Intel (INTC.US) guides thunderstorms, and Bank of America and Komo cut target prices

The Zhitong Finance App learned that Intel (INTC.US) handed over a three-quarter report that exceeded expectations, but the fourth quarter guidance caused the stock price to drop sharply. On October 30, the two investment banks lowered their judgments on the chip giant on the same day: J.P. Morgan Chase remained neutral and the target price dropped from $85 to $80; Bank of America Securities directly lowered from buying to neutral, and the target price was cut from $145 to $100.

Third quarter: the strongest month-on-month ratio in 15 years since DCAI handed over

Revenue for the September quarter was US$16.513 billion, up 20.9% year on year, 0.7% higher than J.P. Morgan's own forecast and basically in line with Wall Street's expectations. What is really impressive is the structure and profit margin: the data center and artificial intelligence (DCAI) business revenue was 7.318 billion US dollars, up 16.9% month-on-month and about 78% year over year, the strongest month-on-month growth in the sector in 15 years, driven by the general strengthening of hyperscale cloud vendors, enterprises and emerging cloud customers; J.P. Morgan estimates that its ASIC business increased by more than 50% month-on-month and more than doubled year-on-year, with an annualized scale of about $2.8 billion.

The profit structure also changed: according to the J.P. Massage segment, DCAI operating profit of US$3,039 million, operating margin of 41.5%, CCG was US$2,285 million, or 28.2%, while the OEM division lost US$1,853 million and profit margin of -27.7%.

However, Bank of America has limited this strength: data center CPUs do benefit from the wave of demand brought about by intelligent (agentic) AI, but they are constrained by tight supply of substrates, etc., and the product technology has not caught up with ultra-micro semiconductors and ARM architecture rivals that are growing 2 to 4 times faster.

Intel's foundry revenue was US$6.697 billion, up 58% year on year; of this, external foundry revenue was about US$410 million (J.P. Morgan's 411 million, Bank of America 406 million), and the month-on-month increase mainly came from a non-recurring customer project. The OEM segment's operating loss narrowed to $1.9 billion ($2.3 billion in the same period last year). The gross profit margin of 44.1% was about 200 basis points higher than market expectations of 42.1%, thanks to reduced costs and improved product structure for Intel 3 and 18A; adjusted earnings per share of $0.52, significantly higher than the consensus of $0.40 to $0.41.

The only shortcoming was the client: CCPG's revenue was $8.097 billion, down 8.8% month-on-month, and J.P. Morgan said storage shortages weighed down the overall market size. J.P. Morgan also noted that 18A client CPUs already account for about one-third of client revenue (less than 5% in the first quarter).

In terms of cash, the third quarter had cash flow of 5.08 billion US dollars from operating activities, about 4 billion US dollars in capital expenses, and about 54 billion US dollars in cash and equivalent at the end of the quarter. There were no dividends or repurchases during the quarter.

The fourth quarter guidance is the one that crashed the market

The problem lies in the outlook. Intel's fourth-quarter revenue guidance was $15 billion to $16 billion (median value of $15.5 billion), about 9% lower than the $17.05 billion market consensus; the adjusted earnings per share guide was $0.25, far below the consensus of $0.43 to $0.44.

The drag mainly comes from clients: J.P. Morgan said CCPG dropped by more than 20% month-on-month, while Bank of America more specifically estimates that PC shipments fell 25% to 30% month-on-month. Management downgraded expectations for the PC market size to fall 12% in 2026 and another 10% in 2027, and expect OEM inventory revisions to continue until at least the first quarter of 2027. DCAI is expected to increase by more than ten percent month-on-month, which can only be partially hedged.

The gross margin guide was 38.0%, which is also far below the market threshold of 42.8%, falling about 600 basis points from month to month; J.P. Morgan said the decline was roughly shared by falling production, diluting fixed costs, deteriorating product structure, and Israel's influence.

Two additional variables: Israel Fab 28 and inventory cycle

The temporary shutdown of the Fab 28 plant in Israel was included in the guidelines, resulting in a fourth-quarter revenue impact of about 300 million US dollars and a gross margin shock of about 200 basis points (focusing on 10 nm server products). J.P. Morgan Chase expects another impact of about 300 million US dollars in the first quarter, but most of the waste loss will roll out, and gross margin is expected to return to around 40% in the first quarter.

The two companies were both cautious in their judgment on inventory: J.P. Morgan said that customers were aggressively reducing inventory weeks, and the correction continued until the first quarter of 2027; Bank of America extended the revised forecast until the first half of 2027.

A year of money spent: capital expenditure reached a high level, free cash flow turned negative

The forward-looking investment cycle has accelerated markedly. Management said that capital expenditure in 2027 will increase “strongly” from the 2026 $20 billion baseline. J.P. Morgan believes that more than $26 billion is a reasonable assumption (the same Bank of America model is 26 billion and 34 billion in 2028); equipment orders increased 40% month-on-month, Arizona's Fab 62 is progressing at an accelerated pace, and Ohio is also exploring further acceleration.

