← back to scans
Will Goldman Sachs serve as lead-left underwriter on OpenAI's IPO?
48%
Agent
73%
Market Price
-24.5%
Edge
42%
Confidence
10,112
Volume
5.0c
Spread
442
Days to resolution
6
Markets in event
final rationale
Mean of 2 draws (0.52, 0.45) after one crux research pass. (0) The condition is not met. EDGAR shows no public OpenAI S-1 as of 2026-10-16, only a confidential draft submitted in June 2026, and no credible outlet has named the lead-left bank. (a) About 14.5 months remain until 2028-01-01. Altman has ruled out a 2026 listing, so an IPO, including a public S-1 at least 15 days before the roadshow, has to happen in 2027. (b) Status quo is No: if nothing happens, there is no IPO and the market resolves No. (c) No path: OpenAI raises a private round at around $1.4T, revenue runs about $20B below expectations, and the listing slips into 2028. Or Morgan Stanley wins lead-left, or a joint top billing triggers the partial-payout tie rule. (d) Yes path: the IPO prices in 2027 as CFO Friar guided. Goldman takes the top-left slot, building on its SpaceX lead-left win, while Morgan Stanley takes Anthropic, which the FT reports it is frontrunner for. Kalshi's 73% looks like it prices mostly the conditional lead-left question; the 'Goldman is in the S-1' claims come from weak aggregator sites. I estimate P(IPO before 2028) at about 0.65 and P(Goldman sole lead-left given an IPO) at about 0.70–0.75, which gives roughly 0.47. I shade that toward the market anchor and land at 0.52.
price vs. prediction
market price
agent forecast at t+0
3 follow-ups
1. Planner Opus 5.5 · picks the research tools
Tools chosen:
kalshi_direct
resolution_source
kalshi_related
polymarket_related
article_search
Planner reasoning
The Kalshi price on this market is the anchor. Sibling markets on other banks for lead-left, plus OpenAI IPO timing markets on Kalshi and Polymarket, give the joint probability that an IPO happens before 2028 and that Goldman gets the lead-left spot. The resolution source would be the final prospectus on SEC EDGAR. One targeted search covers any reported bank mandates for OpenAI's IPO.
2. Research parallel tool calls + 4 web research loops (Sonnet 5.5) · read raw by the forecasters
at a glance
RESEARCH DIGEST (built in code: tool summaries and each report's bottom line; the forecasters read the full tool outputs and reports)
Tools
- kalshi_direct: ## This Market's Kalshi Data **Goldman Sachs** (KXOPENAILEADLEFT-28JAN01B-GSX) - Current price (probability): 73.00% (live; bid 73% / ask 78%) - 7-day price change: -6.00% - 30-day price change: +8.00% - Average daily volume: 161 contracts - Price range: 25.00% - 83.00% - Data points: 39 days
- resolution_source: already_resolved=unclear; matches_criteria=True; last_updated=2026-10-16 (EDGAR search as of today); Forge article ~2026-09-24. As of Oct 16, 2026, EDGAR has no OpenAI S-1 ('No matching companies'), so there is no public or final prospectus with a lead-left underwriter on its cover. Secondary reports say OpenAI filed confidentially around May–June 2026 and is working with Goldman Sachs and Morgan
- kalshi_related: 1 related markets / summaries. series KXOPENAILEADLEFT: 0 markets (skipped 6 no-signal) | series KXOPENAIIPO: 0 markets (skipped 0 no-signal) | keyword 'OpenAI IPO': ok
- polymarket_related: Polymarket search over 1 keyword(s), kept 8 matches. keyword 'OpenAI IPO': 8 markets
- article_search: Fetched 10 unique articles across 1 queries (lookback=120d). 'OpenAI IPO banks hired underwriters Goldman Sachs Morgan Stanley lead-left': 10 articles
Research loops, bottom lines
- current state: Goldman and Morgan Stanley are the reported co-leaders, and no source confirms who is lead-left. A public S-1 cover will not exist until at least late 2026, and management now points to 2027. Two timing risks could stop resolution as Yes: an IPO slipping past 2028-01-01, or a shared top billing, which would give Goldman a partial payout under the rules.
