Earlier this month, Intel reported its Q2 2024 results, missing both top and bottom-line expectations, and provided weaker-than-anticipated guidance for the next quarter. More concerning, however, were the disappointing gross margin figures, the announcement of a suspended dividend, and plans to cut 15% of its workforce. These measures are part of Intel’s strategy to turn around its manufacturing business and deleverage its balance sheet. As a result, Intel experienced its largest one-day price drop on record, falling 26.1%.
In this note, we highlight the StarMine models that accurately predicted these developments and provided early warning signals about the heightened credit risk and deleveraging efforts.
Weak Buy-Side Sentiment
The StarMine Smart Holdings (SH) model, which measures buy-side sentiment by predicting future changes in institutional buying and selling, showed a sharp decline for Intel in July, with its score dropping to 15 (regional percentile ranking out of 100). Following the earnings release on August 2nd, the score further plummeted to 3 (Exhibit 1). For the past year, buy-side sentiment for Intel has ranged from weak to neutral, with the last top-decile ranking occurring in May 2022.
Exhibit 1: StarMine Smart Holdings – Intel
Source: LSEG Workspace
Exhibit 2 illustrates why Intel may have failed to meet the investment criteria of Portfolio Managers and Analysts. The SH model uses 25 fixed factors across seven categories, and Intel ranked in the bottom quintile for 14 of these factors. Categories such as Profitability, Growth, and Leverage scored particularly poorly, providing clear early warning signs to investors. Intel’s forward 5-year EPS CAGR currently stands at 2.1%, compared to 13.1% for TSM and 28.5% for Nvidia.
Intel also received a score of 1 in the Analyst Revisions category, as analysts sharply revised FY1 and FY2 estimates downward by 75.3% and 40.5%, respectively, post-results. No other semiconductor company globally has experienced such steep downward revisions over the past 30 days.
Exhibit 2: StarMine Smart Holdings Screening Factors – Intel
Source: LSEG Workspace
We can further corroborate buy-side sentiment by examining sell-side sentiment using the Analyst Revision Model (ARM), which is designed to predict future changes in analyst revisions based on revisions to EPS, EBITDA, Revenue, and Recommendations over multiple time periods. Heading into the earnings release, ARM showed an increase in the model score, reaching a neutral score of 42. However, this score plummeted to 1 after analysts reacted to the earnings release (Exhibit 3).
This example underscores the value of considering both buy-side and sell-side sentiment when assessing the overall sentiment toward a company. In this scenario, the SH model provided a more accurate prediction compared to ARM, signaling that institutional investors were likely to avoid Intel well before the Q2 release.
Exhibit 3: Analyst Revision Model (ARM) – Intel
Source: LSEG Workspace
Predicting a Suspension in Dividends
Intel’s decision to suspend its dividend could have been partially anticipated by closely examining its fundamentals. The Earnings Quality (EQ) model, designed to measure the reliability and sustainability of a company’s earnings, has consistently signaled concerns over the past year. As shown in Exhibit 4, Intel’s EQ score has remained in the bottom decile since October 2022, with its current score of 3.
Exhibit 4: Earnings Quality – Intel
Source: LSEG Workspace
In 2021, Intel announced its intention to reenter the semiconductor chip manufacturing market, competing with industry giants like Taiwan Semiconductor Manufacturing (TSMC) and Samsung Electronics. This strategic shift requires substantial capital expenditure and investment in new property, plant, and equipment (PP&E).
Exhibit 5 highlights the impact of this decision on Intel’s EQ components, which include Accruals, Cash Flow, Operating Efficiency, and Exclusions. Notably, both PP&E and Capex are significantly above the industry median, while operating margins fall below it.
Exhibit 5: Earnings Quality Components – Intel
Source: LSEG Workspace
Intel’s free cash flow (Cash Flow from Operations minus Capex) has been negative for the last two fiscal years, with figures of -$9.6 billion in 2022 and -$14.3 billion in 2023. According to LSEG I/B/E/S estimates, this trend is expected to continue, with a forecasted FCF of -$18.1 billion in FY1 and -$6.1 billion in FY2. Intel’s struggle to convert EBITDA into free cash flow will continue to pressure its liquidity.
Furthermore, Intel’s return on invested capital (ROIC) is projected to be -0.1% this year, 1.4% next year, and 3.1% the following year, significantly lagging the industry average of double-digit returns. The company’s net debt (short-term plus long-term debt minus cash) is estimated at $35.3 billion this year, the second highest in the industry.
