Original author: Rita

Introduction

Bernstein released a quarterly review of the semiconductor industry on June 23. Key points: AI has become the "only game" in the semiconductor sector, with strong fundamentals, but valuations and overcrowding are at historically high levels. The report simultaneously recommends NVDA and AVGO (rated "outperform"), believing that although they have underperformed this year, they are the most core beneficiaries in the AI supply chain, with current valuations being "ridiculously cheap." It has upgraded AMD but remains cautious on QCOM due to pressures in its mobile business.

AI Demand Fuels Record Gains in the Semiconductor Sector

The Philadelphia semiconductor index (SOX) has risen 155.6% over the past year, increasing 106.6% year-to-date. In the same period, the S&P 500 only rose 9.2%. The premium of SOX relative to the S&P 500 has reached 62%.

This rise has been driven by fundamentals rather than a bubble. Bernstein's data shows that the forward EPS of SOX has increased by 75% from the beginning of the year, while the expansion of valuations themselves is only a small part of this increase.

The differentiation within the semiconductor sector has reached an exaggerated level. From the beginning of the year to June 22, memory chips rose by 500%, while CPUs and optical solutions each rose by 220%, and GPUs and ASICs only rose by 115%. The entire AI supply chain is making money, but the profitability and the degree of profitability are not uniform. The upstream and downstream of the supply chain benefit the most, with new production lines requiring memory and semiconductor equipment, which are in relatively short supply. GPUs only rose by 115%, despite NVDA holding the vast majority of the market share in AI chips.

Actual Purchasing Power Under High Valuations

The forward P/E ratio of SOX is now 34.1 times, while that of the S&P 500 is 21.0 times, a premium of 62%. This sounds expensive, but one must look at specific companies. NVDA's adjusted EPS expectations for 2026 is $9.19, and $12.52 for 2027. According to Bernstein's target price of $315, the P/E for 2027 is 25 times, while the entire sector's forward P/E is 34 times. NVDA is not the most expensive; it is relatively cheap.

Bernstein analyst Stacy Rasgon used a phrase: "ridiculously cheap."

His reasoning is straightforward: NVDA's Blackwell chip series is expected to achieve $1 trillion in revenue by 2027. AVGO has a similar situation, with a target price of $550, but if it achieves the $10 billion AI-related revenue target by 2030, the current valuation looks very cheap.

This is why Bernstein rates both companies as "outperform." Although they have underperformed this year, they are the most core links in the AI demand chain. In comparison, Apple's forward P/E is about 28 times, Microsoft's is 30 times, while NVDA's is at 25 times. Considering the continuity of the Blackwell and Rubin generations of products, and AVGO's monopoly in switch chips, these valuation discounts appear extremely unreasonable. The funds overlook a core fact: without NVDA and AVGO's chips, the entire AI infrastructure cannot operate.

The Dual Story of CPUs and QCOM’s Single Dilemma

AMD has recently been upgraded by Bernstein to "outperform." What is the reason for the upgrade? Because AMD not only has opportunities in AI/GPU but also in the CPU's proxy AI trend. CPU shipments are beginning to improve in Q1 2026, slightly above personal computer shipments. Bernstein believes AMD's fundamentals are sufficient to support a per-share earnings target of $20 by 2028, and the current stock price still has room to rise relative to this target.

QCOM, on the other hand, is in a single dilemma. Smartphone shipments are down 3% year-on-year in Q1 2026, and rising memory chip prices mean increased costs for smartphones, negatively impacting pricing power for chip suppliers. Bernstein acknowledges that the previous downgrade of QCOM was a "bad decision," but still maintains a "market perform" rating. The issue is that the weakness in consumer electronics has become a foregone conclusion, making it difficult for QCOM to find new growth engines. Even if analysts in the future come up with a new story for the data center, compared to AMD’s dual drivers and the structural position of chip manufacturers, QCOM's narrative has limited persuasiveness.

Realistic Considerations for Sub-sectors

Semiconductor equipment (AMAT, LRCX, KLAC) continues to be favored, with strong demand for capacity construction, and all three companies are rated "outperform," with target price increases between 30% and 70%.

The situation for analog chips (ADI, TXN) is more complex. They are indeed in a recovery cycle, having achieved double-digit growth for over a year, but their data center business accounts for only a small percentage, about 10%. The P/E ratios for TXN and ADI are between 30 and 40 times, appearing quite expensive. Bernstein gives both companies a "market perform" rating, opting to wait and see.

Crowding and Inventory Risks

Bernstein's industry sentiment indicators show that the overcrowding in the semiconductor sector is at historically high levels. Inventory days have risen again, well above the historical normal range's upper limit. Although channel inventories have decreased, they still remain above average levels. What does this mean? It means that if there are any signs ofweakness in downstream demand, the entire supply chain will face pressure for proactive inventory reduction. PC and consumer segments have already shown signs of weakness, with smartphones experiencing a year-on-year decline. If inventory pressure spreads to data center procurement, the threat of a price war will become real. At that time, the pricing power of companies close to the bottleneck (NVDA, AVGO) will be severely weakened.

The strength of AI demand is undeniable, but the current high valuations in the semiconductor sector have already priced in these good news. While NVDA and AVGO are relatively cheap, this is predicated on believing they will meet analysts' targets. AMD's story is attractive, but execution risks are also present. QCOM has become a forgotten player, with unclear catalysts. Bernstein's stance is selectively bullish; at this time, the importance of stock selection has surpassed simply being on the right side of the market.

免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。