Tech, Two Weeks On: AI deals, Apple signals, Microsoft refunds, cyber alarms, and Starship milestones

AI platforms — Google’s Gemini pushes into Chrome; OpenAI maps next steps

Google began rolling out Gemini features inside the Chrome desktop browser for U.S. users, pairing multi-step task automation with summarisation and tighter links to apps like Calendar and YouTube. The move follows a U.S. antitrust ruling that avoided a forced Google breakup but tightened some practices, so bundling will face scrutiny. However, Google’s near-term bet is clear: put a first-class AI assistant where billions already browse.

OpenAI, meanwhile, set out an autumn slate of platform and model updates, including security research tooling and infrastructure expansions. The company also detailed a continued Microsoft partnership and new safety work tied to GPT-5 and companion releases. As a result, developer attention is split between browser-native assistants and model-centric platforms.

Apple — earnings resilience and a November hardware hint

Apple reported fiscal Q4 revenue of roughly $102.5 billion and profit of $27.5 billion, underscoring solid iPhone and services performance into the holiday window. On the subsequent call, Tim Cook said Apple is “making good progress” on a next-generation Siri, aligning with broader industry momentum in on-device and cloud-assisted AI. However, Apple has not yet shipped the new assistant, choosing caution on quality and privacy.

Retail chatter points to an overnight refresh on November 11–12, a sign that new hardware or accessories may be imminent. As a result, watchers expect incremental Mac or audio updates rather than a full-scale platform reveal this late in the year. Meanwhile, multiple reports continue to track talks on using Google’s Gemini in “Private Cloud Compute” as a bridge while Apple advances its own models.

Microsoft — refunds in Australia and data-sovereignty push

Microsoft apologised to Australian customers over AI-related plan changes and pricing confusion and offered refunds after local coverage and customer complaints. The episode highlights how fast-evolving AI bundles can outpace clear communication, especially when features cut across Microsoft 365, Copilot and add-on tiers. However, the signal is constructive: the company is willing to unwind charges and simplify offers.

In parallel, Microsoft announced a wider in-country processing push for regulated customers. New “local” options will bring end-to-end processing to more markets by late 2025 and 2026, complementing the EU Data Boundary and adding GPU support for high-performance workloads. As a result, data-residency and sovereignty features are becoming competitive differentiators for cloud AI.

Cybersecurity — breaches climb and regulators lean in

Reports from Asia show a stark jump in exposed credentials in the Philippines during Q3, a reminder that rapid digitisation without commensurate security invites risk. Elsewhere, European digests flagged a run of attacks, fines and sandbox initiatives as regulators blend enforcement with controlled innovation. Meanwhile, one Polish fintech disclosed a significant incident, reinforcing that financial platforms remain prime targets. As a result, CISOs continue shifting budgets toward identity, data loss prevention and rapid incident response.

Luxury and insurance breaches earlier this year still cast a long shadow because leaked data fuels subsequent fraud. The broader pattern is familiar: repeated credential reuse and supply-chain gaps drive cascading compromises. However, coordinated disclosure and regulator guidance can limit downstream harm when paired with multifactor authentication and credential-stuffing defenses.

Consumer tech & policy — TikTok in flux; Alexa’s privacy debate

U.S. policy pressure on TikTok remains live. A federal appeals-court decision earlier this year upheld a divest-or-ban law, and separate reporting has described framework talks to shift ownership. The legal road is not finished, but the past year’s moves suggest the product’s U.S. future hinges on control and data-access guarantees. Meanwhile, the European Parliament has continued probing platform risks in parallel.

Amazon’s shift to Alexa+ earlier in the year also keeps sparking privacy debate, as multiple outlets documented the removal of a setting that stopped voice-recording uploads for analysis. Amazon says only a tiny fraction used it, and cloud processing enables richer AI features. However, critics argue the change reduces user agency and could erode trust without clearer controls.

Space & chips — Starship’s progress and the AI-hardware drumbeat

SpaceX completed Starship’s eleventh test flight on October 13, including releasing mock satellites and achieving a controlled splashdown. NASA has pressed for faster milestones ahead of Artemis lunar missions, and SpaceX now outlines a speed-up to hit lunar-lander commitments. As a result, 2026 ambitions hinge on iterative upgrades and regulatory green lights.

On the silicon front, Nvidia’s 2025–2027 roadmap remains a core narrative, with Blackwell Ultra shipments in 2025 and next-gen architectures queued behind it. While stock-price punditry varies, consensus across technical coverage is simple: hyperscale AI demand is still outpacing supply, keeping accelerator cycles in the spotlight into 2026.

The bottom line — consolidation, clarity and credibility

Across these two weeks, one theme ties stories together: the race to integrate AI everywhere, while proving trustworthiness. Google is embedding Gemini into a default surface—Chrome—just as OpenAI courts developers with platform tools and safety updates. Apple is signalling fresh hardware and a more capable Siri without rushing release. Microsoft is learning that clear pricing and sovereign processing matter as much as flashy features. Meanwhile, cyber incidents and policy fights—from fintech hacks to TikTok—keep testing whether consumer trust can keep pace with innovation.

For readers and businesses, the near-term checklist is pragmatic. Update browser and account controls as AI features arrive by default. Recheck cloud and email configurations for data-location and retention policies. Use passkeys or strong 2FA, not just passwords. And, as holiday-season gadgets drop, scan the privacy settings on day one. The best tech wins of 2025 will be the ones that combine capability with credible guardrails.

Sources: Reuters/OpenAI