For those using LSEG Workspace, dozens of pre-populated charts within the EQ model are available by clicking the ‘Chart’ icon, as illustrated in the image above. Intel is expected to spend $61.6 billion on Capex over the next three years, with $25.8 billion allocated for this year alone. The Capex-to-revenue ratio for Intel has reached an all-time high of 47.5%, significantly exceeding the industry median of 5.7% (Exhibit 6).
For context, the only semiconductor company expected to outspend Intel on Capex over the next three years is TSMC, which is forecasted to spend $102.7 billion. However, TSMC is projected to grow revenues by 22% annually over this period, compared to Intel’s 4% per annum.
Exhibit 6: Capex vs. Revenue – Intel
Source: LSEG Workspace
Additionally, Intel’s gross margins have been in a secular decline over the past decade, with both operating and net margins sharply decreasing in the last two years, significantly underperforming compared to industry peers. In Q2, Intel reported a gross profit margin of 38.7%, well below analyst expectations of 43.4%—the largest miss in eight quarters.
Exhibit 7: Gross, Operating, and Net Margins – Intel
Source: LSEG Workspace
StarMine Implied Rating Suggests Further Downgrades from Rating Agencies
StarMine Credit Risk models are designed to detect credit default risk in the form of a technical default or bankruptcy over a 12-month period.
The StarMine Combined Credit Risk model is an optimal combination of our three credit risk models (Structural, SmartRatios, Text Mining), which each monitor different phenomena to detect an impending credit default. While a strongly bearish view from one individual credit risk model might be enough to influence the overall combined score, further downgrades from the other models can help strengthen that conviction.
CCR assigns a probability of default, which is mapped to a Model Implied Rating, which can be used to compare against rating agencies such as S&P and Moody’s.
In Exhibit 8 below, the gold line represents Intel’s share price, the green line shows S&P’s rating, and the blue line reflects the StarMine Implied Rating.
Since March 2024, the StarMine Implied Rating (blue line) has downgraded Intel twice, from “BBB” to “BB+,” prior to the company’s Q2 earnings release. Meanwhile, the S&P rating (green line) has remained unchanged at “A-” during this period. Following the Q2 results, the CCR model further downgraded Intel four times, bringing the rating down to “B,” while S&P has still not adjusted its rating. Our research indicates that when the StarMine Implied Rating significantly diverges from an agency rating, there is a 4-5 times higher likelihood that the agency will revise its rating towards the StarMine Implied Rating.
Exhibit 8: Combined Credit Risk – Intel
Source: LSEG Workspace
Intel currently has the lowest CCR model score among its peers, indicating the highest probability of default, as shown in Exhibit 9. The StarMine SmartRatios Credit Risk model utilizes common financial ratios, but with the special sauce of also including forward-looking StarMine SmartEstimates®. StarMine SmartEstimates improve upon the consensus estimates in that more weight is given to sell-side analysts with a better track record of accurately predicting financials with the timeliest estimates.
Exhibit 9: Combined Credit Risk – Intel vs. Peers
Source: LSEG Workspace
Reducing Headcount by 15%
As part of its deleveraging strategy, Intel announced plans to reduce its workforce by 15%. To put this in perspective, we used the Excel add-in from LSEG Workspace to analyze Intel’s employee count and compare it to sales as a measure of efficiency.
In Exhibit 10, which compares Intel with its U.S. peers, we find that Intel has one of the lowest efficiency ratings in the group. Despite having the largest workforce by far, with 124,800 employees (as shown in the bottom right of the graph), Intel’s sales per employee value stands at $441,037.
For comparison, Nvidia, Broadcom, and Advanced Micro Devices have significantly higher sales per employee values, at $5.0 million, $2.9 million, and $1.2 million, respectively. Even with a 15% workforce reduction, Intel will still have 106,080 employees—more than twice the number of the next largest company, Qualcomm, which has 50,000 employees.
Exhibit 10: Sales Per Employee – U.S. Semiconductor Group
Conclusion
Intel is facing significant challenges in its turnaround efforts, with disappointing results on both the top and bottom lines, reduced guidance for the next quarter, and a substantial miss on gross margin expectations. Several StarMine models effectively identified early warning signs ahead of the earnings release, highlighting areas of concern. The StarMine Combined Credit Risk model suggests the possibility of further credit rating downgrades, which could increase the pressure on Intel’s ability to manage its debt obligations moving forward.
Get unique value-add analytics and predictive financial modeling, dedicated to making investment research smarter with LSEG StarMine data.
LSEG Workspace is a complete solution for research and analytics. It places the most comprehensive market information, news, analytics and trading tools available into a desktop.