- base rate: The reference class is two banks, Goldman and Morgan Stanley, with Goldman winning lead-left on the closest comparables (Tesla 2010 and SpaceX 2026). SpaceX also showed that relationship-based favourites can lose. Outside-view risk to YES is mostly Morgan Stanley winning the order (market snapshots suggest about 20–25%) plus the chance the IPO slips past 2027 (guided for 2027, but not firm).
- yes case: The strongest YES path is that Goldman wins the tie-break against Morgan Stanley on OpenAI, because the FT reports Morgan Stanley leading for Anthropic, and Goldman's SpaceX lead-left win is the key precedent. This requires an OpenAI IPO pricing in 2027 and a single bank, not a joint pair, on the cover. No source confirms the order.
- no case: The strongest NO case has three parts. The IPO has no scheduled date and could miss 2028-01-01. The lead-left order is unreported. Anthropic reportedly going to Morgan Stanley leaves Morgan Stanley as a credible OpenAI lead-left too, and the claims that Goldman is already in the S-1 come from weak aggregators.
Crux follow-up
- The evidence supports a 2027 IPO as the stated plan. The 2027 date is Friar's internal guidance, though, and Altman has set no firm date, so slippage past 2027 is plausible. On lead-left, Goldman's SpaceX win makes it a plausible favourite. Reputable reporting through September still describes the OpenAI order as contested, and Morgan Stanley is the FT's frontrunner on Anthropic. The market's high Goldman price rests mostly on low-quality claims about a confidential S-1 and appears to ignore IPO
tools
| tool | status | time | items | summary |
|---|---|---|---|---|
| kalshi_direct | OK | 0.2s | 1 | ## This Market's Kalshi Data **Goldman Sachs** (KXOPENAILEADLEFT-28JAN01B-GSX) - Current price (probability): 73.00% (live; bid 73% / ask 78%) - 7-day price change: -6.00% - 30-day price change: +8.00% - Average daily volume: 161 contracts - Price range: 25.00% - 83.00% - Data points: 39 days |
| resolution_source | OK | 9.9s | 1 | already_resolved=unclear; matches_criteria=True; last_updated=2026-10-16 (EDGAR search as of today); Forge article ~2026-09-24. As of Oct 16, 2026, EDGAR has no OpenAI S-1 ('No matching companies'), so there is no public or final prospectus with a lead-left underwriter on its cover. Secondary report |
| kalshi_related | OK | 1.1s | 1 | 1 related markets / summaries. series KXOPENAILEADLEFT: 0 markets (skipped 6 no-signal) | series KXOPENAIIPO: 0 markets (skipped 0 no-signal) | keyword 'OpenAI IPO': ok |
| polymarket_related | OK | 0.5s | 8 | Polymarket search over 1 keyword(s), kept 8 matches. keyword 'OpenAI IPO': 8 markets |
| article_search | OK | 0.2s | 10 | Fetched 10 unique articles across 1 queries (lookback=120d). 'OpenAI IPO banks hired underwriters Goldman Sachs Morgan Stanley lead-left': 10 articles |
research loops
·
4/4 ok ·
16 searches ·
2 page reads
| dimension | status | time | searches | page reads | report |
|---|---|---|---|---|---|
| current state | OK | 18s | 4 | 0 | 3618 chars |
| base rate | OK | 26s | 4 | 2 | 4783 chars |
| yes case | OK | 29s | 4 | 0 | 4517 chars |
| no case | OK | 22s | 4 | 0 | 4847 chars |
report: current state
- [2026-06-08] OpenAI announced it had confidentially submitted a draft S-1 to the SEC. It gave no timing and no bank lineup. https://techcrunch.com/2026/06/08/following-anthropic-openai-files-confidentially-for-ipo/
- [2026-05-20] WSJ, CNBC and Axios reported that OpenAI was working with banks including Goldman Sachs and Morgan Stanley on the filing. JPMorgan was reported as also involved. These were press reports, not company statements. https://www.cnbc.com/2026/05/20/openai-ipo-filing.html
- [2026-06-10] Fortune reported that it was "widely reported" that OpenAI and Anthropic had chosen Goldman and Morgan Stanley as their two top bookrunners. It said neither had decided who gets lead-left. Goldman holds lead-left on SpaceX's IPO (listed 2026-06-12). https://fortune.com/2026/06/10/goldman-sachs-morgan-stanley-openai-anthropic-ipos/
