AI adoption didn't creep up quietly. In roughly three years, it went from a Silicon Valley experiment to something 99% of Fortune 500 companies use in some form, a market on pace to roughly sextuple by 2030, and a genuine rewrite of how search itself works. Below are the numbers that matter most right now, organized around three questions: how far AI adoption has actually spread, how big the market has gotten, and specifically how AI is changing search — since that last question is what this whole site is built around.
Every stat is attributed to a named source and dated as precisely as the original research allows. Where sources disagree — and on several of the numbers that matter most, they disagree by a lot — this piece shows the range and names who's measuring what, rather than quietly picking whichever figure sounds most impressive. That's the same standard we hold agencies to elsewhere on this site, so it felt wrong to do anything less rigorous here.
Two years ago, "we're exploring AI" was still a hedge companies used on earnings calls. Today it's closer to a null statement — the numbers below show adoption moving past the experimentation phase into something closer to default infrastructure, even though exactly how fast depends heavily on what counts as "using AI" in the first place.
Share of companies using AI in some form, 2017 vs. today — Exploding Topics / Stanford HAI
That's up sharply from 55% just a year earlier, according to Stanford's own tracking. The jump suggests AI moved from pilot project to standard infrastructure in a single year.
Source: Stanford HAI, 2025 AI Index Report
Hiring and HR processes are the most common entry point. At this scale, AI use is effectively universal among the world's largest companies.
Source: Qualtrics
This figure includes both live deployments and active pilots, not production use alone. It suggests the remaining holdouts are shrinking fast.
Source: Exploding Topics
"Large" means 1,000+ employees, where deployment tends to move slower due to procurement and compliance. Even so, that's nearly half of big companies with AI already live.
Source: IBM
Generative AI specifically, not just AI broadly, has become mainstream inside a normal business function. Marketing, customer service, and content are the most common entry points.
Source: Stanford HAI
This is forward-looking intent, not current usage. Combined with the 78% already using AI above, it suggests near-universal adoption is close.
Source: Exploding Topics, via National University
That's the baseline worth remembering when today's adoption numbers feel inflated. Enterprise AI adoption has roughly quadrupled in under a decade.
Source: Exploding Topics
"Using AI" means something different to a company running one chatbot than one with AI across every department. When you see a single adoption number in the wild, check what actually counts as "using" first.
Methodology note — see sources for the full range
Every AI market-size forecast is really a forecast about scope as much as growth — whether you count chips and data centers alongside software changes the number by hundreds of billions of dollars. The figures below span some of the most widely cited estimates in circulation, and they disagree with each other more than most coverage of "the AI market" lets on.
Three widely-cited "global AI market by ~2030" forecasts — same category, different scope
That's up from an estimated $214 billion in 2024 — more than a six-fold increase in six years. It's one of the most frequently cited AI market projections in circulation.
Source: MarketsandMarkets, via Forbes Advisor
Grand View Research, Precedence Research, and UNCTAD each define "the AI market" differently — some include hardware and infrastructure, others don't. The forecast you see often depends more on scope than on genuine disagreement about AI's trajectory.
Sources: Grand View Research; UNCTAD
MarketsandMarkets and Precedence Research diverge by more than double for the very same year. The gap comes down to whether hardware and infrastructure spending count as part of "the AI market."
Sources: MarketsandMarkets; Precedence Research
PwC's often-cited estimate splits that gain between productivity improvements and increased consumer demand. It remains one of the largest single economic-impact figures attached to AI.
Source: PwC
This is a narrower, U.S.-specific estimate rather than the global PwC figure above. It still represents one of the largest projected impacts of any single technology on a national economy.
Source: Statista
IDC expects that figure to more than double past $500 billion by 2028. Investment is accelerating faster than most market-size forecasts assumed even two years ago.
Source: IDC
BlackRock's estimate covers data centers, chips, and the physical build-out AI models need to run. It's a reminder that "the AI market" isn't only software.
Source: BlackRock
Stanford HAI tracked this jump specifically in venture and private funding, separate from public infrastructure spending. It's one of the steepest single-year increases in the data.
Source: Stanford HAI
That's a projected 43.8% compound annual growth rate for AI systems that act autonomously rather than just respond to prompts. It's a small slice of the broader AI market today, but one of the fastest-growing.
