The Forecast
This question resolves on one narrow fact that the final prospectus cover will make unambiguous: whether Goldman Sachs is the uppermost-and-leftmost underwriter name on Anthropic's US IPO before January 2028. Reporting from the Financial Times on September 4, 2026, citing four people familiar with the matter, placed Morgan Stanley in the lead-left position and Goldman in the stabilization-agent role, a role division that is the mirror image of the SpaceX IPO earlier in 2026. The evidence leans against Goldman, with the caveat that the same reporting says the appointment was not finalized and could still change, and no public S-1 has appeared on EDGAR to settle it.
The market asks a question that the settlement source answers mechanically. When Anthropic files its public S-1 and then its final prospectus, the underwriter names appear across the bottom of the cover page in a fixed order, and the name furthest up and furthest left is the lead-left. Kalshi resolves off SEC EDGAR filings. There is no interpretation involved once the document exists, which makes this a question about one already-largely-made corporate decision rather than about a contested future outcome.
The decision appears to have already tilted. On September 4, 2026 the Financial Times reported that Morgan Stanley was the frontrunner for the lead-left mandate and Goldman Sachs was expected to serve as stabilization agent, the role that manages the first trading sessions after the listing. Four people familiar with the matter were cited. Supporting detail in the same reporting strengthens the read: Morgan Stanley had been leading price discussions with prospective investors, which is lead-left work, and Morgan Stanley leads the syndicated revolving credit facility that expanded to $15 billion, with Goldman, JPMorgan and Citigroup in prominent but subordinate positions in that syndicate. The role split also reads as a deliberate alternation from the SpaceX IPO in June 2026, where Goldman held lead-left and Morgan Stanley was the stabilization agent.
The outside view for a question of this type is not a coin flip between two banks. When credible reporting citing multiple sources identifies a specific bank as holding the lead-left position weeks before filing, that reporting is usually right, because the mandate is effectively awarded by the time price conversations with investors begin. The reference class is reported-and-then-confirmed lead-left assignments on large US IPOs, where the reported frontrunner prevails in the large majority of cases. That base rate is estimated rather than counted here, and that limitation is stated openly.
What cuts the other way is real but modest. The FT's own sources stressed the choice was not finalized. The IPO has slipped twice, from late September to mid-October to November per reporting on September 24, and a longer runway is more time for a mandate to move. Anthropic's founders are seeking a 50.1% combined voting structure, which signals a company willing to break convention on deal terms. And there is genuine residual probability that the deal does not happen at all before January 2028, which resolves this NO regardless of which bank was favored.
The tie provision in the rules deserves attention. If Goldman and Morgan Stanley appear as co-lead-left with equal top billing, Goldman's side pays $0.50 rounded down, which is $0. That edge case therefore resolves as a loss for YES holders under the payout mechanics, and it is not a negligible scenario on a deal where both banks are being kept close.
Latest evidence: 2026-09-28
The Outside View (2)
How often comparable situations resolved this way, before the specifics. “Counted” means the cases are named and counted; “estimated” means no count was available and the rate is an estimate, said so.
Large US IPOs where credible multi-source financial reporting named a specific bank as lead-left frontrunner in the weeks before the public S-1 filing, measured by whether a different bank ultimately took the cover-page lead-left positionThe base rate usedestimated
Base rate: the reported frontrunner prevails in roughly 80-85% of such cases, so a non-frontrunner bank named as holding a different senior role takes lead-left in roughly 15-20% · 17%
Cases: No systematic count of reported-versus-realized lead-left assignments was available in the record for this scan; the rate is an estimate reflecting that mandate decisions are typically made before price conversations with investors begin, which is the stage reporting described.
Fit: Fits well on structure: this is exactly the situation of a reported frontrunner ahead of filing. Breaks in two places — the reporting itself flagged the choice as not finalized, which is more hedged than a typical frontrunner report, and this deal is unusually large and unusually delayed, both of which give the mandate more time and more reason to move than the typical case.
Source: estimated — no published study of lead-left reporting accuracy was located
Two-horse lead-left contests between Goldman Sachs and Morgan Stanley on mega-cap tech IPOs, with no information about which is favoredestimated
Base rate: approximately even between the two, with a small residual for a third bank or no deal · 45%
Cases: Not counted. Both banks are described in the record as holding the largest tech-IPO market shares.
Fit: This is the correct prior only in the absence of deal-specific reporting. Since substantial deal-specific reporting exists, using this as the prior would discard the best evidence available. It is listed to show what the uninformed baseline looks like and how far the reporting moves from it.