- [2026-08-13] A Prediction Hunt article claimed the confidential S-1 "names" Goldman, Morgan Stanley and JPMorgan as lead underwriters, and that Goldman was at 82% on prediction markets. Confidential drafts are not public, so this is unverified. It also gives no lead-left order. https://www.predictionhunt.com/news/goldman-sachs-leads-openai-ipo-at-82-aug-13-2026
- [2026-08-19] CFO Sarah Friar told staff OpenAI "will be a public company in 2027". It could list sooner if the business keeps inflecting. The company is "confidentially under file". https://www.cnbc.com/2026/08/19/open-ai-ipo-timing-2027-friar.html
- [2026-09-12] Altman told Fortune that a 2026 IPO would be "ill-advised", citing AI safety. Asked about 2026, he said "I would say not 2026". https://fortune.com/2026/09/12/sam-altman-openai-ipo-delay-ill-advised-moment-safety-concerns/
- [2026-10-08] The FT reported that OpenAI told investors its annualized revenue was near $50B at the end of September, about $20B below the expected $70B. Anthropic reportedly is targeting a listing above $2T after the Nov 3 midterms. https://www.pymnts.com/news/artificial-intelligence/2026/anthropic-narrows-revenue-gap-with-openai-before-ipo/
- [2026-10-14] A tech-insider.org summary says OpenAI's lead banks are "not disclosed". It says no OpenAI public S-1 exists and none has a roadshow date. Anthropic's reported leads are Morgan Stanley, Goldman and JPMorgan. https://tech-insider.org/openai-anthropic-ipo-first-kalshi-odds-swing/
- [2026-10-09] Bloomberg reports OpenAI is angling for a new private round at a $1.4T valuation. Forge shows a secondary price of $954.70 on 10/15. https://finance.yahoo.com/technology/article/appetite-for-openai-anthropic-ipos-will-be-strong-despite-some-negative-headlines-ives-204229507.html
- [about 2026-09-13] The Kalshi lead-left market (expires Jan 1, 2028) showed Goldman at 79%, Morgan Stanley 18% and JPMorgan 4%. That snapshot is stale. https://kalshi.com/markets/kxopenaileadleft/openai-lead-left-ipo/kxopenaileadleft-28jan01b
- [2026-08-14] Executive turnover (Simo, Lightcap, Dresser) is a reported execution risk before the IPO. https://www.cnbc.com/2026/08/14/open-ai-ipo-red-flag.html
BOTTOM LINE: Goldman and Morgan Stanley are the reported co-leaders, and no source confirms who is lead-left. A public S-1 cover will not exist until at least late 2026, and management now points to 2027. Two timing risks could stop resolution as Yes: an IPO slipping past 2028-01-01, or a shared top billing, which would give Goldman a partial payout under the rules.
NOT FOUND: any public OpenAI S-1 or cover page; any official or credible report naming the lead-left bank; any announced IPO date, roadshow or exchange; any news from the last 7 days on the bank order.
report: base rate
- [2026-06-10] Fortune reported that both OpenAI and Anthropic had chosen Goldman and Morgan Stanley as top bookrunners but had not decided who gets lead-left. It cited Kalshi at 73% for Goldman on OpenAI at that time. Fortune also noted that a Polymarket market on Anthropic was about even. https://fortune.com/2026/06/10/goldman-sachs-morgan-stanley-openai-anthropic-ipos/
- [2026-05-19] Reference case, SpaceX (2026): Goldman took lead-left, followed by Morgan Stanley, Bank of America, Citigroup and JPMorgan. The prospectus listed 23 banks. Reuters and CNBC had reported this before the public filing, and the cover confirmed it. https://www.cnbc.com/2026/05/19/spacex-picks-goldman-sachs-to-lead-record-breaking-ipo-sources-say.html
- [2026-05-19] Tesla (2010): Goldman was lead-left, with Morgan Stanley, JPMorgan and Deutsche Bank next. Goldman or Morgan Stanley has led most mega tech IPOs, and Fortune calls them the two banks with by far the biggest tech-IPO market share. Both are named for OpenAI in reports, but this is a two-bank reference class and not a clean base rate. https://www.cnbc.com/2026/05/19/spacex-picks-goldman-sachs-to-lead-record-breaking-ipo-sources-say.html
