CEO of AI $1.5 Billion Startup Accused of Massive Fraud by Justice Department

The US Department of Justice has charged the CEO and CFO of iLearning Engines, a $1.5 billion AI startup, with fabricating virtually all customer relationships and revenues since 2019, a case that exposes how institutional investors can be deceived during sector booms when due diligence focuses on market narrative rather than customer verification. The fraud underscores a systemic vulnerability in venture funding and public markets when valuations outpace auditable fundamentals, particularly in nascent technology sectors where investor appetite exceeds scrutiny.

• iLearning founder Harish Chidambaran received $500 million in common stock plus $12.5 million in restricted stock units despite faked revenues.
• The company reported $421 million in 2023 revenue entirely fabricated through sham contracts with non-existent or shell customers.
• FBI data shows 22,000 AI-related fraud complaints in 2025 with estimated losses near $900 million, up 33 percent year-over-year.

• $1.5B iLearning Engines valuation at time of fraud allegations and executive arrests.
• $421M Reported 2023 revenue, entirely fabricated through sham enterprise customer contracts.
• 22,000 AI fraud complaints filed in 2025, far exceeding prior year complaint volume.

Federal prosecutors have charged iLearning Engines founder and CEO Puthugramam “Harish” Chidambaran and Chief Financial Officer Sayyed Farhan Ali “Farhan” Naqvi with securities fraud, wire fraud, and conspiracy as part of what the US Department of Justice characterizes as a “continuing financial crimes enterprise.” According to the DoJ, the two executives fabricated customer relationships and revenue streams across the company’s entire operational history beginning in January 2019, exploiting investor enthusiasm for artificial intelligence technologies to raise capital and inflate their personal wealth.

Chidambaran was arrested in Maryland and Naqvi in California following the unsealing of federal charges that allege the pair constructed an “intricate web of sham contracts with purported customers” to simulate enterprise adoption and recurring licensing revenue.

Chidambaran Extracted Over $500 Million in Stock While Revenue Was Entirely Fabricated

The scale of personal enrichment alleged in the case is substantial. Chidambaran accumulated over $500 million in common stock holdings in addition to $12.5 million in restricted stock units and a $700,000 annual salary paid during 2023 and 2024, according to federal documents.

These compensation packages were awarded based on financial projections and revenue figures that prosecutors say were entirely fraudulent, meaning investors and company boards made equity and salary decisions using false operational data.

The revenue figures iLearning reported to investors and lenders were comprehensive fabrications. In 2023 alone, the company claimed $421 million in annual revenue purportedly derived from AI platform licensing agreements with enterprise customers.

According to the DoJ, none of these customer relationships existed as represented; instead, the executives created fictitious or shell customer contracts in amounts ranging from tens of millions of dollars each to create the appearance of a rapidly scaling enterprise software business riding the AI wave.

The prosecution alleges that both executives directly marketed iLearning to investors and lenders using these fraudulent financial statements. “As alleged, the defendants exploited investor excitement over the AI boom and presented a rosy financial outlook to investors and lenders that was built on lies,” the DoJ statement reads.

“While the defendants pitched iLearning as a way to revolutionize training and education through AI, the truly artificial part of the defendants’ story was iLearning’s customers and revenues.”

Venture Investors Relied on Financial Statements Rather Than Customer Verification

The iLearning case exposes a recurring weakness in institutional investment processes during periods of sector euphoria.

Venture capital firms, growth equity investors, and lenders conducting due diligence on technology companies in hot markets often prioritize revenue growth rates and market positioning over granular customer verification, particularly when those customers are claimed to be large enterprises under confidentiality agreements.

This dynamic becomes more acute in software and AI markets where customers are often large institutions themselves, creating an asymmetry of information that executives can exploit.

A venture partner cannot easily call fifty enterprise customers to verify contracts when those customers claim confidentiality; instead, audited financial statements and management representations become the primary evidence.

When executives control both the financial reporting infrastructure and customer relationship documentation, the pathway to fraud becomes substantially easier than in asset-based or transaction-heavy businesses where third parties provide independent verification.

The $1.5 billion valuation iLearning achieved reflects how far an AI company can scale on narrative alone when investors are competing for allocation in a sector perceived as transformative.

Venture funding rounds and secondary market valuations in the AI space have repeatedly outpaced revenue verification, creating conditions where fraudulent claims about customer adoption can justify extraordinary share valuations before they are exposed.

AI Fraud Complaints Reached 22,000 Cases in 2025 With Losses Exceeding $900 Million

The iLearning case is not an isolated incident but part of a broader pattern of fraud spanning the AI sector. According to the FBI’s latest Internet Crime Report, over 22,000 complaints related to AI fraud were filed in 2025 alone, representing a 33 percent increase from the prior year.

Estimated losses from these complaints total approximately $900 million, a figure that understates actual institutional losses because many large-scale schemes involving private companies like iLearning are prosecuted through securities fraud statutes rather than classified as Internet Crime Complaint Center (IC3) complaints.

This trend reflects two structural factors. First, the genuine commercial promise of AI technologies has created a broad investor base with limited technical expertise, making it easier to deploy fraudulent claims about AI capabilities or customer traction.

Second, the speed at which AI companies scale during boom periods creates time pressure for due diligence; venture firms and growth equity investors often conduct compressed investment processes when competing for access to companies they believe are category-defining.

The iLearning fraud also demonstrates that regulatory arbitrage persists between private and public markets: private companies can operate with less disclosure rigor than public companies, allowing executives to sustain false financial statements through multiple funding rounds before any external audit or customer audit occurs at IPO or acquisition stage.

What Institutional Investors Should Verify Before Allocating Capital to AI Companies

The federal charges against Chidambaran and Naqvi offer concrete lessons for institutional capital allocators. The DoJ’s allegations center on the fact that iLearning’s customer base was substantially or entirely fabricated despite representing the core of the company’s valuation thesis.

This suggests that due diligence processes that rely primarily on financial statements and management interviews without independent customer verification are insufficient in sectors where execution risk is high and management incentives to inflate metrics are acute.

For institutional investors conducting due diligence on AI companies, several verification steps become critical. Direct customer interviews with named accounts, particularly large enterprise deals, should be standard practice rather than optional.

Customer concentration analysis should flag any single customer representing more than 10-15 percent of revenue; iLearning’s alleged scheme created numerous customer relationships worth tens of millions of dollars annually, which should have triggered verification questions.

Reference checks should include customers’ own business outcomes and implementation timelines, which would immediately contradict claims of massive enterprise adoption of a platform that is barely operational.

Equity compensation verification also matters: executives receiving hundreds of millions of dollars in stock options tied to revenue growth should trigger heightened scrutiny of revenue recognition policies and revenue quality rather than revenue quantity alone.

Federal prosecutors have not yet announced sentencing dates for Chidambaran or Naqvi, and the case will likely proceed through trial unless plea agreements emerge; institutional investors should monitor whether any venture capital firms, growth equity funds, or venture debt providers file claims against their directors and officers insurance policies for negligence in due diligence, as such filings would establish clearer liability standards for investor conduct in AI sector allocations.