Source: estimated from the structural fact that these two banks dominate tech IPO lead-left mandates
For this window: No time conversion is required. The question is a one-shot contest resolved by a single document, not a rate per period. The base rate already describes the probability that the non-frontrunner takes the position. The only window adjustment worth noting cuts slightly toward more uncertainty rather than less: the 459-day close window means the prospectus could be filed at any point, and a longer pre-filing delay gives a mandate more opportunity to move than the typical case in the reference class, which is already reflected in choosing 0.17 near the upper end rather than the lower end of the estimated range.
Starting point: 17%
What Moves It (7)
Specific evidence about this question that the starting point does not already carry. A claim’s weight can never exceed its grade: an inferred adjustment counts at most moderate, a contested one at most slight.
Mirror-image rotation from the SpaceX IPOdocumented
toward no · moderate
Evidence: Goldman held the lead-left role on SpaceX's June 2026 offering while Morgan Stanley was the stabilisation agent; the reported Anthropic split is the exact inverse, with Morgan Stanley lead-left and Goldman stabilisation agent.
Not already counted because: The reference class treats frontrunner reports as generic; it does not carry the structural coherence of a deliberate alternation between the same two banks across consecutive mega-deals, which makes this particular report more likely to be describing a settled arrangement.
Source: Private Banker International, September 2026, reporting the FT; corroborated by an FT-sourced summary noting Morgan Stanley was joint lead bookrunner and sole stabilisation agent on SpaceX's June IPO where Goldman held lead-left
Morgan Stanley already performing lead-left functionsreported
toward no · moderate
Evidence: Morgan Stanley had in recent weeks led discussions with potential investors about the price at which Anthropic could sell shares — the valuation and allocation advisory work that defines the lead-left role — as reported September 4, 2026.
Not already counted because: The base rate counts reported intentions; this is reported behavior already underway, which is a stronger signal than a stated plan because reversing it would mean unwinding work already done with investors.
Source: Financial Times via Investing.com and Traders Union, September 4, 2026
Morgan Stanley leads the $15 billion credit syndicatedocumented
toward no · slight
Evidence: Morgan Stanley leads the revolving credit facility that expanded to $15 billion, with Goldman Sachs, JPMorgan and Citigroup in prominent but subordinate syndicate positions; coverage described the tiered structure as a proxy competition for IPO roles.
Not already counted because: This is an independent, structurally documented indicator of relative bank standing, separate from the FT's sourcing on the mandate itself — and the agent role on a syndicated facility is a commitment already executed rather than an expectation.
Source: TechTimes, September 4, 2026; Forbes, September 7, 2026
The mandate was explicitly described as unfinalizeddocumented
toward yes · slight
Evidence: People familiar with the matter said the choice of Morgan Stanley for the lead-left role had not been finalised and could still change, per the September 4, 2026 reporting.
Not already counted because: The reference class assumes a typical frontrunner report; this one carries an unusually explicit caveat from the same sources, which means the report is weaker evidence than the average case in the class.
Source: Financial Times via Private Banker International and Traders Union, September 4, 2026
Repeated timeline slippage with no prospectus fileddocumented
toward yes · slight
Evidence: The confidential draft was filed June 1, 2026 and a public prospectus was expected after Labor Day; as of September 26, 2026 SEC EDGAR remained empty, and reporting on September 24 pushed the expected debut from October to November.
Not already counted because: The base rate is drawn from cases where filing followed the report promptly. Extended delay means more time for the mandate to move and more opportunity for bank competition to reopen, which the class does not carry. It also raises the residual probability of no deal at all, but that cuts toward NO, so this adjustment is held small.
Source: Forkast, September 26, 2026; TipRanks, September 24, 2026
Founder willingness to depart from conventiondocumented
toward yes · slight
Evidence: Anthropic is asking shareholders to approve a structure giving its seven co-founders a combined 50.1% of the vote on most corporate matters, per TechCrunch, September 25, 2026.
Not already counted because: The reference class assumes a conventional issuer accepting standard syndicate structures. A founder group restructuring voting control ahead of listing is more likely than average to impose an unconventional arrangement, including a shared or reshuffled top billing — though note the tie case pays $0 for Goldman, so this only helps YES in the outright-reversal branch.
Source: TechCrunch, September 25, 2026
Residual probability the IPO does not complete before January 2028inferred
toward no · slight
Evidence: No public S-1 exists as of September 26, 2026; the New York Times reported September 28, 2026 that liability questions around AI safety risks are a live issue for Anthropic's and OpenAI's listing plans, and Bloomberg's September 25 commentary characterized the IPO market underneath these mega-deals as weak.
Not already counted because: The reference class conditions on IPOs that completed; it does not price the branch where no prospectus is ever filed. Every such path resolves this market NO regardless of bank standing.