- [2025-12-19] Prior-forecast track record: Reuters said Morgan Stanley was the front-runner for SpaceX lead-left because of its Musk ties. Goldman won anyway. Polymarket had favoured Morgan Stanley until the odds swung to Goldman on 12 May. This shows the favourite and relationship narrative can be wrong. https://finance.yahoo.com/news/morgan-stanley-seen-front-runner-174313230.html
- [2026-06-08] OpenAI confirmed a confidential draft S-1. It said it had not decided on timing and that it "may be a while". Reports say Goldman and Morgan Stanley are leading the work. No source I found says the lead-left order has been decided or disclosed. https://www.cnbc.com/2026/06/08/openai-confidentially-files-for-ipo-prepping-wall-street-for-ai-debut.html
- [2026-08-13] Prediction Hunt, a secondary aggregator, showed Goldman at 82% on Kalshi and 76% on Polymarket for this contract family (KXOPENAILEADLEFT, resolves 2028-01-01). Morgan Stanley was about 21–23%, Citigroup about 2–4% and JPMorgan about 1–2%. Goldman had been as low as 53% earlier and was 59% three days before. On 2026-08-05 it showed Morgan Stanley at 22–32%. Treat these as stale snapshots. The page also claims the S-1 "names" a lead; that claim is unsupported and the cited sources are weak. https://www.predictionhunt.com/news/goldman-sachs-leads-openai-ipo-at-82-aug-13-2026
- [2026-08-19] CFO Friar told employees OpenAI "will be a public company in 2027", or sooner if the business keeps accelerating. In September Altman said no 2026 IPO. Altman called a 2026 listing "ill-advised" (reported via Fortune). The timing risk to a 2028-01-01 deadline is therefore moderate: the IPO is guided into 2027 but could slip. https://fortrovepartners.com/openai-ipo-timeline-valuation-tracker/
- [2026-07-13] FutureSearch (a forecasting model, track record unknown) put the median OpenAI IPO date at 2027-07-15. It also said confidential SEC review typically runs three to six months. That leaves room before the deadline but with limited buffer. https://futuresearch.ai/anthropic-openai-ipo-dates-valuations/
- [2026-05-20] Axios said both OpenAI and SpaceX work with Goldman, Morgan Stanley and JPMorgan. It also said confidential filings come a couple of months before the public S-1, and offerings about a month after that. The lead-left order is normally revealed only on the public S-1 cover or via leaks just before it. https://www.axios.com/2026/05/20/openai-ipo-spacex-musk
- [2026-06-10] Possible complication: Goldman and Morgan Stanley are reportedly bookrunning both OpenAI and Anthropic. Each could take lead-left on one deal. A reported Anthropic IPO window of October–November 2026 would come first. https://fortune.com/2026/06/10/goldman-sachs-morgan-stanley-openai-anthropic-ipos/
- [2026-10-16] The tie rule is a minor factor. Co-lead-left ties on a prospectus cover are rare, because the cover usually shows one uppermost-left name. A tie would pay Goldman 50% in a two-way tie.
BOTTOM LINE: The reference class is two banks, Goldman and Morgan Stanley, with Goldman winning lead-left on the closest comparables (Tesla 2010 and SpaceX 2026). SpaceX also showed that relationship-based favourites can lose. Outside-view risk to YES is mostly Morgan Stanley winning the order (market snapshots suggest about 20–25%) plus the chance the IPO slips past 2027 (guided for 2027, but not firm).
NOT FOUND: Any report or filing naming OpenAI's actual lead-left bank. A public S-1 for OpenAI. Current (October 2026) prices on the sibling markets. A rigorous historical base rate of lead-left outcomes in mega-IPOs.
report: yes case
YES means: Goldman Sachs is the single uppermost-and-leftmost bookrunner on the cover of OpenAI's final US IPO prospectus, and the IPO prices before 2028-01-01. A tied top billing would pay Goldman only a partial Yes.
