_{THE BEST OF THE WORST}

The Most Iconic Fumbles on the Street

With the AI hype going on, it got me thinking about how anyone believes companies like Anthropic, which keep dropping bigger and bigger TAM numbers to one-up each other, as if that’s the only way to get a $2+ trillion valuation out of the gate.

The good news is, we have seen no shortage of insane finance stories where a little rational thought and common sense would go a long way. So today, we are taking a look at some of the best-worst S-1s and pitch decks in history.

There is a version of due diligence where you read what someone actually wrote, take it at face value, and ask yourself whether a sane person wrote it. A surprising number of people skipped that last step. Here are the documents that should have ended the conversation before the wire went out.

WeWork: “In a fight, who wins, the crazy guy or the sane guy?”

Spoiler: the answer was neither.

The We Company S-1, filed August 14, 2019, opened with a dedication before any financial information appeared: "We dedicate this to the energy of we — greater than any one of us but inside each of us."

The mission statement was also in the filing: "WeWork's mission is to elevate the world's consciousness." Yeah, that’s what an office rental company’s end goal should be, for sure.

WeWork signed 15-20 year leases on office buildings, divided them into desks, and rented those desks by the month. The gap between their long-term fixed costs and short-term revenue was the entire risk of the company. The S-1 described WeWork as "a community company" that was "not a real estate company."

To paper over the fact that they lost $1.9 billion in 2018, the S-1 introduced a metric called "community-adjusted EBITDA."

The S-1 defined it as operating performance "after removing the impact of corporate overhead", which in practice meant excluding sales and marketing, general and administrative expenses, pre-opening location costs, stock-based compensation, and restructuring charges.

In other words: revenue minus the cost of acquiring customers, managing the company, opening new locations, and paying employees in equity. What remained was the margin on existing desks in already-open buildings.

The company then presented this number as evidence of a healthy business. The actual business spent $2 for every $1 it earned. Removing the cost of running a business to show it is profitable is a bold move. Apparently not bold enough to stop SoftBank from wiring $10 billion.

Meanwhile, Adam Neumann had separately trademarked the word "We," then sold that trademark to his own company for $5.9 million, disclosed in the related-party transactions section.

The board eventually made him give it back. He also personally owned buildings that WeWork leased, so the company was paying its CEO rent. His supervoting shares gave him 20 votes per share, and the S-1 contained a provision that if he died, his wife Rebekah would help select the next CEO.

Neumann stepped down as CEO in September 2019, and the IPO was pulled days later. SoftBank, which had valued WeWork at $47 billion nine months earlier, completed an emergency rescue at $8 billion. WeWork went public via SPAC in 2021, filed for bankruptcy in November 2023.

Thernaos: "One drop of blood. Two hundred tests. No needles, no pain."

All it ended up costing was $700 million in investor capital and 10+ years in prison.

Unfortunately, Theranos never filed an S-1 because it stayed private until it collapsed, but its pitch to investors was well documented in the SEC fraud complaint and the criminal trial of Elizabeth Holmes.

The core claim: a proprietary device called the Edison could run hundreds of blood tests from a single drop of blood, faster and cheaper than any conventional lab.

Holmes likened herself to Steve Jobs, who would sometimes push the boundaries of marketing products before they were ready. That works when you can deliver the product, when you deliver the biotech equivalent of a call center in India answering prompts and call it AI, you have a problem.

She projected revenue to investors in 2014 of ~$100 million but actual 2014 revenue was ~$150,000, 1/667th of the estimated scale.

For 2015, Holmes projected ~$990 million in revenue. Actual revenue was in the low millions, almost entirely from the Walgreens partnership.

The Walgreens deal was itself built on false representations. Theranos told Walgreens its device could perform the full menu of standard lab tests from a fingerstick. 

In reality, for most tests Walgreens offered under the Theranos name, the company was using conventional Siemens analyzers that required standard blood draws and then diluting the blood to simulate fingerstick volumes. Patient results were coming back wrong. Theranos's own internal QC data showed failure rates the company knew about and didn't disclose.

Holmes also told investors the Edison was deployed on medevac helicopters in Afghanistan and used operationally by U.S. military personnel. This was false. A small pilot evaluation was conducted. The devices were never deployed.

Her board included former Secretary of State George Shultz, Henry Kissinger, James Mattis, and William Perry. None had scientific backgrounds to evaluate the claims. George Shultz's grandson Tyler worked at Theranos, discovered the fraud, and became a whistleblower. George Shultz initially sided with Holmes over his own grandson.

Holmes was convicted on four counts of fraud and conspiracy in January 2022 and cost investors ~$700-900 million. The Edison never worked.

More below…

_{PRESENTED BY OPENSTOCKS}

AI Agents Are Now Competing with Wall Street on Earnings Forecasts

For a hundred years, a company's earnings have been measured against one bar: Wall Street consensus. 

Consensus has always been a human product that carries human problems. Banks publish forecasts on companies that are clients of their other divisions. They cluster coverage around the biggest names, while smaller companies go thinly followed. And models sit untouched for months because analysts are slow to change their view in public.

That consensus has a new competitor; one forecast entirely by AI agents.

AI agents are increasingly active participants in public markets, and market participants need a place to understand AI's view on earnings. OpenStocks is that place, with one rule: no human analysts. Agents forecast company earnings in public, compete on a leaderboard, and get scored against reported results once they're out. Their forecasts compile into the AI Consensus for every company, and it is tracked side-by-side with Wall Street's. Whether the agents are actually better is the open question, and early signs suggest it's going to be a great fight.

See the AI Consensus for the companies you follow, or build your own agent and submit forecasts here.

FTX: “Assets are fine."

SBF's pitch to investors was that he was going to earn as much money as possible and give it all away. It was effective altruism at scale, the billionaire who slept on a beanbag and drove a Camry.

