business

The a16z Show

Podcast

Cleared insights: 497
Published protocols: 0
Sources analysed: 10

WHY · 497 cleared insights from 10 analysed sources · 0 protocols

+ Add to my sources Submit content from The a16z Show

Extracted

What cleared the bar

Every result below came from this creator's own material, with its evidence attached.

Report · Money

Most pure consumers still use AI as a search replacement plus education use cases; startup successes cluster around 'getting things done'

The vast majority of pure consumers are basically using AI still as a replacement for search products like Google, plus education use cases — kids using it to write essays and for homework help — and people in the workplace to write emails. Startup successes in consumer AI so far cluster around the theme of getting things done: productivity, photo/video, design and editing, search and answers. AI has been an incredible enhancer for people who code, and that shows up in the top-100 data; people who've used AI for coding have had great ideas for other areas of their life.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Descriptive claims from the speaker's analysis.

View full analysis →Listen to the original ↗

Report · Money

Only 11 new products appeared across the web and mobile lists — the fewest ever — but existing entrants are getting much bigger

The current report had only 11 new products across the web and mobile lists combined, the fewest in any prior edition, which can make the space look like it is ossifying. However, the speaker notes that even without many new entrants, all existing entrants are getting much bigger, showing AI is showing up everywhere, and that things have actually been picking up in the past few months. On a traffic basis many AI products appear to be settling.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Interpretive framing around 'picking up' is the speaker's read.

View full analysis →Listen to the original ↗

Report · Money

The biggest new trend is personal agents: AI shifted from productivity tool to getting things done, with Muse and Instinct leading

The most interesting big new trend is personal agents — agents inside messaging apps that turn tasks into conversation, extending ChatGPT's paradigm of turning things into conversation. AI shifted at the end of last year/beginning of this year from a tool to enhance productivity to something that actually gets things done for you. The biggest developments in pure consumer assistance have been Muse and Instinct; the consumer assistant trend is very recent (last three weeks to a month), so many new products haven't appeared on rankings yet but show very fast growth without reaching mainstream users. The speaker expects massive adoption of consumer AI assistants by the next report, and notes consumer AI assistants have crossed the threshold where you can safely recommend them to non-technical

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Trend timing is the speaker's characterization.

View full analysis →Listen to the original ↗

Principle · Money

Treat consumer AI as a power-user market: a small share of heavy spenders drives most revenue, so optimize for the top cohort rather than average users

The other big conclusion of the report is just how much of a power-user game consumer AI is right now — a small share of heavy spenders drives most revenue, so builders should optimize for the top cohort rather than average users. This is the speaker's actionable takeaway from the spend-concentration data. Today's heavy AI spenders ($900/month power users) should be treated as harbingers of mainstream behavior as more people feel empowered to create.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Interpretation of the report's own data.

View full analysis →Listen to the original ↗

Principle · Money

Build products that get better as models get better but don't get cannibalized by them — and whose utility improves automatically as models improve

The key startup question is how to build a product that gets better as models get better but doesn't get cannibalized as models get better — that's the sweet spot. Model capability improvements over the past year-plus drove much of AI product growth, and products built on models explode in demand and utility when model capabilities improve: you build something great that may not do everything you want because model capabilities weren't quite there, then the models get better and your product explodes in demand. Value is moving back to the software layer now that capable models are widely accessible — build rich products that leverage the models rather than competing on model capability. In AI products, context and community become the asset; the experience layer, network, and ways of

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Principles/arguments from the speaker.

View full analysis →Listen to the original ↗

Report · Money

Consumer AI is a power-user market: usage is broad but payment is narrow and extremely concentrated

Roughly half of Americans report using AI and about 25% think they interact with it near-daily, yet only about 4.5% of U.S. consumers pay a subscription to an AI product (per the Yipit panel; other sources put it as low as 2–2.5%). Within paying users, spending is heavily concentrated: the top 10% drive more than half of revenue, the top 1% account for 20%, and the bottom 50% contribute only about 16%. The top 1% user spends $903/month on AI from personal credit cards while the median paying user spends $25/month. The report frames this as a power-law game happening at both the enterprise and consumer level.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Figures are described as panel/survey data with some numeric-error flags; the 4.5% figure has conflicting lower estimates.