- [2026-09-04] The FT reported that Morgan Stanley is the frontrunner for Anthropic's lead-left slot, citing four people. Goldman was expected to take the stabilization-agent role instead. The report said Morgan Stanley and Goldman would also pursue roles on OpenAI's IPO, "which could come early next year". This is the best structural YES argument. Morgan Stanley taking Anthropic would let OpenAI, or Goldman, claim OpenAI for Goldman. It would also suit Goldman, who is "balancing" its roles after winning SpaceX. This is inference. Neither company has said it avoids giving both deals to one bank. Morgan Stanley's Anthropic slot was "not finalized". https://finance.yahoo.com/markets/stocks/articles/anthropic-weighs-morgan-stanley-goldman-223747740.html
- [2026-09-07] GlobalData's Private Banker International repeated the FT report and added that Goldman held the leading role on SpaceX while Morgan Stanley was stabilisation agent. The page text was too long to read in full, so I confirmed the FT claims through the Yahoo/Stocktwits version above. The same report says Anthropic's reported leads, besides Morgan Stanley, are JPMorgan, Citigroup and Barclays, because they provided debt financing. https://www.privatebankerinternational.com/news/morgan-stanley-goldman-top-roles-on-anthropics-ipo/
- [2026-10-15] Euronews says Anthropic is aiming for a November listing after the midterms. It says Goldman and Morgan Stanley are "reportedly" leading that deal, and that the prospectus is still unpublished. If Anthropic lists first with Morgan Stanley on the cover's left, Goldman's best path is the OpenAI deal, and Anthropic's debut would show the market how the lead-left role pays. https://www.euronews.com/2026/10/15/anthropic-revenue-jumps-46-to-168bn-as-ai-firm-eyes-record-2-trillion-ipo-valuation
- [2026-08-12] Fortune states flatly that "Goldman is one of OpenAI's underwriters for its IPO". This is a looser claim than the one reported in the colleagues' bullets. https://fortune.com/2026/08/12/sam-altman-goldman-sachs-openai/
- [2026-06-08] CNBC notes that Goldman and Morgan Stanley are "the two firms listed at the top of SpaceX's filing". CNBC's earlier report said Goldman held lead-left on SpaceX, with Morgan Stanley second. This is the closest precedent for a Musk-linked rival of OpenAI. Goldman has just shown it can win the mandate for the largest tech IPO ever, which is the best-documented evidence for Goldman ahead of Morgan Stanley. https://www.cnbc.com/2026/06/08/openai-confidentially-files-for-ipo-prepping-wall-street-for-ai-debut.html
- [2026-08-?? to 2026-10] A July/August Fortune and IBTimes line says Goldman and Morgan Stanley have been in the preparations. The July/August IBTimes item adds that Citigroup and JPMorgan "have also been discussed" as possible additions. That implies no third bank is positioned to outrank Goldman, so the contest is two-way. https://www.ibtimes.co.uk/sam-altman-openai-ipo-goldman-sachs-1814030
- [2026-05-20] The WSJ, CNBC and Bloomberg reports listed Goldman first in "Goldman Sachs and Morgan Stanley" for OpenAI. CNBC's source said OpenAI was "working with banks including Goldman Sachs and Morgan Stanley". Order of mention is weak evidence of lead-left status and not confirmation. https://www.cnbc.com/2026/05/20/openai-ipo-filing.html
- [2026-09-21] Forge shows OpenAI at about $894B on 2026-09-21, and it says Altman has ruled out a 2026 listing. The IPO is guided to 2027, which still falls before the 2028-01-01 deadline. Slippage is the larger risk to a Yes, but the guidance leaves a buffer of at least several months. https://forgeglobal.com/insights/openai-upcoming-ipo-news/
BOTTOM LINE: The strongest YES path is that Goldman wins the tie-break against Morgan Stanley on OpenAI, because the FT reports Morgan Stanley leading for Anthropic, and Goldman's SpaceX lead-left win is the key precedent. This requires an OpenAI IPO pricing in 2027 and a single bank, not a joint pair, on the cover. No source confirms the order.
NOT FOUND: any report or filing naming OpenAI's lead-left bank. A public OpenAI S-1. An IPO date for OpenAI. Any Goldman statement about OpenAI. The Prediction Hunt claim that the confidential S-1 "names" the banks is unverified, and I could not find a primary source for it.
report: no case
YES means: Goldman Sachs is the single uppermost-left underwriter on the cover of the final prospectus for OpenAI's US IPO, priced before 2028-01-01. A tied top billing would pay Goldman only a fraction.