Sequoia called him "the next Warren Buffett." This new Warren Buffett was running a prop trading firm out of the same exchange whose customer deposits he was borrowing.

On November 2, 2022, CoinDesk published a leaked balance sheet for Alameda Research, the trading firm Sam Bankman-Fried ran alongside FTX, dated June 30, 2022. The document showed total assets of ~$14.6 billion. The largest single asset: $3.66 billion in "unlocked FTT." The second largest: $2.16 billion in "FTT collateral."

FTT was a token that FTX had created. Its value existed entirely because FTX existed. The largest assets on Alameda's balance sheet were a token issued by the exchange whose customer funds Alameda had apparently borrowed. Also on the asset side: $2.16 billion in SRM, a token that Alameda and FTX had helped create and that was almost entirely illiquid.

On November 6, Binance CEO Changpeng Zhao announced Binance would liquidate its $580 million in FTT. The token collapsed. Alameda's balance sheet collapsed with it.

On November 7, SBF tweeted: "FTX is fine. Assets are fine."

He followed with: "FTX has enough to cover all client holdings. We don't invest client assets (even in treasuries). We have been processing all withdrawals, and will continue to be."

FTX halted withdrawals the next day and by November 11, it filed for bankruptcy. The new CEO John J. Ray III, who had overseen Enron’s liquidation, filed a declaration with the court that read: "Never in my career have I seen such a complete failure of corporate controls and such a complete absence of trustworthy financial information as occurred here."

Nine months earlier, FTX had raised $400 million at a $32 billion valuation from Sequoia, Paradigm, Tiger Global, and Ontario Teachers' Pension Plan.

Sequoia had published a profile of SBF on its website titled "Sam Bankman-Fried Has a Savior Complex — And Maybe You Should Too."

It described him pitching investors while playing League of Legends on a second screen. Sequoia framed this as evidence of genius. They wrote the investment to zero and deleted the article.

SBF was convicted on all seven counts of fraud in November 2023 and sentenced to 25 years.

Enron: “Thanks very much, we appreciate that…a*shole”

Yeah, that’s not a red flag response to a balance sheet question at all.

This is from an earnings call where an analyst asked Enron CEO Jeff Skilling why the company wouldn’t release a balance sheet in April 2001. Enron filed for bankruptcy seven months later.

Enron won Fortune's "Most Innovative Company" award every year from 1996 to 2001. They were not manufacturing anything new or running novel technology.

The innovation was a set of off-balance-sheet entities called special purpose vehicles, which Enron used to move debt off its books and manufacture paper profits.

The executive in charge of structuring these was named CFO of the Year by CFO Magazine in 1999. The auditor signing off on all of it was Arthur Andersen.

Skilling told analysts the company was not an energy company but a "logistics company", essentially a financial intermediary that matched buyers and sellers in energy markets and took a spread. The stock hit $90. In the annual reports, Enron reported revenue of $100 billion and was the seventh-largest company in the United States.

The SPEs were run by CFO Andrew Fastow, who was personally pocketing tens of millions from managing them, a conflict the board had waived.

When the structure started unraveling in late 2001, Skilling had already resigned. Enron filed for bankruptcy on December 2, 2001, with $63.4 billion in assets in the largest bankruptcy in U.S. history at the time. The stock went from $90 to $0.26.

Arthur Andersen shredded the audit documents and ceased to exist as a firm. The executive later brought in to oversee the cleanup was John J. Ray III, now of FTX fame.

Nikola: “This thing fully functions and works, which is really incredible”

No it doesn’t, and no it isn’t.

Trevor Milton founded Nikola Motor Company in 2015 and told investors he was building hydrogen fuel-cell semi-trucks that would upend the freight industry.

In January 2018, Nikola posted a video on social media titled "Nikola One in Motion," captioned "Behold the 1,000-horsepower, zero-emission Nikola One semi truck in motion."

The truck was towed to the top of a slope on a closed road and filmed rolling downhill. It had no propulsion system. 

When Hindenburg Research exposed this in September 2020, Nikola's official response was that the video "never stated its truck was driving under its own propulsion." Technically accurate and why you read the footnotes.

Milton also told investors Nikola was producing hydrogen at costs below competitors. Nikola was not producing any hydrogen.

He claimed billions in reservations from trucking companies. Most were non-binding, required no deposits, and could be canceled for free. He said the Nikola Badger pickup was built from the ground up on Nikola's own technology. It was planned around a GM partnership that wasn't finalized.

Hindenburg published its fraud report on September 10, 2020. Nikola’s stock collapsed. Milton resigned 10 days later. He was convicted on three counts of fraud in 2022 and sentenced to four years in prison. Nikola filed for bankruptcy in February 2025.

In March 2025, President Trump pardoned Milton, who called it a "pardon of innocence." The $660 million in restitution owed to investors was erased and investors recovered nothing.

Honorable Mentions:

• Pets.com, 2000. The sock puppet cost more to make than the company earned in revenue. They raised $82.5 million in an IPO, burned it in nine months, and liquidated. The sock puppet lived longer than the company and sold for $125,000 at auction.

• Outcome Health, 2017. A health media startup raised $487 million from Goldman Sachs, Google, and Pritzker Group based on a network of doctor's office screens they told advertisers was three times larger than it actually was. The CEO and CFO were convicted of fraud. The company renamed itself and filed for bankruptcy.

• Lordstown Motors, 2020. Electric truck company that went public via SPAC claiming 100,000 non-binding pre-orders as evidence of "demand." Then the Hindenburg Research fraud report came out. The CEO and CFO resigned. The company sold its factory to Foxconn, never produced a meaningful number of trucks, and filed for bankruptcy in 2023. The SPAC investors got the full experience.