Apple will reportedly allow other AI chatbots to plug into Siri

Apple's iOS 27 will introduce an 'Extensions' feature allowing users to connect third-party AI chatbots (e.g., Google Gemini, Anthropic Claude, OpenAI ChatGPT) to Siri on iPhone, iPad and Mac. The integrations will also work with a planned standalone, AI-upgraded Siri app and follow a reported partnership with Google to help power and train parts of Siri; Apple is expected to detail OS updates at WWDC starting June 8.

Analysis

Opening an OS-level voice assistant to third-party LLMs shifts the locus of value from device-level model ownership toward distribution and inference monetization. If even a small percentage of the billions of daily assistant queries route to a cloud provider, that can generate tens-to-low-hundreds of millions in incremental annual inference revenue while improving that provider’s models via scale effects; the winner is the firm that captures both inference volume and training access.

For the device owner, the tradeoff is between control and utility. Ceding model hosting reduces long-run Services margin leverage (pressure over 1-3 years) but can increase retention if the assistant materially improves utility — a modest retention lift (50–150bps annually) across the installed base would offset a non-trivial fraction of lost Services margin through higher hardware upgrade cadence and accessory spend. The split between on-device and cloud inference is the key micro — it determines whether silicon vendors or cloud AI vendors capture most upside.

Catalysts and risks are asymmetric in timing and severity. Near term (weeks–months) the market will re-price on product announcements and early opt-in/adoption metrics; medium term (3–12 months) usage telemetry and developer adoption will reveal monetization; long term (12–36 months) regulatory scrutiny around data-sharing and anti-competitive bundling could force contractual or technical changes that reverse early winners. Poor UX or weak opt-in rates can erase perceived gains within quarters, while favorable telemetry compounds advantages over years.

Key Decisions for Investors

• Long GOOG (Alphabet class A) 12-month directional exposure: allocate 1–2% portfolio via a 12-month call spread (buy ~25–35% OTM calls, sell ~60–80% OTM calls) to capture upside from increased inference/Cloud AI monetization while limiting premium spent. R/R: capped downside = premium (~100% of cost); potential multi-bagger on scale-driven rev acceleration if adoption is broad within 12 months.
• Pair trade — long GOOG / short AAPL equal-dollar position for 3–12 months: this isolates the distribution/training beneficiary vs the device/platform owner. Size to 1–3% net portfolio; stop-loss on pair if relative moves exceed 10% intra-month. R/R: asymmetric toward GOOG given higher per-toke monetization potential; hedge reduces macro beta.
• Convex infrastructure play: buy NVDA 6–12 month calls (small position 0.5–1% portfolio) to capture incremental demand for inference accelerators and cloud GPU cycles. R/R: limited cash premium vs high upside if cloud inference volumes ramp materially; cut if market signals point to predominantly on-device inference.
• Event/volatility play on the device OEM (AAPL): buy a short-dated (1–3 month) straddle or wide strangle around the next product/OS announcement with position size 0.5–1% to capture surprise either that preserves control or demonstrates improved utility. R/R: premium risk if announcement is telegraphed; potential 30–100%+ return on binary positive surprise.