- [2026-09-04/05] Anthropic's bank order points away from Goldman. The FT and Reuters (as relayed by Octagon) said Morgan Stanley is the pick for lead-left on Anthropic, with Goldman as stabilization agent. The FT noted the Morgan Stanley selection "has not been finalized". Octagon is a secondary, auto-generated page and I could not open the FT original (403). If both banks co-lead both deals, issuers may split the lead-left slots. Anthropic going to Morgan Stanley could mean OpenAI goes to Goldman, which helps YES. The other reading is that Morgan Stanley wins both, or that OpenAI weighs Goldman's SpaceX lead-left against a wish to differ from Anthropic. https://www.octagonai.co/markets/financials/ipos/which-bank-will-lead-anthropic-s-ipo
- [2026-09-12 to 2026-10-09] The IPO has slipped out of 2026. Altman said "not 2026". Press on 10/7 and 10/9 describe OpenAI pursuing a private round at about $1.4T to bridge to a 2027 listing. Smartasset (10/3) says no date, size, exchange or valuation has been disclosed. A private round at $1.4T raises the IPO valuation bar. Altman has reportedly called anything under $1T a nonstarter. A round that raises the price expectation and a weaker market could both push a listing into 2028. https://smartasset.com/investing/openai-stock-ipo
- [2026-09-25] One recap says the 2027 date was only an internal CFO expectation, not a scheduled launch, and "no later IPO has been scheduled". So the 2028-01-01 deadline leaves about 14.5 months with no formal timetable. https://www.neoteo.com/en/openai-rules-out-a-2026-ipo-amid-reported-talks-at-up-to-15-trillion
- [2026-09-20] The claim that Goldman "occupies" lead-left in OpenAI's S-1 comes from Prediction.com. The same article says Morgan Stanley is "co-lead". A confidential draft is not public, so the site cannot have seen it. I treat it as unsupported. It also contradicts the 2026-06-10 Fortune report that the order was undecided. Aggregator articles may inflate YES prices. https://prediction.com/news/morgan-stanley-hits-23-in-openai-ipo-market-sep-20-2026
- [2026-06 to 2026-09] OpenAI has never confirmed its banks. One summary says the company has not confirmed Goldman or Morgan Stanley directly, and that the "lead underwriter" labels come from media. The Bloomberg and Reuters Anthropic reports named Morgan Stanley first and Goldman second. Reports on OpenAI have listed Goldman first in some places, but that is not a cover order. https://techjacksolutions.com/ai-brief/openai-submits-confidential-s-1-to-the-sec-officially-enteri/
- [2026-06-03] Anthropic's bank selection was reported by Bloomberg, with Citigroup later added. Syndicates this large are also growing. SpaceX had 23 banks. OpenAI's lineup could add Citi or Bank of America, and a late reshuffle is possible. The order stays a live variable until the public S-1 is filed. https://www.bloomberg.com/news/articles/2026-06-03/anthropic-said-to-pick-morgan-stanley-goldman-sachs-to-lead-ipo
- [2026-06 to 2026-10] Both rivals' disclosure timing also helps NO. OpenAI's confidential draft was submitted in June, and nothing public has followed. A public S-1 must be filed at least 15 days before the roadshow. A 2027 IPO means the cover is not visible for months. The Kalshi contract has no early-resolution trigger short of that cover, and a pre-filing leak is not binding under the rules. https://techjacksolutions.com/ai-brief/openai-submits-confidential-s-1-to-the-sec-officially-enteri/
- [2026-10-09] The market backdrop is weaker for OpenAI. Polymarket odds of a $1.5T OpenAI market cap fell from 80% to 58% on the delay. The cited reasons include post-SpaceX volatility and weaker AI profitability sentiment. This is the kind of drift that turns a 2027 target into a 2028 one. https://cryptobriefing.com/openai-market-cap-odds-drop-ipo-delay/
- [2026-10-08] The revenue shortfall of about $20B against the $70B expected (per the FT, in the colleagues' report) adds to the valuation-gap risk. This is the newest figure I did not re-fetch. A smaller valuation gap needs a private round to succeed first.
BOTTOM LINE: The strongest NO case has three parts. The IPO has no scheduled date and could miss 2028-01-01. The lead-left order is unreported. Anthropic reportedly going to Morgan Stanley leaves Morgan Stanley as a credible OpenAI lead-left too, and the claims that Goldman is already in the S-1 come from weak aggregators.
NOT FOUND: any official or credible report naming OpenAI's lead-left bank; an OpenAI public S-1; a scheduled IPO date; a readable copy of the FT or Reuters Anthropic bank-order stories; current October prices on sibling markets.
