How Google Quietly Rebuilt Its AI Edge and Won Apple's Siri Deal

Headlines in early 2026 framed Google as moving from AI's "biggest loser" to a potential leader. The company's market capitalization briefly surpassed Microsoft, nudged Apple, and approached NVIDIA, set against Microsoft's sharp market cap drop, NVIDIA's stagnation, OpenAI's reported panic, and Apple's choice of Gemini for Siri.

The Three‑Horse Narrative Before Google’s Resurgence

For a long period the dominant AI story centered on three players: OpenAI with ChatGPT, Microsoft as the major funder, and NVIDIA as the primary hardware supplier. Investors funneled money into Microsoft and NVIDIA on the expectation of clear profit paths tied to OpenAI and ChatGPT, which was viewed as the catalyst of the AI boom and a major chip buyer. Other companies—AMD, Oracle, Amazon, Meta—existed in the market but were not central to that narrative.

Google's Early Foundations and Public Perception

Google researchers authored the 2017 paper "Attention Is All You Need," cited over 173,000 times as of 2024, a foundational contribution to modern AI. Despite that, seven of the eight co‑authors left Google, with six founding startups and one joining OpenAI; Aidan Gomez cited limited freedom for product innovation. Internally, Jeff Dean framed Google as facing greater "reputational risk" and moving more conservatively than startups, while CEO Sundar Pichai emphasized "slimming down" after past failed investments such as Google Glass, Project Loon, and others. Early public AI products like Bard and initial Gemini launches encountered problems—Bard's demo hallucination reportedly removed $100 billion from market cap—and reinforced a perception that Google had made wrong moves from 2023–2025.

The Turning Point: Apple Chooses Gemini

In January 2026 Apple selected Google Gemini to power Siri, a decision presented in the brief as shockingly decisive and bypassing alternatives like ChatGPT, Llama, and Grok. Siri had long struggled with delayed upgrades and internal disarray; "Apple Intelligence" was described as embarrassing and Siri was treated as a critical area needing a major upgrade. The Apple–Gemini choice is framed in the brief as a single decision that changed the tech hierarchy.

Google’s Quiet Full‑Stack Transformation

Google is described as having built a full AI pipeline: LLMs (Gemini), custom chips (TPUs/Trillium), global data centers and networking, and software tooling. That stack serves billions of tokens daily via Cloud APIs, reportedly lifting Google Search engagement by about 20%. Enterprise API call volumes are said to rival OpenAI's, and ad tools are auto‑optimizing billions in advertiser spend, indicating broad internal deployment of AI capabilities.

Why Apple Picked Gemini: Stability and Supply Chain Considerations

The brief frames Apple's choice as driven by stability and supply‑chain resilience. OpenAI's dependence on NVIDIA chips, and NVIDIA's dependence on TSMC, are presented as potential points of instability, along with reports that OpenAI might need $1 trillion in NVIDIA/AMD deals. Apple needed a rapid, significant Siri upgrade and preferred a partner with less apparent risk of collapse due to GPU dependencies—attributes the brief assigns to Google.

TPU Dominance and Trillium Chips

Google's Trillium chips—its sixth‑generation TPUs—are presented as offering 4.7× the performance per watt of an NVIDIA H100. Anthropic is reported to be renting 1 million of Google's chips, which the brief frames as cheaper than standard GPUs. These hardware advantages are presented as making Google a logical partner for large‑scale AI deployments like Siri.

Gemini's Reach and Monetization

The standalone Gemini app reportedly reached 750 million monthly active users in early 2026 and had 1.18 billion monthly visits in 2025. Around 70% of Google Cloud customers are said to use Gemini, and Google Cloud growth reached 35% in the quarter referenced. The brief reports a $106 billion Google Cloud backlog, with Google as a whole hitting $105 billion in quarterly revenue, 31% profit margins, and $53 billion in AI spending—framed as evidence that AI is becoming a revenue source for Google rather than only a cost center.

Microsoft’s AI Spending and Financial Strain

Microsoft is described as struggling to monetize AI despite early access to ChatGPT. The brief cites a one‑day market cap drop of $360 billion and a capital spend increase of 66%, with $37.5 billion spent in one quarter—two‑thirds on chips. Eric Clark, a portfolio manager, is quoted noting revenues up 17% while cost of revenues rose 19%, framing concern that AI spending has not yet translated into commensurate revenue.

Investor Sentiment, Valuation, and the Closing Gap

From 2021–2024 Google's stock rose about 20% and remained relatively sideways until mid‑2025, after which it "went straight up" as AI success was perceived as underpriced. By contrast, NVIDIA and Microsoft had valuations the brief describes as already "priced for hype" and thus more constrained, leading to slower stock movement. Investors are characterized as penalizing high AI burn rates without clear returns—seen in Microsoft's fall—and rewarding companies showing concrete monetization. The brief frames a view that "boring incumbent players" with established brands, user bases, and infrastructure will fare best, and that Google is catching up to where it should be and "just became a real threat" in the AI landscape.