Source: Forkast, September 26, 2026; New York Times, September 28, 2026; Bloomberg, September 25, 2026 — inference from these that a non-trivial no-deal branch exists
What Would Move It Next (6)
Anthropic's public Form S-1 appearing on SEC EDGAR with the underwriter syndicate listed on the cover page, showing whether Goldman Sachs or Morgan Stanley occupies the uppermost-and-leftmost positionExpected weeks to months out; reporting on September 24, 2026 pointed to a November 2026 listing, which would require a public filing well before then
toward no · decisive
Where to watch: SEC EDGAR full-text search for Anthropic PBC registration statements
Financial Times, Bloomberg or Reuters reporting that Anthropic has reversed or reopened the lead-left mandate in Goldman Sachs's favour
toward yes · strong
Where to watch: Financial Times and Bloomberg IPO coverage of Anthropic underwriter roles
Confirmation that Goldman Sachs has been formally designated stabilization agent on the Anthropic offering, which under standard practice is held by a bank other than the lead-leftWould appear in the preliminary or final prospectus underwriting section
toward no · strong
Where to watch: The underwriting section of Anthropic's S-1 or S-1/A on SEC EDGAR
A further Anthropic IPO delay pushing the listing beyond 2026 into 2027, which extends the window for a mandate reshuffle
toward yes · slight
Where to watch: Wall Street Journal, Bloomberg and Reuters IPO calendar coverage
Anthropic announcing withdrawal, indefinite postponement, or a non-IPO listing route such as a direct listing for its US public debut
toward no · decisive
Where to watch: Anthropic company announcements and SEC EDGAR filings including any RW withdrawal request
Reporting or a filed prospectus showing Goldman Sachs and Morgan Stanley sharing equal top billing on the Anthropic cover page rather than one bank uppermost-and-leftmost
toward no · moderate
Where to watch: The cover page of the preliminary prospectus on SEC EDGAR
The claims behind this analysis, each with its verification status — including what is contested, unverified, or could not be established.
What each grade meansLarge US IPOs where credible multi-source financial reporting named a specific bank as lead-left frontrunner in the weeks before the public S-1 filing, measured by whether a different bank ultimately took the cover-page lead-left position: the reported frontrunner prevails in roughly 80-85% of such cases, so a non-frontrunner bank named as holding a different senior role takes lead-left in roughly 15-20%
Fits well on structure: this is exactly the situation of a reported frontrunner ahead of filing. Breaks in two places — the reporting itself flagged the choice as not finalized, which is more hedged than a typical frontrunner report, and this deal is unusually large and unusually delayed, both of which give the mandate more time and more reason to move than the typical case.
— ESTIMATED BASE RATEestimated — no published study of lead-left reporting accuracy was located · cases: No systematic count of reported-versus-realized lead-left assignments was available in the record for this scan; the rate is an estimate reflecting that mandate decisions are typically made before price conversations with investors begin, which is the stage reporting described.
Two-horse lead-left contests between Goldman Sachs and Morgan Stanley on mega-cap tech IPOs, with no information about which is favored: approximately even between the two, with a small residual for a third bank or no deal
This is the correct prior only in the absence of deal-specific reporting. Since substantial deal-specific reporting exists, using this as the prior would discard the best evidence available. It is listed to show what the uninformed baseline looks like and how far the reporting moves from it.
— ESTIMATED BASE RATEestimated from the structural fact that these two banks dominate tech IPO lead-left mandates · cases: Not counted. Both banks are described in the record as holding the largest tech-IPO market shares.
Mirror-image rotation from the SpaceX IPO: Goldman held the lead-left role on SpaceX's June 2026 offering while Morgan Stanley was the stabilisation agent; the reported Anthropic split is the exact inverse, with Morgan Stanley lead-left and Goldman stabilisation agent.
The reference class treats frontrunner reports as generic; it does not carry the structural coherence of a deliberate alternation between the same two banks across consecutive mega-deals, which makes this particular report more likely to be describing a settled arrangement.
✓ DOCUMENTEDPrivate Banker International, September 2026, reporting the FT; corroborated by an FT-sourced summary noting Morgan Stanley was joint lead bookrunner and sole stabilisation agent on SpaceX's June IPO where Goldman held lead-left · moderate toward no
Morgan Stanley already performing lead-left functions: Morgan Stanley had in recent weeks led discussions with potential investors about the price at which Anthropic could sell shares — the valuation and allocation advisory work that defines the lead-left role — as reported September 4, 2026.
The base rate counts reported intentions; this is reported behavior already underway, which is a stronger signal than a stated plan because reversing it would mean unwinding work already done with investors.