3. Forecasters 2 draws · Opus 5.5 + GPT · plain mean · re-drawn after crux research
crux research
The first round disagreed: the draws' mean 0.53 sits 0.20 from the same-question market "this kalshi market" at 0.73.
One more research loop went after the crux and every draw below was redone with its report.
first-round draws:
55%, 50%
· crux loop: 4 searches, 3 page reads
report: crux
CRUX: The disagreement turns on two things. One is whether OpenAI actually lists before 2028-01-01. The other is whether Goldman, not Morgan Stanley, is the single top-left bank if it does. No primary source settles the second point: no public S-1 exists, and the "Goldman is lead-left" claims come from weak sources.
Evidence:
- **2026-06-08:** OpenAI announced it had confidentially submitted a draft S-1. It said the filing "gives us the option to go public sooner" and that timing was undecided. https://www.cnbc.com/2026/06/08/openai-confidentially-files-for-ipo-prepping-wall-street-for-ai-debut.html
- **2026-05-20, CNBC/WSJ:** OpenAI was working with "banks including Goldman Sachs and Morgan Stanley". No lead-left ordering was reported for OpenAI. The same article said sources put Goldman in the lead-left spot on SpaceX's prospectus. https://www.cnbc.com/2026/05/20/openai-ipo-filing.html
- **Prediction-site claims about OpenAI's bank order:** Pages such as predictionhunt (2026-08-13) say the confidential S-1 "names" Goldman, Morgan Stanley and JPMorgan as leads. A confidential draft is not public, so these pages cannot have seen its cover. Treat them as unreliable on lead-left ordering. https://www.predictionhunt.com/news/goldman-sachs-leads-openai-ipo-at-82-aug-13-2026
- **2026-09-07, FT via Private Banker International:** Morgan Stanley was the frontrunner for Anthropic's lead-left, with Goldman as stabilisation agent. Anthropic's choice was not final. The same piece said Morgan Stanley and Goldman were "expected to compete" for OpenAI roles, so OpenAI's lead-left was not settled as of then. It also said Goldman led SpaceX and Morgan Stanley was its stabilisation agent. https://www.privatebankerinternational.com/news/morgan-stanley-goldman-top-roles-on-anthropics-ipo/
- **2026-06 (Fortune, via Cryptopolitan):** Goldman and Morgan Stanley were the top two bookrunners on both AI deals, and neither company had named a lead-left. https://www.cryptopolitan.com/goldman-morgan-stanley-openai-anthropic-ipos/
- **2026-08-19, CNBC:** CFO Friar told employees OpenAI "will be a public company in 2027", or sooner if the business keeps inflecting. https://www.cnbc.com/2026/08/19/open-ai-ipo-timing-2027-friar.html (snippet only; the page would not load)
- **2026-09-12/13 and 2026-10-07:** Altman told Fortune "I would say not 2026" and called now an "ill-advised moment" to go public. TokenPost's 2026-10-07 write-up says he named 2027 as the earliest window but gave no firm date. OpenAI's official line is that timing is undecided. https://www.tokenpost.com/news/business/27478 and https://www.cnbc.com/2026/09/12/anthropics-amodei-proposes-plan-to-slow-the-pace-of-advancing-ai-capabilities.html
- **2026-09 (Forge, citing the NYT):** OpenAI was leaning toward 2027 because of its $1T+ valuation goal. A quicker IPO at a lower valuation was reportedly a "nonstarter". The same piece says OpenAI is considering another primary funding round. https://forgeglobal.com/insights/openai-upcoming-ipo-news/
BOTTOM LINE: The evidence supports a 2027 IPO as the stated plan. The 2027 date is Friar's internal guidance, though, and Altman has set no firm date, so slippage past 2027 is plausible. On lead-left, Goldman's SpaceX win makes it a plausible favourite. Reputable reporting through September still describes the OpenAI order as contested, and Morgan Stanley is the FT's frontrunner on Anthropic. The market's high Goldman price rests mostly on low-quality claims about a confidential S-1 and appears to ignore IPO-timing risk.