Source: market.us
Enterprise adoption is one story; ordinary people picking up a chatbot on their own time is another. The data below suggests individual use of generative AI spread faster than almost any consumer technology in memory, though how fast varies enormously by country, income, and how honestly people report their own habits.
That's a rough global average, and it hides enormous variation by country and income level, covered below. Still, it puts generative AI's reach in a striking range only a few years after ChatGPT's launch.
Source: Simplilearn
That's up from 15.1% just six months earlier, per Microsoft's own tracking. The pace of growth, not just the raw number, is the notable part.
Source: Microsoft AI Economy Institute
Both figures already outpace how quickly personal computers were adopted at the same stage after launch. Personal use outpacing work use cuts against the assumption that AI adoption is mainly employer-driven.
Source: St. Louis Fed
Both countries outpace the U.S. and most of Europe on working-age population adoption. Smaller, wealthier, digitally-dense economies appear to adopt generative AI fastest.
Source: Microsoft AI Economy Institute
24.7% versus 14.1%, per Microsoft's tracking. The gap mirrors existing infrastructure and connectivity divides rather than introducing a new one.
Source: Microsoft AI Economy Institute
That's higher than most Western economies, including the U.S. It challenges the assumption that AI adoption simply tracks national wealth.
Source: Worklytics
Separate research found only about a third of people realize how often they're actually interacting with AI-powered features day to day. Most people who say they "don't use AI" are probably using it anyway, just unlabeled.
Sources: Pew Research; AIPRM, via National University
This is the section that actually explains why a site like this one exists. Search results pages have quietly become destinations in their own right rather than doorways to the rest of the web, and the numbers below trace how much of that shift is AI-driven versus a decade-long trend that AI simply accelerated.
Share of searches ending without a click, with vs. without an AI Overview — Similarweb
This SparkToro/Datos figure is generally treated as the most methodologically rigorous of several competing estimates. It means most searches are now resolved entirely on the results page.
Source: SparkToro / Datos, 2024 Zero-Click Study
The range comes down to differences in panel data, device mix, and whether Google-owned properties count as "a click." Whichever number you use, the direction is the same: fewer searches send traffic anywhere.
Sources: multiple zero-click studies, 2025–26 — see sources list
That compares to roughly 60% for ordinary results without an AI Overview present. The AI-generated answer itself is the single biggest driver of zero-click behavior.
Source: Similarweb
Pew Research's figure is the starkest number in this entire roundup. Being cited inside an AI Overview earns a brand impression almost never a visit.
Source: Pew Research
Sistrix's Germany-specific tracking and BrightEdge's broader estimate disagree by more than double. Both agree the share is rising quickly regardless of the exact baseline.
Sources: Sistrix; BrightEdge
Google reported at I/O 2026 that AI Mode query volume is more than doubling every quarter. It's still a small share of total Google searches today, but growing faster than any other single search feature.
Source: Google, I/O 2026
Mobile searchers are more likely to want a quick answer resolved directly on the page rather than a destination to visit. Since mobile makes up the majority of global searches, it pulls the overall zero-click average upward.
Source: Similarweb, device-split analysis
That's instead of going straight to a traditional search engine, per multiple 2025–26 consumer surveys. It's a meaningful behavioral shift, even though AI tools' absolute traffic still trails Google by a wide margin.
Sources: multiple consumer surveys — see sources list
This is a stated preference, not measured behavior, so it's worth treating as intent rather than fact. Still, a majority preferring an AI-first option is a meaningful signal for where search habits may be heading.
Source: consumer survey, via National University
For a while, "AI chatbot" was basically synonymous with ChatGPT. The numbers below show that's still mostly true by user count, but genuinely less true by market share than a year ago — and worth reading with one persistent caveat: almost every figure here comes from the company being measured, not an outside auditor.
Disputed AI-chatbot market share ranges, low-to-high estimate — Similarweb / First Page Sage
That's more than double the roughly 400 million reported just a year earlier. It's OpenAI's own reported figure, not an independently audited count.
Source: OpenAI
Similarweb's tracking shows real erosion even as ChatGPT's absolute user count keeps growing. Competitors are taking share of a fast-growing pie rather than shrinking ChatGPT itself.
Source: Similarweb
Different trackers measure different things — web traffic, chatbot referrals, or self-reported usage — and land on different answers. Treat any single "market share" figure for ChatGPT as directional, not exact.