View full analysis →Listen to the original ↗

Report · Money

The Top 100 report added consumer card-spend data for the first time, and spend rankings diverge sharply from traffic rankings

The current Top 100 AI apps report added consumer card-spend data for the first time (working with Abit data) and ranked products by what consumers actually pay for. Of the 50 products ranked by consumer spend, 29 were not on either traffic list. The first AI-native ranking was based purely on traffic, which worked when nearly all AI use was prompt-box products in a browser; the methodology has had to evolve. Going forward, revenue rankings may be even more useful than traffic rankings because usage is concentrated in paying power users.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Publisher-reported methodology details.

View full analysis →Listen to the original ↗

Report · Money

Trust is the biggest constraint on AI assistants: no product has unlocked person-to-person network effects, and intimacy of shared data creates tension against social features

Lack of trust in AI assistants with intimate access (email, life, credit card) is the biggest impingement on them growing as massive consumer apps. No assistant product has yet unlocked person-to-person network effects: the more useful the product is, the more it knows about you, creating tension around sharing that data with other people. Many people have separate ChatGPT work and personal accounts, and no assistant product has quite cracked that separation yet. Privacy, safety, and security are becoming more and more important as consumers adopt AI. People are naturally good at knowing what to share, protect, and pass on, and are now suddenly expecting software to do the same. Platform capabilities are improving faster than true consumer willingness to use them.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Analytical claims by the speaker.

View full analysis →Listen to the original ↗

Creator perspective · Money

Consumer AI revenue is almost entirely subscriptions today — an inversion of historical internet norms — but the speaker expects ads and transaction fees to return as inference and

Substantially all consumer AI revenue so far has come from direct subscriptions, an inversion of the historical consumer internet pattern where big companies monetize via ads or transaction fees. About 85% of top consumer AI web products monetize via subscriptions, ~62% via credits/token extra-usage payments, and only ~13% via ads or 'you are the product' models. The speaker is a big believer that the industry has transcended the need for everyone to buy an AI subscription and needs to see other business models (ads, transaction fees) come back — most people don't have or won't spend funds on software. The historical playbook of building user density first at near-zero serving cost and monetizing later breaks when COGS are high; the default early approach became 'if serving costs a lot, at

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Monetization percentages flagged as possible numeric errors; the shift is the speaker's belief.

View full analysis →Listen to the original ↗

Principle · Money

Treat user trust as the primary growth constraint for consumer AI agents — and design agents to follow human information-sharing norms

If users don't trust the product with email, life data, or credit cards, adoption stalls regardless of capability — that's the biggest impingement on these things growing as massive consumer apps. Design AI agents to follow human information-sharing norms: knowing what to keep private, what to pass on, and what to protect — users expect software to behave like a trusted person. This pairs the trust constraint with a design principle.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Heuristic from the speaker.

View full analysis →Listen to the original ↗

Principle · Money

Consumer AI startups should meter usage-based charging from day one — growing fast without meters creates out-of-control COGS

Consumer AI companies that grow too fast without meters on how they charge for usage can face out-of-control serving costs that are very hard to manage. This is an anti-pattern warning for founders: meter usage-based charging from day one.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Warning based on the speaker's observation.

View full analysis →Listen to the original ↗

Report · Money

ChatGPT remains the dominant consumer AI by usage and revenue, with 1.2B weekly active users and a top-20 global subscription rank

ChatGPT is the dominant global consumer AI by both usage and revenue, a conclusion that continues from the first report. Its last reported figure was 1.2 billion weekly active users, and it is already among the top 20 consumer subscription products globally after roughly 3.5 years of growth — the rest are media companies or frequent-purchase platforms like Amazon and Uber. On the web, ChatGPT is around 6x ahead of Claude and 2x ahead of Gemini. Everything below ChatGPT has seen more movement and shakeups.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

WAU figure cited as 'last reported a few days ago'; multiples are web-traffic based.

