Writing Team
The three biggest AI bets in history are about to meet Wall Street. And the man running the fourth is cheering them on while quietly planning his own exit.
- Perplexity CEO Aravind Srinivas confirmed a 2028 IPO target regardless of how Anthropic or OpenAI perform on public markets
- Anthropic confidentially filed for an IPO last week at a valuation approaching $1 trillion; OpenAI filed the same week
- Srinivas called the listings "important for the AI industry" and predicted they'll go well, because "they're doing well"
- He also warned: six months without a frontier model capability advance would be a serious problem for both companies
- A new trend called "tokenmaxxing" is emerging, where employees inflate AI usage to signal productivity rather than generate value
What Happened
Perplexity CEO Aravind Srinivas told CNBC his company is targeting a 2028 IPO, independent of what happens when Anthropic and OpenAI hit public markets. This follows Anthropic's confidential IPO filing last week at a valuation near $1 trillion, and OpenAI's confidential filing the same day Srinivas spoke. SpaceX is also listing this week, making the next month one of the most consequential IPO windows in tech history.
Srinivas said he believes the Anthropic and OpenAI listings "will go well, because they're doing well." He also acknowledged, without sugarcoating, that if they don't go well, there will be "ripple effects."
Why This Should Make You Nervous
There is a version of this story that is inspiring: the frontier AI companies are so clearly valuable that institutional investors are lining up to own a piece of history. Srinivas is basically saying that version is true.
But listen to what he actually said as a warning: "If for six months you don't see a model capability advance from one of these two companies, then it's a problem for them." That's not a confidence statement. That's a treadmill. A trillion-dollar company whose valuation depends on never slowing down is not a stable investment thesis. It's a pressure cooker.
The other thing worth sitting with is a new word he introduced: "tokenmaxxing." Employees are reportedly increasing their AI use not because it produces better work, but because high usage signals productivity. This is the enterprise equivalent of leaving your office light on so your boss thinks you're working late. The fact that this is already a named phenomenon, weeks before the companies behind it go public, is a detail that deserves more than a passing mention.
What AI's IPO Parade Means for Marketers and Growth Leaders
Marketing teams have been absorbing vendor pressure to buy AI tools for two years. The IPO wave will intensify that pressure considerably. When Anthropic and OpenAI have public shareholders to answer to, the pitch cycle gets louder, not quieter.
That matters if you're a CMO or growth leader trying to separate real utility from performance theater. The tokenmaxxing pattern Srinivas described isn't just an enterprise curiosity. It's what happens inside your team when AI adoption becomes a KPI instead of a problem-solving tool. High usage numbers look good on a slide. They don't tell you whether the work got better.
Our advice: set outcome metrics before your team starts attributing everything to AI. The companies going public are selling capability. Your job is deciding which capabilities actually move your numbers. If you want help building an AI adoption framework that measures real output rather than token counts, our growth consulting team can help you think through the actual ROI question.
The Part Nobody Wants to Say Out Loud
A trillion-dollar valuation for a company whose moat depends on continuous frontier model advancement is, at minimum, a hypothesis. The venture capital world has been stress-testing that hypothesis with private money. Now it's about to test it with yours.
Srinivas may be right that these IPOs will go well. SpaceX's listing this week will tell us a lot about whether retail and institutional investors are still in a mood to buy the future on installment. But "they're doing well" is not due diligence.
The AI industry needs these listings to succeed, and many of the people saying so have a financial interest in that outcome. That doesn't make them wrong. It does mean you should read the prospectus when it drops, and not just the press release.
If your organization is still figuring out how to build an AI strategy that doesn't depend on vendor hype cycles, the A-Eye Spy archive is a good place to start. We cover the story without the cheerleading.
Want growth strategy that doesn't hinge on a $1 trillion bet? Talk to Winsome Marketing.