OpenAI 想把公司股份分给公民，像阿拉斯加发钱一样。这事要是成了，可能是 AI 公司第一次让普通人直接分钱。
OpenAI 提出向美国政府转让 5% 股权，意图让公民直接受益，类似阿拉斯加石油分红。对比挪威 1.9 万亿美元主权基金、新加坡 GIC 等，多数主权财富基金收益用于政府预算而非个人。OpenAI 在 2025 年亏损 210 亿美元，收入 130 亿美元，5% 股权当前估值约 426 亿美元，但距离盈利仍需 5 年以上。该提议能否落地、分红形式（UBI 或政府收入）均未确定。
If this actually pays out to citizens, it could be…
If this actually pays out to citizens, it could be a first for AI.
The only real precedent is Alaska, where resource wealth turns into direct cheques. Residents get an annual dividend from oil royalties invested on their behalf.
Pretty much every other sovereign wealth fund works differently. Norway's $1.9 trillion fund, Singapore's GIC, Saudi Arabia's PIF, the UAE's MGX - they recycle returns into government budgets and public services, not people's pockets.
OpenAI's proposal is a genuine shift. Hand the government equity in the company, explicitly framed as something that should benefit citizens. But there's no guarantee the US adopts an Alaska-style payout.
So what does it become - a UBI mechanism, or just another revenue stream for general spending? Alaska writes cheques. Here, there's no promise of anything that direct.
Timing matters as well. OpenAI reportedly lost $21 billion in 2025 on $13 billion in revenue. Training costs and data centres are eating cash, and profitability still looks a long way off. Any meaningful dividends are years away - five or more wouldn't surprise me. So nobody should expect near-term payments.
Still, it's better to debate this now than once the money is already flowing. Build the structure before the profits arrive, rather than scrambling to retrofit it later. This administration has roughly two and a half years before the 2028 election - whether it gets implemented now or kicked to the next one is anyone's guess.
Then there's scale. Five per cent of OpenAI at today's valuation is about $42.6 billion. Is that remotely enough to cushion long-term job displacement by then?
To clarify what 5% likely means - we're guessing here based on how equity stakes typically work, nothing's confirmed. It would probably be 5% ownership of the company itself, meaning 5% of any profits or dividends OpenAI eventually pays out. Whether that requires OpenAI to go public first, or if they could structure payouts while staying private, I don't know. The details haven't been released.
Obviously OpenAI's valuation by the time they're actually profitable could be vastly different. It might go up significantly if things go well, making that 5% stake worth far more. It could also tank if they stumble, or the company might not exist at all. But if the bet pays off, you'd expect that slice to grow substantially.
For comparison, the UAE has already raised $50 billion for AI investment via MGX alone. The bill for widespread disruption could easily dwarf what's on the table here.
What do you think - is this the right model, or do we need something completely different?