The AI bull market has entered the “era of cash out”! Big Mo and Little Mozi look at S&P 8,000 points, and the violent counterattack of semiconductors and the Korean stock market verifies that “profit owners are rising.”

The Zhitong Finance App learned that since August, against the backdrop of a massive rebound in the semiconductor sector and technology stocks dominated by AI computing power infrastructure themes, the recent sharp volatility in the global stock market has rapidly subsided. The two major Wall Street financial giants, Morgan Stanley and J.P. Morgan Chase, recently released a tacit research report saying that the primary driving force driving the S&P 500's upward trend is expanding from valuation expansion, switching to profit improvement+AI commercialization and implementation. Last week, J.P. Morgan Chase (“Komo”) raised its target for the end of 2026 from 7,800 points to 8,000 points, and raised the EPS profit trajectory for this year and next two years. Morgan Stanley (or “Daimo”) previously also raised its 2026 target to 8,000 points and the 12-month target to 8,300 points, and made it clear that the increase was mainly due to profit rather than valuation. Currently, at least seven Wall Street institutions expect the S&P 500 to reach 8,000 points by the end of 2026.

As profit expectations drive the S&P 500 index to 8,000 points, it resonates strongly with the price behavior after deleveraging the AI computing power infrastructure theme in July: the Philadelphia Semiconductor Index plummeted nearly 29% from the June 22 high to the July 29 low, then quickly rebounded about 20% from the low; the Korean benchmark stock index KOSPI rebounded nearly 22% only about two weeks after the July 30 low and re-entered a technical bull market.

South Korea's sharp decline in July included extremely leveraged ETFs and forced exit factors. The scale of leveraged products dropped from about 50 billion US dollars to 17 billion US dollars; yet the fundamentals of AI Memory (AI driven memory chip super bull market) did not collapse at the same time, and the industry is still discussing the tight supply of DRAM/HBM and the 2027 demand gap. As a result, this AI-led bull market is increasingly in line with the positive feedback of “deleveraging — resetting market positions — re-taking risks — FOMO sentiment is heating up more and more”, not like a dead cat rebound after a round of AI profit cycles peaked.

Another Wall Street financial giant Citadel gave evidence of capital flow showing that this round of counterattack has moved from “fundamental repair” to a stage of “self-strengthening buying.” According to its official August report, the S&P 500 EPS growth rate in the second quarter was about 33%, and the profit improvement path was at least one of the steepest since 2000; at the same time, the 12-month forward P/E fell from about 23.1 times in October last year to 20.1 times, that is, until now, mainly profit expansion rather than multiple expansion (valuation/price-earnings ratio expansion) has boosted the index.

From “buying a shovel” to “who uses a shovel to extract profits”: the ultimate indicator of AI investment becomes ROIC and free cash flow

J.P. Morgan has raised its 2026 S&P 500 target point from 7,800 points to 8,000 points. The reason is that the second-quarter earnings season performance was extremely strong, and there is growing evidence that large-scale investment in artificial intelligence is being transformed into stronger business performance. The bank raised its profit forecast for 2026 and 2027 at the same time. However, high interest rates, geopolitical risks, and huge new supply in the capital market still limited its assumptions about valuation multiples.

As 87% of the S&P 500 constituent stocks have announced results, the J.P. Morgan strategist team led by Dubravko Lakos-Bujas said that the profit picture “remains strong, and this strength is widely distributed across multiple industries.”

This strong performance prompted the bank to raise its 2026 earnings per share (EPS) forecast to $365, representing a 35% increase over the previous year. This forecast is also higher than the current market's consensus estimate of $358. For 2027, J.P. Morgan raised the EPS forecast to $420, which means an additional 15% increase on this basis.

Part of the unusually strong profit growth was due to an increase in the investment value of private enterprises held by listed companies. The J.P. Morgan strategist team predicts that according to the relevant valuation adjustments recorded in the first half of 2026, these valuation changes contributed about $18 to the S&P 500 EPS.

Excluding this effect, the normalized EPS in 2026 would be approximately $347. Even according to this adjusted caliber, the year-on-year profit growth rate still reached about 28%, highlighting the strength of the bottom line of the profitability of AI-driven companies.

One of the most important changes in this earnings season is a shift in market discussions around AI investment in hyperscale cloud computing companies.

Investors are increasingly shifting their focus from the scale of AI capital expenditure to whether these expenses can create an attractive return on invested capital (or ROIC). J.P. Morgan believes that the latest results have provided encouraging evidence that commercial monetization is beginning to appear.

According to J.P. Morgan Chase, the strongest examples come from Google, Amazon, and Microsoft, which “have successfully crossed the high expectations threshold previously set by investors through stronger cloud business growth, expanded order backlog, and improved visibility into operating cash flow.”

J.P. Morgan emphasized that the growth rate of cloud computing business of North American tech giants provides some of the clearest evidence that AI investment is being transformed into stronger customer demand.

The revenue growth rate of AWS, a cloud computing business owned by Amazon, accelerated to 37% year-on-year, while revenue from Microsoft's Azure cloud computing business increased by 43%. Google Cloud, a cloud computing business owned by Google, performed the strongest, with a record increase of 82% in revenue.

