The Numbers Don't Add Up — And That's the Point
MINIMAX reported revenue of $117 million, up 283.1% year-over-year. Gross profit surged 464.8%. Losses narrowed 11% to $358 million. On paper, this is a growth story that would make any venture capitalist salivate. But I've been auditing protocol financials since 2017, and I've learned one immutable truth: ledgers do not lie, only their auditors do.
Let me walk you through what this financial statement actually reveals when you strip away the narrative gloss. Because the gap between the headline numbers and the operational reality is not a small discrepancy — it's a chasm.
Context: The AI Video Generation Landscape in 2026
MINIMAX operates in the most capital-intensive corner of the AI economy: video generation. The Hailuo series, their flagship video model, competes directly with OpenAI's Sora, Google's Veo, and China's domestic challengers like Kuaishou's Kling and ByteDance's Jimeng. This is not a niche — it's a war zone where compute costs run at ten to fifty times the cost of text-based inference.
The company listed on the Hong Kong Stock Exchange, which means it faces international investor scrutiny and disclosure requirements that are more demanding than a private round. Yet the financial statement reveals remarkably little about the operational fundamentals. Revenue, gross profit, net loss — three numbers, and then silence. No R&D breakdown. No customer concentration metrics. No cash position disclosure. For an analyst trained to read between the lines, this selective disclosure is itself a data point.
Based on my audit experience with Layer 2 protocols, I've learned that the most important numbers are often the ones a project chooses not to disclose.
Core Analysis: The Unit Economics Tell a Different Story
Let me calculate what the disclosed figures actually imply. Revenue of $117 million against gross profit of $20.8 million yields a gross margin of approximately 17.8%. To put that in perspective, mature SaaS companies typically operate at 70-80% gross margins. Even capital-intensive cloud providers like AWS maintain margins above 60%. A 17.8% gross margin means that for every dollar of revenue, over 80 cents is consumed by direct costs.
What are these direct costs? For an AI video generation company, the answer is almost certainly compute — GPU depreciation, cloud infrastructure, electricity, and network bandwidth. Video inference is computationally brutal. Generating one minute of high-definition video can require thousands of GPU inference calls. This is the fundamental structural problem with the AI video business model: the cost of goods sold scales nearly linearly with usage, and there is no path to the kind of margin expansion that software investors expect.
Now, let's consider the growth trajectory. Gross profit grew 464.8% — faster than the 283.1% revenue growth. This suggests unit economics are improving, which could come from better model architecture, more efficient inference, or a shift toward higher-margin enterprise contracts. But here's the critical caveat: even with this improvement, the gross margin remains catastrophically low. The improvement is from a base of roughly 8-10% to 17.8%. That's progress, but it's progress from "unsustainable" to "still deeply problematic."
The net loss of $358 million is the more alarming figure. That's 3.06 times the revenue. The company is burning roughly $3 for every $1 it brings in. At this pace, the annualized burn rate approaches $700 million. Unless MINIMAX has a war chest of several billion dollars, it will need to return to the capital markets within 12-18 months — at terms that will likely be less favorable than its last round.
Here's the pattern I recognize from auditing DeFi protocols during the 2020 bull run: growth that requires continuous capital injection to sustain is not growth — it's a Ponzi scheme with a longer time horizon. I'm not accusing MINIMAX of fraud. But the structural economics — high growth, low margins, massive losses — mean the company's survival depends entirely on its ability to raise ever-larger rounds at ever-higher valuations. If the capital markets close, the company closes.
The efficiency question is central. In my 2020 stress tests of Aave and Compound, I discovered that the protocols' reserve factors were too slow to adjust for volatility. The lesson was simple: efficiency isn't just about speed — it's about survival. MINIMAX's 17.8% gross margin is not a sign of operational excellence; it's a sign that the company is buying market share at an extraordinary cost.
The Contrarian Angle: What the Selective Disclosure Reveals
Here's where I diverge from the consensus take. The market will likely view MINIMAX's revenue growth as validation of the AI video category. I see something different: a company that is strategically withholding information to maintain its narrative.
Notice what the financial statement doesn't disclose. No R&D expense breakdown — critical for understanding whether the technical moat is widening or narrowing. No customer concentration metrics — essential for assessing revenue quality. No cash position — the single most important number for a company burning $700 million annually. No user metrics — no way to distinguish between sustainable subscription revenue and one-time enterprise deals.
This pattern of selective disclosure is not unique to MINIMAX. I've seen it repeatedly in crypto projects during their ICO phases. The 2017 EtherFund audit I conducted taught me this lesson: when a project emphasizes its best metrics while obscuring its operational fundamentals, the hidden numbers are usually worse than the disclosed ones.
Consider the competitive pressure. OpenAI's Sora and Google's Veo are not standing still. If MINIMAX's technical advantage narrows — and in AI, advantages narrow quickly — its growth rate will decelerate precisely when its cost structure demands continued expansion to achieve any path to profitability. The company is running a race where the finish line keeps moving, and the track is made of burning capital.
The deeper issue is what this means for the AI video category as a whole. MINIMAX's 283% revenue growth is being cited as evidence that AI video is a viable commercial category. But a 17.8% gross margin suggests the category's unit economics are fundamentally broken. This is the "efficiency-ethics friction" I've documented throughout my career: the industry is prioritizing growth at the expense of sustainability, and the bill will come due.
Takeaway: The Vulnerability Forecast
Code is law, but human greed is the bug. MINIMAX's financials reveal a company running a capital-intensive race against time. The real question isn't whether MINIMAX can grow — it's whether the AI video category can ever achieve sustainable unit economics.
My forecast is cautious. Over the next 12-18 months, we will see one of two outcomes: either MINIMAX achieves a dramatic gross margin improvement through technical innovation (unlikely but possible), or the capital markets will force a reckoning through down-rounds, dilution, or strategic fire sales. The company's growth rate is impressive, but growth without margin is just a slower form of death.
Yield is the interest paid for ignorance. In this case, the yield is the revenue growth, and the ignorance is the assumption that growth alone constitutes a viable business model. I've audited enough protocols to know that when the fundamentals are this weak, the narrative is always strongest. And narratives, unlike ledgers, always eventually break.
The smart money will watch the gross margin trajectory, not the revenue headline. If MINIMAX can't push margins above 30% within two quarters, the growth story will collapse under the weight of its own cost structure. Until then, I'm watching the numbers — and the numbers don't support the narrative.
We build bridges in the storm, not after the rain. MINIMAX's storm is coming, and the question is whether they've built a bridge strong enough to cross it.
