Zan Digital › Companies
Every piece of Zan Digital analysis that covers Google — 85 pieces, each figure traced to a named source.
- 001AI Capex 2026: The $725B Bet With No Payback MathFour hyperscalers plan roughly $725B of capex in 2026, up 77% on 2025. Here is where the money goes, what it earns today, and the three numbers to watch.$725B
- 002Alphabet's $80B Equity Raise: Why Cash Ran ShortAlphabet raised $80 billion in equity in June 2026. Its filings show free cash flow fell 82% in six months and buybacks stopped at $0. What that signals.82%
- 003Anthropic IPO: The Enterprise Listing That Resets PricingAnthropic filed confidentially on 1 June 2026 for an October listing. A $47B run rate, 85% enterprise revenue, and why the mix matters more than the number.$47B
- 004GPU Depreciation: The $176B Accounting Question in AIHyperscalers depreciate AI hardware over 5 to 6 years. One investor puts the resulting overstatement at $176B through 2028. Both sides of the argument, checked.$176B
- 005The $662B Data Centre Leases Off Big Tech Balance SheetsMoody's found $969B in hyperscaler lease commitments, $662B of it not yet on any balance sheet. What the rule is, why it matters, and when it arrives.$969B
- 006The AI Chip Selloff: Dress Rehearsal or False Alarm?Chip and AI hardware stocks fell sharply in late July 2026 and recovered within a week. Fast recoveries signal conviction or complacency. Five indicators.5 days
- 007Agent Marketplaces: What Distribution Is Actually WorthEvery major platform launched an agent marketplace. In one editor's data, 3 agents took 92% of turns. What marketplace distribution is really worth.92%
- 008AI Moats in 2026: Integration Depth Beats Model DepthThe top 4 AI labs sit within 25 Elo points of each other. Model quality no longer separates products. Integration depth does, and here is the evidence.25 Elo
- 009Data Moats Are Overrated. Workflow Moats Are NotMost proprietary data can be bought or approximated. Palantir grew US commercial revenue 149% in Q2 2026 by owning the workflow, not the data corpus.149%
- 010Horizontal SaaS in 2026: Five Tests Before You StartGoogle folded a $20 per user AI add-on into a $2 price rise. Five tests that decide whether a horizontal SaaS company is still worth starting in 2026.$20
- 011Incumbent Advantage: Why Worse Enterprise AI Keeps WinningGiven a free choice, just 8% of workers pick Copilot. Given only Copilot, 68% adopt it. Here is the incumbent advantage machine, and how challengers break it.8%
- 012MCP: The Standards Fight That Decides the Agent StackMCP went from one company's experiment to a Linux Foundation standard with 97M monthly SDK downloads. What it settles, what it does not, and who benefits.~97M
- 013Micro SaaS Economics: What One Person Can Actually RunClosed micro SaaS deals change hands at a 3.9x median profit multiple. What AI made cheaper, what it did not, and where the one-person software business stops.3.9x
- 014SaaS Categories AI Is Erasing, and 5 Getting StrongerWe scored 15 SaaS categories on origination, write authority, liability and substitutable work. 10 are being erased, 5 get stronger, and here is why.-57%
- 015AI Agent Liability: Who Pays When It Gets It WrongStandard AI contracts cap vendor liability at the fees you paid. Deloitte refunded 22% of a A$440,000 report. What buyers can actually recover in 2026.C$650.88
- 016AI Contract Negotiation: 6 Clauses That Beat DiscountFederal AI contract terms now specify 72-hour incident notice and 30-day model change notice. Six clauses CFOs are winning in 2026, and what each is worth.72 hours
- 017AI Provider Contracts Compared: Anthropic, OpenAI, GoogleAnthropic, OpenAI and Google differ on retention, training rights, indemnity and uptime. None of the 3 publishes a credit-backed uptime SLA on standard terms.99.48% v 99.94%
- 018AI Shrinkflation: Same Price, Fewer Tokens, Quieter LimitsAI shrinkflation is real but mis-stated. Vendors owe API buyers 60 days notice before a model dies. Subscribers got 31 before limits changed. How to detect it.Under 5%
- 019Rate Limits as a Pricing Lever: A 5-Test StandardGitHub's premium request multipliers span 173x, from 0.33 to 57. A 5-test transparency standard for AI rate limits, with four vendors graded against it.173x
- 020AI Chief of Staff: 3 Tasks That Transfer, 4 That Do NotMeeting agents now handle scheduling, notes and follow-up. On 175 real office tasks the best agent finished 30%. Here is what transfers and what does not.30.3%