The backlog of orders has also expanded significantly. The Google Cloud order backlog increased by 52 billion US dollars from the previous quarter to reach 514 billion US dollars. The AWS order backlog reached $496 billion, up 36% from the previous month, and reached nearly 2.5 times the level of a year ago. As these hyperscale cloud computing giants continue to invest heavily in additional production capacity, these figures provide considerable visibility into their future revenue and profits.

This is why J.P. Morgan is able to raise the index profit forecast even as AI CapEx is rushing to about 900 billion US dollars in 2026 and exceeding 1.2 trillion US dollars in 2027; it is really betting on a closed commercial loop of “AI CapEx — cloud revenue — order backlog (backlog) — operating profit” rather than simply betting on GPU shipments.

Morgan Stanley is taking this logic one step further: the next phase of alpha is spreading from AI infrastructure vendors to AI adopters (AI Adopters) — companies that can actually use AI to increase labor productivity, reduce costs, expand profit margins, and form cash flow may get a higher valuation than companies that simply have an “AI concept.” The team of Morgan Stanley strategists led by Michael Wilson believes that investors are also becoming more picky and are increasingly willing to reward companies that can combine profit growth with strong free cash flow and operational efficiency.

This is why in the medium to long term, Morgan Stanley favors Hyperscalers (i.e. cloud computing supergiants) rather than simply computing power-related stocks such as AI semiconductors. Chip stocks may still return to a tactical upward trajectory after Momentum is liquidated in July, but cloud computing giants also have existing cash cow business+AI infrastructure+models/platforms+customer distribution channels+future AI monetization options, and the risk-benefit ratio is more complete. Recently, Morgan Stanley has continued to emphasize that the market is moving from an early cycle of a bull market to a mid-cycle, and that the leading power of the bull market is spreading further from simple high beta to “profit quality+cash flow.”

The bull market has not left technology, but it is shifting from “technology alone” to profit diffusion: “high quality” is the next stage of real scarce assets

What J.P. Morgan provides is the most important multi-faceted evidence of the current AI bull market: the high AI-related revenue growth of AWS, Azure, and Google Cloud and the backlog of orders of hundreds of billions of dollars are beginning to prove that AI capital expenditure is not a pure cost black hole, but rather a transformation into visible revenue. This is why the bank was able to improve index profit forecasts even as AI CapEx heads towards about 900 billion US dollars in 2026 and over 1.2 trillion US dollars in 2027.

The two major Wall Street firms have not actually offered conflicting strategies — J.P. Morgan continues to confirm that the AI supercycle is an engine for exponential profit, while Morgan Stanley tells investors how the bull market should spread in the next phase.

87% of S&P 500 profits exceeded expectations, Russell 3000's median profit growth rate rose to 15%, and profit revisions and improvements in the financial and consumer sectors meant that the market began to expand from Mega-Cap AI (tech giants closely linked to AI) to high-quality finance, consumer, and AI application layers, and a wider range of profitable growth companies. The most critical selection criteria are no longer just growth, but “can growth be converted into cash”: companies that improve EPS (earnings per share) and FCF (free cash flow) at the same time receive significant excess income, while only companies with EPS growth and deteriorating cash flow have begun to be punished by the market.

Judging from the investment period of several months, Morgan Stanley believes that hyperscale cloud computing giants have more attractive risk-return characteristics compared to AI semiconductor forces.

The agency's strategists focused on these companies' “resilient core businesses, attractive relative valuations, and the return on investment and options brought about by AI adoption related to AI computing power infrastructure that the market is not fully aware of.” This combination means that hyperscale cloud computing companies will not only continue to benefit from continued growth in the cloud business, but may also benefit from continued improvements in the return associated with their huge AI investments.

J.P. Morgan Chase and Morgan Stanley are increasingly focusing on “quality” at a time when market leadership/leading forces are spreading. The expansion of the scope of profit recovery has made Morgan Stanley strategists more convinced that market opportunities are no longer limited to a few large capitalization stocks.

As 87% of the S&P 500 constituent stocks surpassed expectations, the median profit growth rate of Russell 3000 constituent stocks reached 15%, the strongest growth rate since 2021, and the profit expectations of various industries were revised and improved, the overall fundamental environment has been significantly strengthened. Meanwhile, Morgan Stanley statistics show that the breadth of S&P 500 index profit forecast revisions has returned to 23%, while 76% of industry groups are recording positive profit revisions. Both indicators are close to their respective cycle highs.

However, investors are also becoming increasingly picky about the specific quality of this growth. As a result, Morgan Stanley favors high-quality companies with strong free cash flow, AI adopters, large financial stocks, and non-essential consumer goods companies. Within the technology sector, hyperscale cloud computing giants are still superior to semiconductor stocks in terms of longer investment periods.

The biggest negative risk is also very clear — long-term US bond yields, oil prices, and capital market financing costs; J.P. Morgan Chase is unwilling to increase the valuation multiplier to more than 20 times, and Morgan Stanley continues to warn of long-term returns, which essentially all indicate that if this bull market continues to rise, the most optimistic scenario will be dominated by Earnings Expansion (continued profit expansion) rather than re-reliance on Multiple Expansion (expansion of valuation multiples).