d3]. Reilly's note, according to Fortune, did not mention the yield surge [10]. His case rests on price and supply: a cyclically adjusted price-to-earnings ratio close to its dotcom peak and a valuation against Treasury bonds near dotcom extremes [4], forward 12-month earnings-per-share growth in line with the dotcom peak [5], market-cap concentration at extreme levels [7], and an issuance boom with a pipeline of IPOs and follow-ons still to come [8].
Fortune reports that others on Wall Street read the yield move as normalization after years of suppression by central bank policy [15]. Sharma's answer is the debt stock, now above 100% of GDP [15]: "Rising public borrowing costs will squeeze other borrowers sooner, and hit the bubbly AI markets harder," he wrote [16].
Ed Yardeni moved too. He cut the odds on his "Roaring 2020s" scenario from 80% to 70% and lifted the bearish outcome from 20% to 30% [17], half again on the bear side [4], and still a minority of the probability. "Admittedly, recent developments in the oil and bond markets are unnerving," he said in a note Saturday [18].
Two things would break the sequence. If hyperscaler free cash flow stays positive through 2027, the external funding channel never binds and a 5% yield costs the AI trade less than Sharma expects [c6, c13]. If the 10-year backs away from 4.97%, the trigger does not fire [c9, c11]. I would take the 21% more seriously than the 7.7%, on the evidence of Reilly's own issuance indicator: he says that in the past, activity of this kind has signalled that a bubble's end is months away, not years [8].
What to watch
- Hyperscaler capital spending and free cash flow guidance for 2027: cash flow that stays positive removes the funding leg of Reilly's case.
- Whether the 10-year closes decisively above 5% or stalls below it, which is the level Sharma named as the trigger.
- Whether the IPO and follow-on pipeline Reilly flags gets absorbed by buyers or backs up.