India Returns to the Spotlight as Healthcare Matches Semiconductors in Overweight Allocation: BofA Asia Survey Reveals Capital Flows After AI Theme Takes a Hit

Stock News11-18 19:45

According to a Bank of America Asia fund manager survey released on November 17, as economic growth and corporate earnings expectations weakened, surveyed fund managers broadly slashed their overweight allocations to East Asian semiconductor sectors and global technology stocks. India's stock market, which has extremely low exposure to AI computing power and semiconductors, turned to a net overweight for the first time since January this year, while healthcare recorded its highest net overweight since the option was introduced in October 2023, tying with semiconductors and key tech hardware as the most favored allocation sectors for the first time.

Overall, the BofA fund manager survey shows that this round of large-scale repositioning by global fund managers is characterized by cooling preferences for technology themes such as semiconductors, while simultaneously raising the relative position of traditional defensive sectors like healthcare, consumer staples and utilities in investment portfolios.

The BofA Asia fund manager survey covered the period from November 6 to 12, 2026, and was published on November 17, a few hours after BofA's global fund manager survey. The combined regional questionnaires had 94 respondents managing a total of $241 billion in assets. Of these, 21 listed Asia-Pacific as their investment specialty. However, BofA did not separately disclose the assets under management for these 21 individuals.

Market data shows that the Philadelphia Semiconductor Index, widely regarded as the global bellwether for AI computing power, had already entered a technical bear market back in July. Based on closing prices, the index hit a historic high of 14,634.72 points on June 22, then pulled back 20.23% by July 17, crossing the commonly defined bear market threshold. On November 5, the index fell to 9,011.14 points, then rebounded 17.23% over the following four trading days, but dropped another 5.28% on Monday, November 16, and edged up only 0.34% on Tuesday ahead of Nvidia's earnings release, closing at 9,642.88 points. By that calculation, the index sits 34.11% below its June historic closing high and 27.05% below its October 6 interim high. Over the same period, the Nasdaq 100 fell 2.67% on Monday and 0.72% on Tuesday, down 4.87% cumulatively since the start of November. The previous day, Monday, the Dow, S&P 500 and Nasdaq Composite fell 1.41%, 1.92% and 2.49% respectively, with chips and semiconductors bearing the brunt of the broader market selling pressure.

BofA's global fund manager survey covering November 6 to 12 explains why such heavy de-risking has not yet triggered its contrarian buy signal. The 169 global survey respondents managed a combined $461 billion, and the net proportion overweight technology stocks dropped from 37% to 12%, a 25-percentage-point decline in a single month — the largest drop since June 2006. But average cash positions also unexpectedly fell from 3.9% to 3.6%, with only 3% expecting a recession in the coming year, highlighting that overall equity market risk exposure remains elevated. Meanwhile, 34% cited an AI bubble as the biggest tail risk, 31% chose a hyperscaler credit event, and 52% expect at least one major cloud provider to cut capital expenditure in the next six months.

BofA strategist Michael Hartnett, known as Wall Street's most accurate strategist, maintained a tactical defensive stance based on this: historically more attractive buying points for tech stocks tend to appear after positioning falls further to a net underweight and cash positions rise significantly.

India Returns to Overweight, Healthcare Ties for First Place with Semiconductors

How healthcare and chips arrived at the same position says more about shifting capital preferences than the tied ranking itself. In Asia ex-Japan markets, the net overweight for healthcare and pharmaceuticals rose 9 percentage points month-on-month to 29%, while semiconductor and key tech hardware allocation themes each fell sharply by 26 percentage points to 29%. The BofA survey showed that non-retail discretionary consumer, retail e-commerce and utilities rebounded by 31, 21 and 16 percentage points respectively, though all three sectors remain net underweight. Industrials and materials saw even larger allocation declines, dropping 35 and 29 percentage points respectively. It is evident that global fund managers are doing two things simultaneously: reducing concentrated positions in semiconductors and chips — sectors closely tied to AI computing infrastructure amid slowing growth — along with some cyclical industries, while replenishing previously underweight defensive domestic-demand sectors.