Sources: multiple platform trackers — see sources list
The range depends on whether the count includes only the standalone Gemini app or also Gemini-powered features inside Search and Workspace. Alphabet does not publish one consolidated figure.
Source: Google / Alphabet
That makes Gemini the fastest-growing of the major AI platforms by share, even though it remains well behind ChatGPT. Multiple trackers agree on the direction even where the exact number differs.
Sources: Similarweb; First Page Sage
The gap depends on whether the count includes only the core app or all of Perplexity's surfaces, including its agent products and browser. Either way, Perplexity has more than doubled its user base within about a year.
Sources: multiple platform trackers — see sources list
That's up from 80% just two years earlier, per OpenAI's own enterprise figures. Enterprise AI adoption is consolidating around a small number of major providers rather than staying fragmented.
Source: OpenAI
The single-vendor era of enterprise AI procurement appears to be over. Most large companies are deliberately spreading usage across multiple providers rather than committing to one.
Source: enterprise LLM spend research
OpenAI, Google, and Anthropic each publish their own usage figures, typically timed to funding rounds or major announcements. Third-party trackers like Similarweb offer an independent check, but even they disagree with each other by wide margins.
Methodology note
AI adoption isn't spreading evenly across the economy, and it was never going to. The industries moving fastest tend to share two traits — large budgets and data-heavy operations — while the ones lagging behind are usually held back by regulation, procurement complexity, or both.
Share of organizations using AI, by industry — Stanford HAI, via Simplilearn
Financial services (71%) and healthcare (64%) follow closely behind. These three sectors combine large budgets with data-heavy operations that suit AI well.
Source: Stanford HAI, via Simplilearn
Government brings up the rear at 48%, likely due to procurement and regulatory complexity. The pattern roughly tracks each sector's existing digital maturity.
Source: Simplilearn
That's a larger projected gain than any other individual sector in the same research. Physical, process-heavy industries appear to have some of the most room to benefit.
Source: Forbes, via National University
PwC puts healthcare's long-term adoption rate at 40% and automotive's at 18%. Both combine safety-critical decisions with large amounts of structured data — exactly where AI tends to be deployed most cautiously but most thoroughly.
Source: PwC
Workforce statistics on AI tend to split into two camps that rarely get cited together: displacement estimates that sound alarming, and net-job-creation estimates that sound almost too reassuring. Both sets of numbers below come from credible research, and reading them side by side is more useful than picking whichever one matches your priors.
One widely-cited global jobs estimate tied to AI-driven automation — WEF-style projection, via National University
That's McKinsey's widely-cited estimate of jobs meaningfully affected by AI-driven automation through 2030. "Displaced" doesn't necessarily mean eliminated — many roles are expected to shift rather than disappear.
Source: McKinsey
That's a net gain of 12 million jobs globally by one widely-cited projection. It's a useful counterweight to displacement-only headlines, though both figures come from modeling, not observed outcomes.
Source: WEF-style estimate, via National University
Forbes frames this as transformation, not elimination — most jobs get reshaped by AI rather than replaced outright. Historically, that's been the more common pattern through past major technology shifts.
Source: Forbes
That's a majority of workers reporting some level of job-security concern tied specifically to AI. It sits in real tension with the more optimistic net-job-creation estimates above.
Source: Forbes Advisor
That's a meaningful acceleration on top of typical productivity growth, which usually moves in fractions of a percentage point. Statista's modeling attributes most of the gain to AI-augmented, not AI-replaced, work.
Source: Statista
PwC's estimate is among the higher productivity claims in circulation, and likely varies enormously by role and task. It's best read as a ceiling for well-suited use cases, not an average.
Source: PwC
Everything past this point should be read differently than the rest of this article. These are long-range forecasts with wide error bars, not near-term measurements, and they're included because they're widely cited — not because they're settled.
This is a long-range forecasting estimate, not a near-term prediction, and it carries wide uncertainty bands. It's included here because it's one of the most frequently cited AGI-timeline estimates in circulation, not because it should be treated as settled.
Source: AI-timelines research (Arxiv)
McKinsey's analysis extends this beyond knowledge work into medicine, creative fields, and civic governance. The framing is collaboration between humans and AI, not replacement.
Source: McKinsey
This is one of the more concrete, near-term applications inside otherwise long-range forecasting. It's a useful reminder that "the future of AI" includes unglamorous infrastructure problems, not just chatbots.
Source: McKinsey