○ REPORTEDFinancial Times via Investing.com and Traders Union, September 4, 2026 · moderate toward no
Morgan Stanley leads the $15 billion credit syndicate: Morgan Stanley leads the revolving credit facility that expanded to $15 billion, with Goldman Sachs, JPMorgan and Citigroup in prominent but subordinate syndicate positions; coverage described the tiered structure as a proxy competition for IPO roles.
This is an independent, structurally documented indicator of relative bank standing, separate from the FT's sourcing on the mandate itself — and the agent role on a syndicated facility is a commitment already executed rather than an expectation.
✓ DOCUMENTEDTechTimes, September 4, 2026; Forbes, September 7, 2026 · slight toward no
The mandate was explicitly described as unfinalized: People familiar with the matter said the choice of Morgan Stanley for the lead-left role had not been finalised and could still change, per the September 4, 2026 reporting.
The reference class assumes a typical frontrunner report; this one carries an unusually explicit caveat from the same sources, which means the report is weaker evidence than the average case in the class.
✓ DOCUMENTEDFinancial Times via Private Banker International and Traders Union, September 4, 2026 · slight toward yes
Repeated timeline slippage with no prospectus filed: The confidential draft was filed June 1, 2026 and a public prospectus was expected after Labor Day; as of September 26, 2026 SEC EDGAR remained empty, and reporting on September 24 pushed the expected debut from October to November.
The base rate is drawn from cases where filing followed the report promptly. Extended delay means more time for the mandate to move and more opportunity for bank competition to reopen, which the class does not carry. It also raises the residual probability of no deal at all, but that cuts toward NO, so this adjustment is held small.
✓ DOCUMENTEDForkast, September 26, 2026; TipRanks, September 24, 2026 · slight toward yes
Founder willingness to depart from convention: Anthropic is asking shareholders to approve a structure giving its seven co-founders a combined 50.1% of the vote on most corporate matters, per TechCrunch, September 25, 2026.
The reference class assumes a conventional issuer accepting standard syndicate structures. A founder group restructuring voting control ahead of listing is more likely than average to impose an unconventional arrangement, including a shared or reshuffled top billing — though note the tie case pays $0 for Goldman, so this only helps YES in the outright-reversal branch.
✓ DOCUMENTEDTechCrunch, September 25, 2026 · slight toward yes
Residual probability the IPO does not complete before January 2028: No public S-1 exists as of September 26, 2026; the New York Times reported September 28, 2026 that liability questions around AI safety risks are a live issue for Anthropic's and OpenAI's listing plans, and Bloomberg's September 25 commentary characterized the IPO market underneath these mega-deals as weak.
The reference class conditions on IPOs that completed; it does not price the branch where no prospectus is ever filed. Every such path resolves this market NO regardless of bank standing.
— INFERREDForkast, September 26, 2026; New York Times, September 28, 2026; Bloomberg, September 25, 2026 — inference from these that a non-trivial no-deal branch exists · slight toward no
The Financial Times reported on September 4, 2026, citing four people familiar with the matter, that Morgan Stanley was the frontrunner for Anthropic's lead-left mandate and Goldman Sachs was expected to serve as stabilization agent.
This is the single most direct piece of evidence on the question and it points against Goldman.
The same reporting states the Morgan Stanley lead-left selection had not been finalized and could still change.
It caps how large the adjustment against Goldman can honestly be and prevents a decisive grading.
On SpaceX's June 2026 IPO, Goldman Sachs held the lead-left role and Morgan Stanley was the sole stabilization agent — the exact inverse of the reported Anthropic split.
The mirror-image structure makes the reported Anthropic assignment more credible as a deliberate rotation rather than a preliminary guess.
Morgan Stanley led the revolving credit syndicate that expanded to $15 billion, with Goldman Sachs, JPMorgan and Citigroup in prominent but subordinate positions; the facility was widely described as a proxy competition for IPO roles.
Independent corroboration of Morgan Stanley's primacy through a documented financing structure rather than a single reporting thread.
As of September 26, 2026, no public Anthropic S-1 had appeared on SEC EDGAR; the confidential draft was filed June 1, 2026 and reporting on September 24 pushed the expected listing from October to November.
The settlement document does not yet exist, leaving room for the mandate to shift and for the deal itself to slip or fail before January 2028.
Anthropic is asking shareholders to approve a structure giving its seven co-founders a combined 50.1% of the vote on most corporate matters, per TechCrunch on September 25, 2026.
Evidence of a founder group willing to depart from standard deal convention, which slightly raises the chance of an unconventional underwriter arrangement such as shared top billing.