India's shift also deserves closer scrutiny. Its net overweight rose to 5%, roughly on par with South Korea, though still below Japan's 38% and Taiwan's 24%. The biggest concern among respondents regarding India remains the lack of listed targets that directly benefit from AI, cited by 29%; however, the proportion citing weak economic growth as the primary concern dropped from 15% to zero. Actual trading also shows that India has earnings drivers independent of the chip/semiconductor super-cycle in Northeast Asian markets. On November 17, as Northeast Asian markets continued to weaken sharply, India's Nifty 50 rose 0.38% against the trend; as of November 18, the index was still up 9.4% for the month. Falling oil prices improved input costs for some Indian companies, and combined with resilient corporate earnings, foreign portfolio investors poured in approximately $10 billion during the same period.

AI Computing Power Is Cheaper Now, So Why Are Fund Managers Still Focused on AI Revenue Data?

The contradiction facing Asian investors is that valuation attractiveness has improved, but earnings prospects are insufficient to support very high return expectations. The net proportion expecting earnings improvement for Asia ex-Japan companies over the next year dropped from 55% to 5%, and the net proportion expecting rising inflation fell from 50% to 5%. Meanwhile, a net 29% believe stock valuations are cheap, yet the average expected return over the next year dropped to 4.7%, slightly below the historical average.

Regarding China's stock market, the survey showed that a net 19% of surveyed fund managers expect the economy to weaken, but when selecting the two most favored investment themes, AI and semiconductors still received 48% overweight selection, while traditional safe-haven and defensive sector healthcare came second at 29%, a new high since the question was introduced in April 2023. Weakening macro judgments can still accommodate optimism about specific industries, while safe-haven and defensive postures are becoming increasingly evident, particularly reflected in healthcare's selection rate hitting a new high since BofA introduced that question in April 2023.

Japan exhibits the same divergence. The net proportion expecting Japan's economy to improve stands at 33%, falling to the 16th percentile historically, with 62% expecting the Bank of Japan to raise rates in December; respondents still expect Japanese stocks to gain 5.1% over the next year, above the historical average of 4.8%. Corporate earnings remain the primary driver for Japanese stocks, while the importance of the yen exchange rate has risen to match monetary policy normalization. In the selection of the market that will benefit most from the next phase of the AI cycle, Japan rose from 20% to 33%, surpassing Taiwan's 29% for the month; looking back to May, the popularity of Japan and South Korea has already reversed. Japan's most favored domestic sector remains semiconductors, with 67% listing it among their top two overweight sectors, followed by banks at 48%.

Regarding AI itself, the question investors care most about has shifted from "how much more needs to be built" to "how much can be earned once built." Semiconductor cycle confidence, measured by South Korean and Taiwanese export growth, fell to a net negative 5%, at the 12th percentile historically. The proportion choosing to rotate into defensive sectors to hedge AI downside risk rose from 15% to 33%, but 86% still list AI capital expenditure and business monetization as the primary factors affecting Asia-Pacific stock markets over the next year. Within the supply chain, those believing software and platforms, and AI computing power offer the best risk-reward rose from 5% and 10% to 24% respectively; memory dropped from 30% to 19%, network interconnect from 20% to 10%, and data center infrastructure from 15% to 5%. What best illustrates this round of fund manager capital trade-offs is that 67% believe actual AI revenue would most boost confidence to add positions, while only 5% chose accelerated AI capital expenditure by cloud computing hyperscalers.

Overall, the BofA survey shows that across the entire AI supply chain, software and platforms, and AI computing power tied as the top risk-reward picks at 24% each, with memory trailing at 19%; as for what would boost confidence to add positions, 67% of respondents most value evidence that companies have already generated revenue through AI businesses, while only 5% most value cloud giants reaccelerating capital expenditure. Rather than a new round of computing power procurement plans, investors clearly want to see customers already paying for AI products and services.