- 021AI Data Retention Policies, Ranked by What They PromiseEvery major AI vendor's data retention policy, scored on 5 criteria. Anthropic keeps covered-model prompts for 30 days even where zero retention was agreed.30 days
- 022AI for Excel Hits 34.9% on Multi-Sheet Model TestsOn real multi-sheet financial models, the best AI for Excel scores 34.9%. Here is what four benchmarks measure, where they disagree, and how to test.34.89%
- 023AI Note-Takers: $5,000 a Violation If Consent Is WrongCalifornia sets damages at $5,000 per violation. A federal court held on 13 August 2026 that training on recordings makes a note-taker an eavesdropper.$5,000
- 024Autonomous Coding Agents Tested on Real RepositoriesTop systems resolve 61.5% of public benchmark tasks but 51.5% on private commercial codebases. What Devin, Factory and agent buyers should take from it.61.6%
- 025ChatGPT vs Claude vs Gemini Enterprise: Admin ControlsChatGPT, Claude and Gemini Enterprise are close on model quality. Governance decides: admin controls, 180-day audit logs, 10 data residency regions.180 days
- 026Claude Code vs Cursor vs Copilot: The 2026 VerdictGitHub Copilot's workplace adoption fell from 29% to 21% while Claude Code rose to 39%. A buyer-side comparison built on adoption data, not benchmarks.39%
- 027Deep Research Tools Compared: Citation Accuracy AuditedCitation accuracy across deep research agents ranges from 77.96% to 93.68%, and the volume leader ships 13.3% dead URLs. What to measure before you buy.93.68%
- 028GitHub Copilot Market Share: The 51% Drop Nobody Can SourceThe claim that GitHub Copilot fell from 67% to 51% of developers is not in the survey it cites. The measured decline is 29% to 21%. Here is the evidence.21%
- 029Glean vs Copilot: The Permissions Gap Nobody DemosGlean vs Copilot is decided by permission plumbing. Microsoft syncs connector ACL changes only on a full crawl, daily by default. Glean falls back to 24 hours.30M
- 030Harvey vs Legora vs General AI: A Use-Case ScorecardLegal AI and general assistants both hit 80% research accuracy against a 71% lawyer baseline. How to score Harvey, Legora and ChatGPT by use case.80% vs 71%
- 031Model Routing Savings: 31.7% Measured, Not 85%The largest independent routing benchmark measured 31.7% cost savings, not 85%. Here is the quality gap across five enterprise tasks, and where it hides.31.7%
- 032Perplexity Enterprise vs Google AI Mode for ResearchPerplexity had the lowest error rate of 8 AI search tools tested and still missed 37% of citations. For research teams, source control decides this one.37%
- 033Agent Drift Is Silent. Only 14% Monitor Production AIOnly 14% of teams monitor LLM apps in production. Agent drift comes from model updates, data shifts and prompt sprawl. How to detect it before customers do.14%
- 034AI Agent Payback: Why Every Function Says 6 MonthsSix commissioned studies report AI agent payback in under 6 months across six functions. The one with 420 respondents says 15. What that gap means.6 of 7
- 035AI Procurement Checklist: 5 Sections Most RFPs MissGartner expects over 40% of agentic AI projects to be cancelled by 2027. A 5-part procurement checklist for capability, data, cost, evals and exit terms.40%+
- 036AI Vendor Lock In: What Switching Models Really CostsAnthropic retired 9 models in 10 months, one on 61 days notice. The API call is the cheap part. Here is what model switching actually costs to redo.40%
- 037Context Engineering Replaced Prompt Engineering. Mostly.Anthropic shipped a 1M token window at standard pricing in March 2026, and its own best model still misses 21.7% of MRCR v2. Context engineering is why.78.3%
- 038Data Freshness: The AI-Ready Data Layer Nobody OwnsTop models override their own correct answer over 60% of the time when the retrieved document is wrong. Data freshness is a governance job, not a model one.60%
- 039Legacy Integration: The Real Reason AI Agents Fail35% of large enterprises name data and integration as the top barrier to scaling AI agents. Why write access to legacy systems is the real constraint.35%
- 040LLM Evals: The Only Asset That Survives a Model SwapAnthropic retired 9 Claude models in 10 months and gives 60 days notice. Your eval suite is the only asset that survives. How to build one from scratch.60 days
- 041Multi-Agent Systems: 15x Tokens for a Narrow PayoffMulti-agent systems use about 15x the tokens of a chat, and lose to single agents under matched budgets. The 4 tests that decide the architecture.15x
- 042RAG Is Not Dead: 11 of 13 Models Fail at 32K TokensLong context did not kill retrieval. At 32K tokens, 11 of 13 models fell below half their short-context accuracy. How to diagnose and fix bad RAG.11 of 13