NOT FOUND: A public OpenAI S-1 or any prospectus cover. Any Bloomberg, WSJ, Reuters or FT report naming a single OpenAI lead-left bank. A firm OpenAI IPO date. The Aug 19 CNBC page did not load in full, so its content is from search snippets.
claude-opus-5-5 · high effort
42% conf
| Yes | 52% |
| No | 48% |
(0) The condition is not met. EDGAR shows no public OpenAI S-1 as of 2026-10-16, only a confidential draft submitted in June 2026, and no credible outlet has named the lead-left bank. (a) About 14.5 months remain until 2028-01-01. Altman has ruled out a 2026 listing, so an IPO, including a public S-1 at least 15 days before the roadshow, has to happen in 2027. (b) Status quo is No: if nothing happens, there is no IPO and the market resolves No. (c) No path: OpenAI raises a private round at around $1.4T, revenue runs about $20B below expectations, and the listing slips into 2028. Or Morgan Stanley wins lead-left, or a joint top billing triggers the partial-payout tie rule. (d) Yes path: the IPO prices in 2027 as CFO Friar guided. Goldman takes the top-left slot, building on its SpaceX lead-left win, while Morgan Stanley takes Anthropic, which the FT reports it is frontrunner for. Kalshi's 73% looks like it prices mostly the conditional lead-left question; the 'Goldman is in the S-1' claims come from weak aggregator sites. I estimate P(IPO before 2028) at about 0.65 and P(Goldman sole lead-left given an IPO) at about 0.70–0.75, which gives roughly 0.47. I shade that toward the market anchor and land at 0.52.
claude-opus-5-5 · high effort
42% conf
| Yes | 45% |
| No | 55% |
(0) Not resolved. As of 2026-10-16, EDGAR shows no OpenAI S-1, so no prospectus cover exists. OpenAI has only a confidential draft from June 2026, and no credible outlet (FT, WSJ, Bloomberg, Reuters) has reported a lead-left bank. (a) About 14.5 months remain, until 2028-01-01. (b) If nothing changes, the answer is No: there is no IPO date. Altman ruled out 2026, and 2027 is only internal CFO guidance with no firm timetable. (c) The NO path is timing slippage. A private round at about $1.4T, a reported ~$20B revenue shortfall, Altman calling anything under $1T a nonstarter, and a weaker AI backdrop could push the listing into 2028. Morgan Stanley could also take the top-left spot, since it is the FT's frontrunner on Anthropic and reporting through September called the OpenAI order contested. (d) The YES path is an OpenAI listing in 2027 with Goldman alone at top-left. Goldman has lead-left precedents in SpaceX 2026 and Tesla 2010. Morgan Stanley taking Anthropic could also lead OpenAI to balance the mandates toward Goldman. I put P(IPO before the deadline) at about 0.6 and P(Goldman sole lead-left | IPO) at about 0.7–0.75, with a small haircut for a possible tie. That gives roughly 0.43–0.45. This is well below the Kalshi anchor of 73% (bid 73 / ask 78, thin volume). The market looks to be pricing the conditional bank choice, drawing on unreliable 'S-1 names Goldman' aggregator claims, and underweighting the timing risk.
4. Final plain mean of 2 draws · second round
48%
(42% confidence)
Yes: 48%
No: 52%
Mean of 2 draws (0.52, 0.45) after one crux research pass. (0) The condition is not met. EDGAR shows no public OpenAI S-1 as of 2026-10-16, only a confidential draft submitted in June 2026, and no credible outlet has named the lead-left bank. (a) About 14.5 months remain until 2028-01-01. Altman has ruled out a 2026 listing, so an IPO, including a public S-1 at least 15 days before the roadshow, has to happen in 2027. (b) Status quo is No: if nothing happens, there is no IPO and the market resolves No. (c) No path: OpenAI raises a private round at around $1.4T, revenue runs about $20B below expectations, and the listing slips into 2028. Or Morgan Stanley wins lead-left, or a joint top billing triggers the partial-payout tie rule. (d) Yes path: the IPO prices in 2027 as CFO Friar guided. Goldman takes the top-left slot, building on its SpaceX lead-left win, while Morgan Stanley takes Anthropic, which the FT reports it is frontrunner for. Kalshi's 73% looks like it prices mostly the conditional lead-left question; the 'Goldman is in the S-1' claims come from weak aggregator sites. I estimate P(IPO before 2028) at about 0.65 and P(Goldman sole lead-left given an IPO) at about 0.70–0.75, which gives roughly 0.47. I shade that toward the market anchor and land at 0.52.
Pipeline Timing
Total pipeline time: 106.9s
Per-tool research timings shown in the Research section above.