Will Twilio disclose specific AI or Voice AI revenue metrics by Q2 2026 earnings?

Q4 2025 earnings confirmed the non-disclosure pattern — management had a natural opportunity to quantify AI revenue and chose not to. The CEO's 'early days' framing and description of products as 'in private beta' actively positions AI as pre-revenue-materiality. Three remaining opportunities exist: Q1 2026 earnings, SIGNAL conference (May), and Q2 2026 earnings. SIGNAL is the strongest candidate because investor conferences sometimes debut new metrics, and Twilio has historically used SIGNAL for product announcements. However, the calibration feedback from the guidance miss (Brier 0.4624) teaches that our ensemble overestimates the probability of management behavior changes. The pattern of selective disclosure (growth rates without bases) is now confirmed through 5+ consecutive quarters, making continued non-disclosure the strong base case. The usage-based model complexity that management cites as a barrier to AI revenue isolation is a genuine structural obstacle, not just a talking point.

I'm giving slightly more weight to the SIGNAL conference catalyst. Twilio's annual developer conference has historically been where major product and strategic announcements happen. If Voice AI revenue has grown from 60%+ in Q4 to potentially 70-80%+ by May 2026, the absolute number could be reaching a threshold where disclosure becomes a competitive advantage — 'look how much AI revenue we generate' positions the stock for AI premium valuation. Gartner naming Twilio 'company to beat in CPaaS AI' creates external pressure for quantification. Additionally, branded calling at ~6x YoY and RCS at ~5x QoQ show management is willing to disclose specific growth rates for new products — the precedent exists if the number is flattering. However, ConversationRelay is described as 'growing off a relatively small base' in Q4 2025, suggesting the absolute number may still be too small to be impressive. The three remaining opportunities (with SIGNAL as the strongest catalyst) keep this slightly above the lower bound.

The strongest argument against disclosure is the structural one: Twilio's usage-based model means AI revenue is deeply intermingled with existing product revenue. When an existing customer uses ConversationRelay, that shows up as voice minutes — disaggregating 'AI voice minutes' from 'traditional voice minutes' requires a new reporting framework that management has shown no interest in building. The CEO's language has actually moved away from quantification ('early days,' 'relatively small base,' 'private beta') rather than toward it. The pattern across 5+ quarters is unmistakable: management will provide percentage growth rates and customer counts for AI products but will not disclose absolute revenue. This is a deliberate strategic choice — if the number were large enough to impress, they would disclose it; if it's small, disclosure would undermine the AI narrative. The calibration lesson is clear: when management consistently behaves in a pattern, predict they'll continue that pattern rather than change it.

The base rate for this kind of metric disclosure is low. Companies typically introduce new reporting segments or metrics at annual investor days or when the metric becomes material enough to require separate disclosure under accounting standards. Twilio has 3 remaining opportunities, with SIGNAL being the most likely venue. However, Q4 2025 was a strong evidence point that management prefers qualitative over quantitative AI disclosure. Voice AI growing 60%+ is impressive as a percentage but 'relatively small base' language suggests the absolute dollar amount would not move the needle on valuation. The Stitch acquisition for identity verification is broadening the AI story beyond voice, which could delay rather than accelerate voice-specific revenue disclosure. Base case remains non-disclosure.

I'm anchoring on the prior prediction of 20% and adjusting for the consumed opportunity. One of four opportunities is now gone with confirmed non-disclosure. Naively, this should reduce probability by roughly 25% (from 20% to 15%). But the SIGNAL conference in May is a unique catalyst that wasn't weighted in the original prediction — developer conferences are where companies launch metrics alongside products. The Gartner 'company to beat' designation could motivate management to back up the claim with numbers at SIGNAL. Against this: the 'early days' framing, the usage-model complexity, and the calibration lesson about overestimating management changes. Net effect: roughly flat from prior at 20%, with the consumed opportunity offset by better understanding of the SIGNAL catalyst.

The question requires 'specific AI or Voice AI revenue metrics' — this is a high bar. It means a dollar amount, percentage of total, or run rate with a disclosed base. Management has shown they can provide impressive qualitative metrics (60% customer growth, 10x from top-10 startups) without ever crossing into quantitative territory. This selective disclosure pattern is well-established and serves management's interests: it creates AI excitement without the accountability of a specific number that analysts could track. The Q4 earnings call doubled down on this approach. With ConversationRelay still described as early-stage, there's no accounting materiality threshold that would force disclosure. Voluntary disclosure is a management choice, and management has clearly chosen not to make it.

Non-disclosure pattern confirmed through Q4 2025. Three opportunities remain: Q1 earnings, SIGNAL, Q2 earnings. SIGNAL is the best chance but management still describing AI as 'early days.' CEO language suggests they won't quantify until the number is large enough to impress. Low base rate for new metric disclosure in a 6-month window.

Simple pattern matching: management has been asked about AI revenue quantification repeatedly and consistently declines. One opportunity consumed with confirmed non-disclosure. Calibration feedback says we overestimate management behavior changes. Three remaining chances but same management team making same strategic choice. Conservative estimate below 15%.

Three remaining chances keep this above 10% — there's always some probability of a surprise disclosure, especially at SIGNAL. Voice AI growth above 60% means the absolute number is getting larger each quarter. If it reaches a threshold where disclosure becomes a competitive advantage, management may flip. But base case is continued non-disclosure. Setting at 18% reflects low but non-trivial probability.