Perplexity: Company Plans IPO in 2028 Regardless of Anthropic and OpenAI's Listing Progress Perplexity：无论Anthropic与OpenAI上市进展如何，公司均计划2028年IPO

Perplexity’s CEO dropped a reassurance pill in an interview: IPO in 2028, regardless of what Anthropic and OpenAI do. The statement sounds tough, but on closer inspection, it seems more like boosting his own courage.

Frankly, Perplexity’s timing for this announcement is quite subtle. OpenAI has already secretly submitted its IPO filing, Anthropic is frequently in the capital market spotlight, and the entire AI community is buzzing about the race to "go public first." At this moment, the CEO steps up to say, "We’ll follow our own pace." On the surface, it appears strategic composure, but deep down, anxiety is hard to hide—after all, when it comes to valuations, the first to list sets the benchmark, and later entrants can only price themselves based on others’ favor.

However, Srinivas has bet on one thing correctly: the AI search space is far from being a settled battle. Perplexity’s greatest advantage isn’t how powerful its technology is, but the awkward yet genuine niche it has carved between Google and OpenAI—users want answers, not ten links to click. This demand isn’t glamorous, but it’s solid. An IPO in 2028 means he still has a four-year window to prove that the search paradigm can truly be disrupted, not just patched over by big tech.

That said, four years in the AI field is almost like a geological era. A company valued at billions today might be crushed tomorrow by some open-source model. By betting on 2028, Perplexity is wagering that it can survive as a small but profitable cash cow in the giants’ shadow, rather than being dragged into a money-burning war and obliterated. This gamble, honestly, is even more thrilling than the technical roadmap itself.

Looking closer to home, China’s A-share semiconductor sector surged today. Inventmicro rose over 7%, with GigaDevice and Eoptolink following suit. On the surface, it looks like a chip rally, but the underlying logic is actually a spillover from the global AI arms race—you need computing power for large models, chips for computing power, and the narrative of domestic chip substitution will always find a market. But such rallies often mean that by the time retail investors pile in, they’re already catching the tail end of the story.

What’s truly interesting is the trending headline: "ChatGPT’s Biggest Ever Update—It’s Not Just for Chatting." If true, OpenAI is essentially transforming ChatGPT from a conversational tool into an operating system. If this move succeeds, it means the AI battlefield shifts from "who’s smarter" to "who has a deeper ecosystem." But the risk is also obvious—Apple once thought the same way, and Siri hasn’t made meaningful progress in a decade. Ambitions to turn tools into platforms have historically seen very few success stories.

One easily overlooked piece of information is: "After adopting AI, a company stopped hiring." This is the AI issue that deserves the most discussion in 2025. The cost-cutting efficiency axe of enterprises will inevitably hit the job market sooner or later. Perplexity’s CEO talks about IPOs, OpenAI discusses platformization, and the media focuses on valuations and growth curves—but who talks about those who are being "optimized away"? The dividends of technological progress have never been evenly distributed, and our discussion on this issue remains stuck on comforting but empty platitudes like "AI will create new jobs."

At the end of the day, today’s AI scene is like a game of Texas Hold’em—Perplexity is pretending to stay calm, OpenAI is going all-in, China’s semiconductor sector is riding the hype, while ordinary users and practitioners can only watch the table, guessing who the real players are and who are just leeks waiting to be harvested.

Perplexity的CEO在采访里扔出一颗定心丸：2028年IPO，不管Anthropic和OpenAI怎么折腾。这话听着挺硬气，但仔细琢磨，更像是给自己壮胆。

说实话，Perplexity选这个时间点表态挺微妙的。OpenAI那边IPO文件都秘密提交了，Anthropic也在资本市场频繁露脸，整个AI圈都在讨论"谁先上市"这场竞赛。这时候CEO跳出来说"我们按自己的节奏来"，表面看是战略定力，骨子里难保没有焦虑——毕竟估值这东西，先上市的就是锚，后来的只能看别人脸色定价。

但斯里尼瓦斯赌对了一件事：AI搜索这个赛道，现在还远没到盖棺定论的时候。Perplexity最大的优势不是技术多牛，而是它在Google和OpenAI之间找到了一个尴尬但真实的缝隙——用户想要答案，不想点十个链接。这个需求不性感，但很实在。2028年上市，意味着他还有四年窗口期来证明搜索范式真的可以被颠覆，而不是被大厂的一个补丁就抹平。

不过话说回来，四年在AI领域几乎等于一个地质年代。今天估值几十亿的公司，明天可能就被某个开源模型降维打击。Perplexity押注2028，赌的是自己能在巨头的夹缝里活成一只小而美的现金奶牛，而不是被卷入烧钱大战粉身碎骨。这个赌注，说实话比技术路线本身更刺激。

把视线拉回国内，A股半导体板块今天集体嗨了。盈方微涨超7%，兆易创新、联特科技跟涨。表面看是芯片行情，但深层逻辑其实是全球AI军备竞赛的外溢效应——你造大模型需要算力，造算力需要芯片，芯片国产化的叙事就永远有市场。只是这种涨法，散户追进去的时候，往往已经是故事的后半段了。

真正有意思的是热榜上那条"ChatGPT史上最大改版，不只聊天"。如果这条消息属实，OpenAI正在做的事情本质上是把ChatGPT从一个对话工具变成一个操作系统。这步棋走对了，意味着AI的战场从"谁更聪明"转向"谁的生态更厚"。但风险也明摆着——苹果当年也是这么想的，结果Siri十年没长进。工具变平台的野心，历史上成功的案例屈指可数。

还有一条容易被忽略的信息："用上AI后，公司停止招人了。"这才是2025年最该被讨论的AI议题。企业降本增效的刀，迟早会砍到就业市场头上。Perplexity的CEO谈IPO，OpenAI谈平台化，媒体谈估值和增长曲线，但谁来谈谈那些被"优化"掉的人？技术进步的红利从来不是平均分配的，而我们对这个问题的讨论，依然停留在"AI会创造新岗位"这种正确但空洞的安慰剂上。

说到底，今天的AI圈就像一场德州扑克——Perplexity在强装镇定，OpenAI在全押，国内半导体在蹭热度，而普通用户和从业者只能看着牌桌，猜测谁是真玩家，谁是待宰的韭菜。