- 043Scaling AI Agents: What the 25% Did DifferentlyOnly 25% of organisations moved 40% or more of their AI pilots into production. Deloitte, McKinsey and BCG agree on what that minority did differently.25%
- 044Agent Identity: 109 Machine Accounts Per Human UserMachine identities outnumber humans 109 to 1, and 41% of identity breaches trace to weak non-human identity management. Here is what actually works.109:1
- 045AI Copyright Lawsuits: The $1.5B Repricing of Training DataAnthropic paid $1.5 billion, about $3,000 a book, to settle. Here is what the live AI copyright cases could do to model economics and licensing prices.$1.5B
- 046AI Data Residency: What EU and India Buyers Demand51% of firms now rate data sovereignty as very important. What EU and Indian buyers actually demand, and where vendor residency claims stop short.51%
- 047AI Export Controls: The 19 Days Claude Went DarkUS export controls pulled Claude Fable 5 offline worldwide for 19 days in June 2026. The order, the reversal, and the precedent it set for your stack.19 days
- 048Data Processing Agreement Gaps: 7 Clauses to RewriteMost DPAs predate agent tooling and model updates. Here are the 7 clauses to rewrite, with the EDPB findings and vendor terms that made them stale.38%
- 049EU AI Act Transparency Rules: The 2 August ChecklistArticle 50 transparency duties applied from 2 August 2026, with fines to €15M or 3% of turnover. Only 9 of 27 member states have both enforcers named.€15M or 3%
- 050Open Source AI Licence: The 4 Clauses Nobody ReadsMost open weight models are not open source. 4 clause types decide commercial use, from Llama's 700 million user cap to terms you must pass downstream.700M
- 051Personal AI Accounts: 67% of AI Use Is Off Your BooksVerizon's 2026 DBIR found 45% of employees now use AI on corporate devices, and 67% of them sign in with personal accounts. Source code leads the leak.67%
- 052Prompt Injection: 12 Defences Broken Above 90%Researchers broke 12 published prompt injection defences with attack success above 90%. Why the SQL injection fix has no equivalent, and what actually works.Above 90%
- 053Prompt Injection: 4 of 26 Models Paid the AttackerZscaler ran 26 models against live injected pages and 4 paid the attacker. Why prompt injection is the agent era's SQL injection, and what contains it.4 of 26
- 054AI Content Saturation: 50% of Articles, 14% of ResultsRoughly 50% of new online articles are AI-written, but only 14% of articles ranking in Google are. Here is what actually survived the content flood.49.9%
- 055AI Pricing Accuracy: Fix What ChatGPT Gets WrongAI assistants ship a significant issue in 45% of answers. Here is why they misquote SaaS pricing, and a six step fix for the pages they actually read.45%
- 056AI Referral Traffic: 4.4x Is a Model, 1.3x Is MeasuredThe 4.4x conversion claim for AI referral traffic is a model output. The datasets that publish a denominator land near 1.3x. What to measure instead.1.31x
- 057AI Search Citations: Top-10 Overlap Fell 76% to 38%Ahrefs measured the same thing twice: 76% of AI Overview citations ranked top 10 in 2025, 38% by March 2026. What now decides whether you get cited.37.9%
- 058AI Shortlist: 94% of B2B Buyers Now Start There94% of business buyers now use AI in the buying process, and about 90% buy from their Day 1 list. What decides whether your name is in that answer.94%
- 059AI Visibility Tracking: Build the Dashboard in a DayThree AI models agree on the top recommended brand just 41.6% of the time. Here is the free AI visibility dashboard you can build in a single day.41.6%
- 060Block AI Crawlers? What robots.txt Actually ControlsCloudflare measured Anthropic crawling 70,900 pages per referral sent back. Blocking AI crawlers is defensible. robots.txt controls less than you think.70,900:1
- 061Brand Mentions Beat Backlinks in AI Search: 0.66 vs 0.22Ahrefs studied 75,000 brands: web mentions correlate 0.664 with AI Overview visibility, backlinks 0.218. What that means for your community budget.0.664
- 062ChatGPT Citations Ignore Google Rank: What They FollowNearly 90% of ChatGPT-cited pages rank 21 or worse on Google. Four measured traits separate cited pages from ignored ones, and authority is not one.90%
- 063Cold Email Deliverability: 5,000 Sends Is a One-Way DoorGmail and Outlook both treat 5,000 sends a day as a permanent classification. AI SDR volume pushed ordinary senders over it. What still lands in 2026.5,000
- 064Cold Email Deliverability: What AI Volume Actually BrokeGmail and Microsoft's bulk sender rules cover consumer inboxes, not B2B. Vendor data on 53.1 million cold emails shows what actually collapsed in 2026.0.3%