Nvidia Earnings Ease Order Concerns, East Asian Semiconductor Rally Still Awaits Earnings Follow-Through

The BofA global survey provides a cautious and somewhat pessimistic risk appetite backdrop for this round of Asian repositioning. The global net overweight for technology among fund managers plunged from 37% to 12%, a 25-percentage-point single-month decline — the largest since June 2006. But the cash ratio also fell from 3.9% to 3.6%, with only 3% expecting a recession in the coming year, while 52% expect at least one hyperscaler to announce capital expenditure cuts in the next six months. BofA strategist Michael Hartnett believes that with technology still net overweight and cash still low, the market remains far from the historically more attractive contrarian buying point. His cautious phrase "not enough panic" in the report reflects his judgment on the timing of position expansion and dip-buying capital entering the market.

The Philadelphia Semiconductor Index entered a technical bear market long ago, and recent rebounds have not recovered major lost ground. On Tuesday, ahead of Nvidia's earnings release, the index edged up 0.34% to close at 9,642.88 points, still 34.11% below its June historic closing high and 27.05% below its October 6 interim high. The Nasdaq 100 fell 2.67% on Monday and 0.72% on Tuesday, down 4.87% cumulatively for the month as of Tuesday. Governance uncertainty surrounding Jensen Huang's departure, combined with tightening export rules and financing concerns, further depressed chip stock valuations. Before the earnings release, Goldman Sachs cut its Nvidia target price from $300 to $220, Citi from $285 to $225, and UBS from $300 to $240, all three maintaining "Buy" ratings. Goldman Sachs particularly noted that the order backlog helps support earnings over the next 12 to 18 months, and leadership changes could have a larger impact on valuation multiples.

The declines in Northeast Asian markets reflected this layer of concern: South Korea's KOSPI fell 3.59% and 2.59% on Monday and Tuesday consecutively, Taiwan's weighted index dropped 6.03% and 2.31% over the same period, and the Nikkei 225 fell 2.42% and 1.34% respectively. Nvidia's earnings released after Tuesday's close provided evidence that near-term computing power procurement remains strong. Third-quarter revenue grew 103% year-over-year to $115.8 billion, data center revenue grew 110% to $107.6 billion, and Vera Rubin contributed nearly $20 billion to sequential revenue growth. Fourth-quarter revenue guidance had a midpoint of $130 billion, about 3.8% above consensus estimates, and excludes China data center computing revenue. The board stated it has no intention of reinstating Jensen Huang, while Kress raised the fiscal 2028 revenue growth forecast from about 70% to about 75%, while emphasizing selective support for new cloud projects. Nvidia rose more than 4% after hours at one point, with Broadcom, Micron, CoreWeave and Nebius subsequently moving higher. On Wednesday, South Korean stocks rebounded 1.99%, while Taiwan and Japan recovered about 0.26% and 0.28% respectively, none recovering the prior two days' losses.

For investors holding Asian AI computing infrastructure supply chain stocks, this earnings report answered "are customers still buying"; the next question revealed by the BofA survey is "why will customers continue buying." Foundries, advanced packaging and high-bandwidth memory are the first to benefit from rack deliveries; for orders to extend into the next round of capacity expansion, cloud providers still need to convert deployed computing power into paid usage and generate revenue sufficient to cover depreciation, electricity and financing costs. This explains why the same group of respondents can continue to overweight semiconductors while simultaneously increasing healthcare allocations, and list AI business monetization as the most valued evidence when adding to investments.

Disclaimer: Investing carries risk. This is not financial advice. The above content should not be regarded as an offer, recommendation, or solicitation on acquiring or disposing of any financial products, any associated discussions, comments, or posts by author or other users should not be considered as such either. It is solely for general information purpose only, which does not consider your own investment objectives, financial situations or needs. TTM assumes no responsibility or warranty for the accuracy and completeness of the information, investors should do their own research and may seek professional advice before investing.

Comments

We need your insight to fill this gap
Leave a comment