Ed Elson and tech reporter Alex Heath spent a recent segment of Prof G Markets arguing about the same set of facts and reaching almost opposite conclusions. Elson thinks OpenAI is a company drifting sideways while leaks about an IPO pile up. Heath thinks the leaks are the story, and most of them are being planted by people with something to gain. <\/p>\n
Elson’s read is blunt. OpenAI, he says, “doesn’t really have its sh-t together,” and he points to a stack of decisions that fit that story: a confidential S-1 filing, a series of reversals on side bets, and a reported $200 million podcast acquisition that looks a lot like scope creep. If you are a retail investor waiting to buy shares one day, that is the version of OpenAI you should probably worry about. A company on the verge of the largest tech listing in history should look like it is narrowing focus, not expanding into audio. <\/p>\n There is context for the anxiety. Anthropic filed a confidential S-1 on June 1, 2026, and its private valuation ran from $380 billion in March to $965 billion by late June. SpaceX has already gone public and carries a market cap around $2.11 trillion, though the S&P Dow Jones committee declined to fast-track it into the index. If OpenAI is the marquee AI IPO, every quarter of drift is a quarter its two biggest peers use to define the story. <\/p>\n Act now: the analyst who called NVIDIA in 2010 just named his top 10 AI stocks \u2014 and Google didn’t make the cut. Grab the names FREE today<\/a>. <\/p>\n The media noise case <\/p>\n Heath’s rebuttal is that most of what you are reading is manufactured. “This IPO thing is really being ginned up in the media by investors,” he argues, and the only concrete fact on the table is the confidential filing itself. OpenAI’s own line, he notes, is that “we do not have a sense of timing.” Anonymous leaks pointing to fall 2025 or 2027 windows are, in his framing, investors messaging each other through reporters. He calls the situation “very dynamic” and would “be shocked if it was in the second half of 2027.” <\/p>\n Story Continues <\/p>\n You can hear the full exchange on Prof G Markets. Heath’s underlying claim is more interesting than the timing debate: the IPO calendar is a place where private-market shareholders lobby for liquidity in public, and the press is the megaphone. <\/p>\n What actually moves the timeline <\/p>\n Both agree the listing happens eventually. Sam Altman personally wants it, and the volume of private capital that has already been raised makes a public float close to inevitable. The variables Heath flags are the ones worth tracking: Anthropic’s ARR trajectory, whether OpenAI’s super app finds a user base, the ad model, hardware, and executive turnover. <\/p>\n The talent signal is already moving. On June 23, 2026, two senior AI researchers left Alphabet (NASDAQ:GOOGL<\/a>), with John Jumper going to Anthropic and Noam Shazeer, a co-author of the “Attention Is All You Need” paper, going to OpenAI. Alphabet shares fell 6%. That move erased roughly $250 billion in market value. That data point quietly resets IPO math, because talent flow is a live read on which platform frontier researchers think will matter in three years. <\/p>\n Retail access is the other subplot. The KraneShares AGIX ETF (NYSEARCA:AGIX<\/a>) holds a roughly 1.4% stake in Anthropic, and Webull (NASDAQ:BULL<\/a>) is rolling out private-market SPVs with Monark Markets for accredited investors. Neither gets you OpenAI directly, but both are the plumbing being built for the day it does list. <\/p>\n Who is right <\/p>\n Probably both, in the annoying way that these debates usually resolve. OpenAI can be a company that is genuinely scattered and a company whose IPO chatter is being managed. On SpaceX as a cautionary tale, Heath’s counter is worth sitting with: trading at $1.16 trillion is not obviously a bad outcome given the revenue underneath it. The question for investors is which signal you weight, the drift or the noise, while you wait for a prospectus that will eventually settle the argument. <\/p>\n Act now: the analyst who called NVIDIA in 2010 just named his top 10 AI stocks \u2014 and Google didn’t make the cut. Grab the names FREE today<\/a>. <\/p>\n<\/a> Deemerwha studio \/ Shutterstock.com <\/p>\n