AI Risks and Super Micro: Unraveling Financial Reporting Issues
Written on
Chapter 1: Overview of Super Micro's Financial Challenges
Super Micro Computer (SMCI) saw its stock plunge by 20% on Wednesday following news of a delayed 10-K filing and a damning report from Hindenburg Research, which accused the company of accounting irregularities involving related parties. This decline represents a staggering loss of over 63% from its peak of more than $1,200 per share in mid-March.
The delay in filing, as noted in SMCI’s 8-K, was attributed to "additional time needed for management to complete its assessment of the design and operational efficiency of its internal financial reporting controls." This raises serious concerns, especially given the context of the deadline.
Financial reporting issues aren't new territory for Super Micro. The company has previously faced delisting, and in 2020, the SEC charged it with "channel stuffing," which refers to prematurely recognizing revenue while understating expenses. This manipulation was executed by recognizing revenue on undelivered goods, shipping without proper authorization, and sending misassembled products.
The critical aspect here is that AI identified increasing risks in financial reporting as early as 2022, risks that were largely overlooked at the time.
The recent stock drop has shattered the previous complacency. Now, SMCI’s board, executives, auditors, and regulators are under scrutiny regarding their credibility.
Section 1.1: The Role of AI in Identifying Financial Risks
In capital markets, transparency is crucial. Red flags prompt necessary inquiries. Properly designed AI tools tailored for specific domains can expedite this process.
For example, Hudson Labs created pioneering software driven by finance-specific large-language models (LLMs) to streamline equity research and derive actionable insights. Their tools analyze over 800,000 SEC filings annually, processing millions of pages and billions of words.
One significant screen used by Hudson Labs highlighted that SMCI ranked among the top 100 companies with the highest related-party risk. They discovered that “related party risk isn’t the only concern at SMCI.” The company's forensic risk score has been elevated since 2019.
Hudson Labs CEO Kris Bennatti stated, “We first flagged transactions at SMCI resembling ‘round tripping’ in 2022.” This term refers to using related party transactions to artificially inflate sales figures.
In February 2022, Andre Castillo, the head of forensic research at Hudson Labs, shared some revealing findings in a research blog, including: Related party purchases from Ablecom and Compuware made up 9.4% of cost of sales in 2021, an increase from 7.3% in 2020. Notably, Compuware functions as both a distributor and manufacturer for SMCI.
Complicating matters, SMCI’s CEO Charles Liang has familial ties to Ablecom; his brother, Steve, is its CEO and owns approximately 29% of the company. Both Charles Liang and his wife, Sara Liu, who is also a board member at SMCI, jointly own over 10%. This nearly 40% control across the three raises questions about the actual value of such transactions, especially since Ablecom sells components back to SMCI at the same price it paid.
Bennatti responded to inquiries about the materiality of these transactions, remarking, “In most cases where related party transactions have been associated with misconduct, the disclosed transactions were less significant than those here. The ability to influence suppliers and customers creates opportunities that may not be reflected in the financial statements, presenting additional risks.”
Section 1.2: Controversies and Regulatory Scrutiny
The Hindenburg report intensified the scrutiny surrounding SMCI, including references from legal filings and claims from anonymous former employees, alongside additional disclosure analyses. Among the report's critical points is the SEC's 2020 fine of $17.5 million, which implicated former CFO Howard Hideshima. Although CEO Liang faced no charges, he was required to repay the company $2.1 million in stock profits per Sarbanes-Oxley regulations. Interestingly, Hideshima has recently appeared as a consultant for Ablecom.
Hindenburg, which revealed its short position in SMCI, pointed out that the related parties appear to engage in minimal other business: 99.8% of Ablecom’s exports to the U.S. since 2020 were directed to Super Micro, and 99.7% of Compuware’s exports were also to Super Micro, according to trade records.
Analysts, including J.P. Morgan’s Samik Chatterjee, viewed Hindenburg’s report as lacking substantial details regarding alleged misconduct. However, he noted, “It’s unsurprising that the company has areas for improvement in governance, transparency, and communication with investors, especially given its recent growth alongside rising AI server demand,” as reported by Investors Business Daily.
CFRA Research analyst Shreya Gheewala downgraded SMCI’s rating to hold and reduced its price target from 729 to 454, arguing, “While the evidence does not definitively indicate significant accounting malpractice or sanction evasion, SMCI's delayed 10-K filing and potential reputational harm raise concerns.”
This turmoil has exposed SMCI's precarious valuation, which was previously hidden in plain sight. The eventual outcomes will certainly provide lessons for others in the industry.
Chapter 2: The Impact of AI on Financial Governance
The first video titled "Is the AI Dream Over? Super Micro Computer (SMCI) Faces Major Allegations and Stock Plunge!" explores the recent controversies surrounding SMCI and how AI has played a role in uncovering these financial issues.
The second video titled "AI is Ruining My Life: Group Therapy for Security Leaders" discusses the broader implications of AI on security and governance in organizations, shedding light on the challenges faced by leaders in navigating these complexities.
Hudson Labs has demonstrated that current AI advancements can significantly enhance the detection of financial risk signals. So, what will break the cycle of human inertia?
The governance section on SMCI’s website lists a nine-member board led by founder Liang, alongside co-founder Liu. The documents include a generic audit committee charter and a business conduct code emphasizing “full, fair, accurate, timely, and understandable disclosures.” However, this rhetoric starkly contrasts with reality.
External watchdogs seem indifferent. The only “critical audit matter” flagged by Deloitte in the last annual report concerned excessive and obsolete inventory reserves. As major public accounting firms invest billions in AI, shouldn't investors expect better, more thorough, and prompt assurance? This raises important questions.
Is any of this truly unexpected? How did governance mechanisms fail? Are regulators taking notice? Where were the auditors during these events? Are these challenges merely a result of the soaring stock market enthusiasm? What defines a “publicly-owned” enterprise?
Will the trifecta of incentives, incompetence, and indifference prevail?