Financial Gravity: Why OpenAI, MiniMax, Manus Are Collapsing — Eric Ries' Framework

What Is Financial Gravity

On June 10, Eric Ries — author of The Lean Startup — hosted an AMA on Hacker News (536 upvotes, 431 comments). He introduced a concept he calls "financial gravity": the invisible pull that drifts good companies away from their original mission. Ries argues this drift isn't caused by a sudden decision to do evil, but by organizational structure slowly dragging them there.

Ries identifies three forces of financial gravity:

Short-termism — quarterly earnings pressure forces abandonment of long-term value.

Money's gravity — the more capital a company raises, the easier its founding purpose shifts.

Structural fragility — without a structure designed to resist gravity, even great people cannot hold the line.

In his new book Incorruptible , Ries studies companies that successfully resist gravity. Costco kept its $1.50 hot-dog combo unchanged for 15 years. Patagonia donates 1% of sales to the planet. Novo Nordisk, a century-old Danish pharma, became the global diabetes leader. Their common trait isn't a visionary CEO but an inherently anti-gravity structure.

OpenAI: The Idealist's Fall

Financial gravity explains OpenAI's trajectory perfectly. Founded in 2015 as a non-profit, open, capped-profit entity with the mission "to ensure AGI benefits all humanity," it attracted the world's top AI researchers. Then gravity pulled:

2019: Created a capped-profit arm, accepted $1B from Microsoft. Reason: AI is too expensive for a non-profit model. This was gravity's first tug — once money enters, a new operating logic takes over.

2023: Sam Altman fired and reinstated, exposing governance fragility.

2024: Began converting from non-profit to fully for-profit. Models closed. API prices rose. Microsoft ties deepened. The "AI for humanity" narrative became a commercial "race to AGI" story.

Ries' AMA mention that he helped design Anthropic's governance signals a structural vacuum in AI companies. OpenAI solved technical problems but never seriously addressed structural ones; when gravity hit, no structure held it.

MiniMax: Snap Collapse

Some gravity pulls slowly; some snaps instantly. MiniMax, a Chinese LLM star backed by Hong Kong capital with a ~$300B valuation, had strong products in music generation and good developer rapport. Then it abruptly raised API prices with no notice or transition. Developers revolted.

Through the gravity lens: when capital's return expectations collide with pricing reality, user trust is the first sacrifice. MiniMax isn't the first or last, but its speed and direct cost validate Ries' observation: "There's a dark side in our industry we rarely discuss. I've seen too many good companies drift in various ways." The pricing move may have had commercial logic, but the way it drifted decided the outcome. A trusted company destroyed that trust with an unannounced pricing shift. Gravity doesn't need time — it just needs a trigger.

Manus: Free Fall

Manus's story is shorter and more ironic. An AI Agent product that went viral overnight, promising to be everyone's automated assistant, was then exposed for registering users' domains in its own name and refusing to return them. A product built to automate everything performed the most non-automated, opaque act: registering domains silently, denying it, then stonewalling. This is both a technical and a structural failure.

Manus's gravity took a different shape: explosive growth outpaced operational and trust infrastructure. When the gap between product promise and delivery widened, trust went into free fall. As the article states: "Free fall needs no structure; resisting gravity does."

Eric Ries' Solution

Financial gravity is a diagnostic framework; Ries also offers a cure. In Incorruptible he details structures that resist gravity. A standout AMA exchange: when asked whether Costco's low prices rely on leadership rather than structure, Ries replied directly — if it's only leadership, what happens when the CEO leaves? Who chooses the successor? How to ensure the successor is right? These are structural questions, not personal-quality questions.

His core thesis: you cannot bet on every CEO being a Jim Sinegal (Costco's legendary CEO). You must design a structure that preserves values even under a mediocre CEO.

LTSE (Long-Term Stock Exchange) — lets companies operate around long-term value without quarterly Wall Street reporting.

Anthropic governance design — concrete practice of his theory in the AI industry.

For AI founders, Ries did what few have: he explained why you'll fall and gave a method to build a parachute.

Questions for AI Product Managers

The article turns Ries' challenge to the reader: you've thought about MVP, PMF, growth loops — but have you thought about anti-gravity structure? When users grow from 10K to 1M, what mechanism prevents you from harming them? When investors demand next-quarter doubling, what pulls your product decisions off course?

Most AI startups die from nothing. But the ones that survive die from financial gravity.

Ries closed his AMA: "I may not have figured out all the answers, but I've spent more time on why good companies go bad than my health probably allows." A framework honed over decades is now being validated by real-world cases. That alone deserves every AI practitioner's serious attention.