Michael Bur called the market’s current phase "denial" and said that stage lasts 6 to 9 months; if that’s right, the next "everything goes kaboom" phase lands in the first half of 2027.
Bur swapped shorts for puts on NVDA, PLTR, and MU expiring at various points next year, after previously predicting the 2008 subprime crash.
The 10-year Treasury pushed above 5.35 a couple days ago, the highest level in more than 24 years and the highest since 2002; oil is flirting with $100 and the S&P is near 7,800.
The Fed raised rates on September 16, the first hike in years since 2023; guidance was read as pointing to another hike this year.
Headline inflation is 3.4% and gasoline is up 27% in a year, while Brent crude is near $100 and mortgage rates are 75.
Q3 earnings for the S&P 500 are expected to grow almost 30%, implying eight straight quarters of double-digit growth.
The S&P 500 4PE is about 19, roughly equal to its 10-year average of 19; earlier this year Ford’s 4PE was around 22, so the market got cheaper, not pricier.
On January 31, 2023 Bur tweeted "sell" when the S&P 500 closed around 4,77; the index is around 7,800 now, up almost 93% in under three years, and $10,000 invested then would be about $19,200 today.
Bur posted in March 2023, "Yeah, I was wrong to sell," underscoring the cost of getting both the exit and re-entry right.
The 10-year has averaged around 5.15% since 1953, fell to about 0.5% in summer 2020, and hit almost 16% in 1981; the napkin math is 3.4% inflation plus a bit over 2% real growth, or about 5.4%, matching a 5.3% 10-year.
When the 10-year averaged 10.6% in the 1980s, the S&P 500 returned about 17.5% per year; at 6.7% in the 1990s, it returned about 18%; in the 2000s the 10-year averaged about 4.5% lower than now and the S&P went nowhere for years.
When the 10-year was 6% or higher, the S&P 500 averaged about 14.5%; when it was below 4%, returns averaged about 7.7%, so high yields are not automatically a panic signal if earnings are rising.
Higher rates hit two ways: debt and discount rate. A company with $8 billion of debt refinancing from 3.5% to 7% doubles its interest bill and cuts pre-tax profit about 30%, while a higher risk-free rate lowers the present value of future cash flows.
The avoid list is unprofitable trash, REITs, home builders, small caps with floating-rate debt, and leveraged cyclicals; the preferred bucket is net-cash compounders, cash-fortress mega caps, companies with real pricing power, and big margins.
Hyperscalers are buying chips that go obsolete fast and amortizing them over five or six years, but if replacement cycles are shorter than the books assume, current profits are overstated and bigger unplanned bills hit later.
Bur said hyperscalers have about $3 trillion of AI hardware and data-center commitments not yet on the balance sheet because accounting rules delay recognition until delivery.
Steve Iceman argued depreciation timing won’t matter if AI demand stays strong and older chips stay in use, and cloud revenue backlogs and earnings are growing like crazy.
March 2000 looked far more extreme than today: the S&P 500 4PE was around 25, the NASDAQ traded at triple-digit multiples, and Pets.com-type IPOs with no revenue could triple in a day.
Today’s setup is not 2000: the S&P 500 4PE is 19 while earnings are growing almost 30% per year, and the average return over the next year after a new S&P high since 1954 is about 14.6%.
Peter Lynch’s line that it’s never a bad time to buy a great company is used to argue that a seemingly expensive entry can become cheap if the business keeps compounding.
JP Morgan data cited here says $10,000 fully invested in the S&P 500 from 2005 to 2024 grew to about $71,750; missing the 10 best days out of roughly 5,000 cuts that to about $32,900.
Seven of the 10 best days occurred within two weeks of the 10 worst days, so trying to sidestep drawdowns can also miss the snapbacks.
$100 a month into the S&P 500 over 25 years produced radically different outcomes across 2000, 2008, 2020 and 2022: panic-selling ended down 94%, pausing DCA and resuming reached 314%, always DCA reached 385%, and doubling down on drops reached 440%.
The S&P 500 has been positive in 100% of all 20-year periods, and a bear market averages about 10 months.
The three DCA paths are: if the 10-year cools back to 4.5%, rate-sensitive stocks and growth should rally and DCA stays normal; if rates stay 5% to 6% and earnings grind higher, weak leverage names bleed out but DCA stays normal; if a Bur-style break sends the market down 20% to 30% or more, the DCA bank gets opened and doubled down.
The readiness checklist is an emergency fund, a preplanned 50/50 split between regular DCA and cash for dips, bond positioning based on retirement distance, and trimming winners into the DCA bank with explicit rules: trim 50% when a position is up 50%, trim 10% when up 100%, trim 20% when up 150%, and trim 30% at the next step.
The key threshold is 7,040, defined as 10% below the 7,819 record/52-week high; for growth names such as PLTR, NVDA, and TSLA the trigger should be 20% instead of 10%, with an example of putting in $1,500 if a $1,000-a-month plan has $500 reserved for the bank and the S&P 500 drops below 740, while the ceiling stays fixed.
The closing argument is that earnings are growing almost 30%, the market is priced at the average 10-year yield, and AI is the biggest technology shift, so the setup favors buying through volatility rather than waiting for perfection.