- 065Comparison Pages: The Rules Before You Name a RivalThe FTC has encouraged naming competitors since 1979, and peer-reviewed work ties citations to a 30% to 40% visibility lift. Here is the template and the rails.1979
- 066Content Marketing 2026: 4.7x Cheaper, 23% Fewer ClicksAI cut the cost of a blog post from $611 to $131. Over the same period, the share of Google searches that send any click fell 22.9%. Here is what survived.4.7x
- 067Gated Content Math: 99% of Your Leads Never CloseForrester puts inquiry-to-close at under 1%, so the gate fails 99% of the time. Here is what the form costs in AI visibility, and what replaces it.Under 1%
- 068GEO vs SEO: 45 Studies on What Actually Gets CitedA survey of 45 GEO studies found body-only rewrites cut citations by 6%. Here is what the peer-reviewed evidence says drives AI citation, and what does not.3 of 54
- 069LinkedIn AI Slop: 40% of Long Posts Are AI-WrittenPangram flagged 40% of long LinkedIn posts as fully AI-written. The engagement data splits by topic: slop wins on inspiration, loses badly on analysis.40%+
- 070Marketing Attribution Broke: 68% of Searches End Unclicked68% of US Google searches ended without a click in early 2026. Here is which attribution methods survive dark discovery, and which ones quietly invent numbers.68.01%
- 071Original Research Is the Only Content Moat LeftA 200-response study costs about $457 at published panel rates. Why original research is the only content moat left, and how a small team runs one.$457
- 072Paid Search CTR Fell 68%. Do Not Move the Budget YetSeer measured a 68% paid CTR fall on AI Overview queries across 3,119 terms. Its larger 5.47M query update reads 16.21%. What that does to your budget.68%
- 073Product-Led Growth Metrics Break When Agents Sign UpNetlify hit 16,000 daily signups and 96% did not come from AI coding tools. Agents still break activation, PQL and expansion. A replacement metric set.50%+
- 074Reddit Marketing and Perplexity Citations: Where It FailsReddit is 46.7% of Perplexity's top-10 citations, per Profound's 680 million citation study. What legitimate participation looks like, and where it stops.46.7%
- 075Sales Engineering Is the New Marketing: 94% Now Verify94% of B2B buyers fact-check AI research before deciding. That makes demo environments and public benchmarks top-of-funnel assets, not sales tools.94%
- 076Software Review Sites Sold for $110M: The ReckoningGartner wrote off $150M on Digital Markets, then sold Capterra, Software Advice and GetApp to G2 for $110M. What that says about software review sites.$110.0M
- 077Zero-Click Search Hit 68%: What to Report Instead68.01% of US Google searches ended without a click in early 2026, up from 60.45% two years earlier. The reporting metrics that replace organic sessions.68.01%
- 078AI Tool Standardisation: What One Tool Really Costs70% of engineers run two to four AI tools at once, per a 906-person survey. Standardising improves governance and costs you goodwill. Here is the trade-off.70%
- 079Code Review Is the Bottleneck: Time in Review Up 441%Faros AI telemetry from 22,000 developers puts median time in PR review up 441.5%. Generation outran review capacity. Here is how to rebuild the workflow.441.5%
- 080Developer Trust in AI Fell to 29% as Usage Hit 84%Stack Overflow's 2025 survey puts developer trust in AI accuracy at 29% while 84% use or plan to use the tools. What that gap really costs tool vendors.29%
- 081Engineering Burnout: AI Raised Pace, Not Headcount49% of software engineers now feel emotionally drained weekly, up from 39%. AI raised delivery pace without changing team size. What managers can fix.49%
- 082Generative UI vs Designers: The Craft Moves to ConstraintsGoogle's generative UI scored 1736 Elo against 1800 for human-designed sites. What runtime interfaces change about design systems, craft and hiring.1736.2
- 083Product Manager Skills When AI Writes 75% of CodeAI now writes 75% of new code at Google. That moves the product manager job upstream, toward specification quality, evaluation design and problem selection.75%
- 084Technical Interviews When Everyone Ships With AI71% of engineering leaders say AI has made technical skills harder to assess. Portfolios and take-homes stopped signalling. Here is what replaces them.71%
- 085Vibe Coding's Cleanup Bill: 74% Report Major Rework74% of technology leaders say at least a quarter of their AI-generated code needs significant rework. Here is what the cleanup costs and how to cap it.74%