HP Called Autonomy a Fraud. A Jury Just Suggested HP Was Mostly Fooling Itself.

In August 2011, HP agreed to pay £25.50 a share for a British software company most Americans had never heard of — a 64% premium over the prior day's close, unanimously blessed by both boards.⁹ By October it controlled 87% of Autonomy's shares.² Fourteen months after that, it told the world the whole thing had been worth roughly nothing: an $8.8 billion impairment, most of it pinned on fraud it said had happened before HP ever signed.¹¹ One of the most admired companies in technology had bought a $11.1 billion asset and discovered, almost immediately, that it could not find the company it thought it had purchased.

The official story is clean and comforting: Autonomy cooked its books, HP was the innocent victim, the fraud was the disaster. Almost every load-bearing word of that has since cracked. HP's own court filings show it projected $7.4 billion in synergies that never existed. Its due-diligence function reported to the wrong executive. And in 2024 a jury looked at the central villain of the story and acquitted him of everything.

The $5.3 billion that was never fraud at all

Here is the number that quietly dismantles the victim narrative, and it came from HP, not its critics. In litigation, HP's own filing revealed that it had modeled $7.4 billion in revenue synergies from owning Autonomy — the extra sales it expected from cross-selling enterprise search into HP's vast install base. Ex-Autonomy executives called the figure 'unprecedented,' and they were not flattering it. When those sales failed to materialize, the gap between the fantasy and the reality accounted for $5.3 billion of the $8.8 billion writedown.⁵ Sit with that ratio. More than half of the famous fraud charge was not fraud. It was HP marking down its own optimism. You cannot accuse a seller of stealing the value you invented after the deal closed.

Who was guarding the gate, and who did they answer to

A deal this size does not collapse on the strength of a bad model alone. It needs the model to go unchallenged — and at HP, the architecture for challenging it was pointed the wrong way. Meg Whitman, who inherited the mess, said publicly that under Léo Apotheker's leadership HP's due-diligence function reported to the Chief Strategy Officer rather than the CFO. "I've never seen that before and it's a change I made right away," she said.¹² Read past the corporate phrasing and the failure is structural and damning: the people whose job was to find reasons not to do the deal answered to the executive whose job was to champion it. That is not a system designed to surface red flags. It is a system designed to digest them. When the skeptics report to the salesman, the deal always looks better than it is.

The jury that reopened a closed case

For a decade the fraud narrative had a legal spine: a 2018 San Francisco conviction. Autonomy's CFO, Sushovan Hussain, was convicted on 16 counts of wire and securities fraud, sentenced to five years, and — after losing his appeal — began serving that sentence in August 2020.¹⁰ That conviction is real, and it is why the accounting-manipulation question remains genuinely unsettled rather than dismissed. But the spine of HP's story was never the CFO — it was Mike Lynch, Autonomy's founder and CEO, the man HP spent thirteen years and untold millions casting as the mastermind. In June 2024 a federal jury in the same city acquitted Lynch on all fifteen counts: fourteen of wire fraud, one of conspiracy. His co-defendant, the VP of Finance, was cleared of everything too.⁶ HP's $5 billion 'accounting improprieties' figure was always its own characterization, asserted in a press release, not adjudicated in a courtroom. After 2024, the courtroom had spoken — and it had not agreed.

But surely the fraud was real — wasn't a CFO convicted?

The honest counter deserves its full weight, because the revisionist take can be sold too cleanly the other way. Hussain's 2018 conviction was not overturned by Lynch's acquittal; two juries can look at related conduct and reach different verdicts about different people, and both can be right. Something was wrong with how Autonomy recognized certain revenue — that is the most defensible reading of the record, and it is why the accounting question stays open rather than resolved in Lynch's favor. So the steelman for HP is this: it walked into a target whose books were, at minimum, aggressively presented, and a thinner-margined enterprise software business than HP believed it was buying. Fair. But notice what that steelman cannot rescue. Even if every dollar of the alleged accounting manipulation were proved, it could explain at most the portion of the writedown not already accounted for by HP's own synergy gap — leaving $5.3 billion that HP conjured itself, ran through a diligence team reporting to the deal's own champion, and then blamed on someone else.⁵ A victim of fraud does not get to bill the fraudster for its own bad math.

The counterfactual hiding inside the 2015 split

Three years after the writedown, HP did the thing the Autonomy deal had been meant to prevent. On November 1, 2015, the company split itself in two: HP Inc., the PCs-and-printers business, kept the old HPQ ticker; Hewlett Packard Enterprise — the servers, software, services, and the smoldering remains of Autonomy — became HPE.⁸ The Autonomy acquisition had been a bet that HP could buy its way into high-margin enterprise software and become more than a hardware company. The split was the formal admission that the bet had failed. Run the counterfactual: had HP's diligence reported to its CFO, had it stress-tested its own $7.4 billion synergy number instead of Autonomy's revenue, the deal might have closed at a sane price — or not at all. The same governance reflex that misfired in 2011 is what eventually forced the company to cut itself in half. The disaster was never really the fraud. It was a company that had stopped being able to tell its own optimism from someone else's lie.

HP spent thirteen years prosecuting the theory that it had been robbed. The jury's verdict, the synergy filing, and the org chart all point somewhere less flattering and more useful: a great company can wreck itself without anyone defrauding it, simply by wanting a deal badly enough to build the case for it out of its own assumptions — and then routing the people paid to doubt it under the people paid to want it. The most expensive thing HP bought in 2011 was not Autonomy. It was the proof that the dangerous fraud is the one you tell yourself.