As a result, free cash flow for 2027 was confirmed as negative — but the two companies gave different sizes of the gap: around $6.6 billion for the J.P. Morgan model and negative $2.6 billion for the Bank of America model. J.P. Morgan also pointed out that the 23 billion US dollar equity financing in August was clearly positioned to support this round of investment.

OEM: Costs and yield are improving, but external customers are still unimplemented

Progress at the executive level is the agreement of the two companies: Intel 18A costs have dropped by about 40% since the first quarter, yield is better than planned, and Intel 3 costs have been reduced by about 15%; the 0.9 PDK for 14A was released as scheduled; 18A-P is being launched on schedule by the end of the year, and “several highly strategic customers” are being contacted; more than 10 long-term CPU agreements have been signed, including the disclosed ORCL.US (ORCL.US).

The difference is whether these are enough. The direct reason the Bank of America downgraded its rating is that the stock price “has not conveyed sufficient progress from external OEM customers.” Its foundry valuation is still contextual — assuming Apple M-series wafers, MediaTek TPU packages, Terafab's IP and package, and potential ARM architecture server CPUs are launched one after another, with external wafer revenue of 40 billion US dollars and advanced packaging of 6 billion US dollars by 2030, for a total foundry revenue of 46 billion US dollars, corresponding to 8% to 10% of the approximately $380 billion foundry market. Bank of America stressed that until the details of the 18A-P/14A transaction are clear, its foundry benchmark model will not be adjusted.

Target price: Komoro and Bank of America all have different methods and time limits

J.P. Morgan's $80 target price (period December 2027) comes from approximately 25 times the price-earnings ratio, multiplied by annualized earnings per share when exiting 2028:4 times its forecast of $0.79 per share for the fourth quarter of 2028, or $3.16. Earnings forecasts were revised downgraded at the same time — revenue/earnings per share fell from $63.3 billion/$1.58 billion to $61.9 billion/ $1.48 in 2026, from $77.7 billion/$2.44 billion to $67.2 billion/ $1.68 billion in 2027, and adjusted earnings per share fell 31% in 2027.

Bank of America's $100 target price is another algorithm: based on the 2030 earnings potential of $5.89 per share (previously $6.14), a price-earnings ratio of 20 times, then a 9% discount rate for two years. The Bank of America split this target price into a market value of about 550 billion US dollars — a product business of about 450 billion US dollars (roughly equivalent to half of the market value of ultrafine semiconductors; the two server CPUs overlap but the latter has an additional GPU business), plus about 100 billion US dollars in foundry business; referring to TSMC's market sales rate of about 11 times that of 100 billion US dollars, 100 billion US dollars implies external business of 8 billion to 10 billion US dollars. The target price period of the two companies is different from the valuation anchor, and it does not constitute a comparison of highs and lows in the same dimension.

Notably, the two companies' downgraded earnings forecasts for 2027 ($1.68 and $1.69) were highly close, and both fell short of market consensus — the consensus quoted by J.P. Morgan Chase was $2.12, and Bank of America tied for Bloomberg's $2.12 and Visible Alpha's $2.09; the two companies were $2.82 and $2.79 in 2028, which is also below the $3.19-$3.20 consensus. J.P. Morgan's explanation for the neutral rating is that risk and return are relatively balanced: team execution has indeed improved since Chen Liwu became CEO. In the future, it is necessary to track data center CPU roadmaps, AI computing power strategies, external OEM customer contacts, and ASIC design order reserves.

The upward risks listed by Bank of America are key external OEM packaging/wafer orders, 18A and 14A yield climbing better than expected, PC being stronger than expected, and geopolitics driving local manufacturing sentiment; downside risks are lower than expected yields, lack of substantial external wafer customers, further weakening of PCs, and accelerated loss of share.

Spillover: AMD is under pressure, equipment stocks benefit

Bank of America has given a clear cross-sector transmission judgment: PC weakness will drag down AMD (AMD.US), which will announce results next week. The PC side will drag down up to $750 million, partially offset by strong server CPU revenue of more than $500 million — the consensus's expectation for AMD's client PCs in the fourth quarter was flat, while Intel fell 25% to 30%; however, the rise in the AMD HeliOS/MI series may form a strong hedge. The strength of Intel's data centers is a positive sign for Nvidia (NVDA.US), Broadcom (AVGO.US), and Mwell Technology (MRVL.US); capital expenditure intentions (growing in 2027, and rising share of equipment) are beneficial to semiconductor devices, especially KLAC.US (KLAC.US), Applied Materials (AMAT.US), and Fanlin Group (LRCX.US).