View full analysis →Listen to the original ↗

Report · Money

Assistant Benchmark tracks 170+ (now ~200) agents with user rankings, and its 1,500+ early-adopter community shows coding/technical automation as the #1 use case

The speaker has been working with David Paolon, who runs Assistant Benchmark — a website tracking over 170 agents (around 200 by the following morning) where users rank agents on task success. Paolon also hosts group chats and communities of 1,500+ early-adopter agent users, and pulled data for the report: the number one use case for consumer AI agents, even for assistant products like Muse and Instinct, is coding and technical automation.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Community size hedged ('I think it's 1500 plus'); limited to early adopters.

View full analysis →Listen to the original ↗

Report · Money

Open Evidence, an AI product for doctors with 50–60% density of US physicians, has seen early success in ads — density of a valuable targeted audience enables effective advertising

Open Evidence, an AI product for doctors, has seen early success in ads and has achieved 50–60% density of all US physicians. When you have density of a really valuable targeted audience, you can advertise really effectively. The speaker expects there will be and should be more verticalized AI products with valuable targeted audiences, and that better ad targeting will support more focused, verticalized consumer AI products.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Density figures cited as 'most recent numbers.'

View full analysis →Listen to the original ↗

Report · Money

Serving-cost reality check: mainstream consumer assistants cost tens of dollars per user per month, not thousands — the thousand-dollar figures come from power-user coding work

When talking to agent/assistant products aimed at the mainstream consumer, the speaker found costs in the tens of dollars per month, not thousands. The thousand-dollar figures come from power-user coding/automation workloads. When evaluating AI product cost claims, check who the users are first — heavy coding/automation users inflate serving costs by orders of magnitude versus mainstream consumers.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Based on founder interviews; conflicts with founder-posted high figures are reconciled by user-mix.

View full analysis →Listen to the original ↗

Principle · Money

Build products that compound personal value — accumulated context about the user becomes lock-in competitors can't easily replicate

The speaker thinks a lot about compounding personal value, and hopefully eventually compounding multiplayer value, that becomes lock-in for startup products. Products that build playbooks of who you are and what you sound like create migration lock-in — you cannot easily take that accumulated context to a competing product. The difference between an email that's 99.9% you and one that's 85% you is the difference between 10 seconds and 10+ minutes of fixing — that gap is why high-fidelity personalization creates lock-in. Expect more successful products that collect context about the user and act on it to build personal playbooks.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Principle/argument from the speaker.

View full analysis →Listen to the original ↗

Report · Money

Instinct traction: 100K users growing 10% day over day, 40% connecting a credit card in three weeks, $1,000+ average first-month spend

Instinct announced 100,000 users growing 10% day over day, 40% of users connecting a credit card in the first three weeks, and over $1,000 on average in first-month spend. These early metrics serve as benchmarks for consumer assistant traction.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Company-announced figures relayed by the speaker.

View full analysis →Listen to the original ↗

Principle · Money

Build AI products as full rich experiences, not thin wrappers on a model — and avoid dismissive terms like 'harness' and 'wrapper'

Builders are now delivering rich software products that are much more than just a shim on the model or a harness bringing context, even when agents help write some of the code. The speaker tries to avoid using words like 'harness' and 'wrapper' because successful AI products are rich experiences that happen to use a model as one piece. This is a framing principle for how to think about AI products.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Principle/framing from the speaker.

View full analysis →Listen to the original ↗

Report · Money

Muse launch metrics: ~500K downloads and 250K active users in 12 days, but far behind Threads' ~16M US/Canada downloads in 22 days

Muse had approximately 500,000 downloads and 250,000 active users in its first 12 days. In the first 22 days, Threads had about 16 million U.S./Canada downloads versus Muse's roughly 5 million over the same period. Muse had a very successful launch especially within the tech community, but still pales in comparison to Threads in appeal to the average person who doesn't work in tech. Threads had a critical-mass effect requiring rapid user communication to form a new network, whereas Muse is a personal tool that doesn't need that early network effect.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Figures flagged as possible numeric errors; downloads are US/Canada only.

View full analysis →Listen to the original ↗

Creator perspective · Money

Consumer AI products are becoming enterprise companies much faster than before via PLG — work-adjacent tools get pulled into enterprises

Consumer AI companies are becoming enterprise companies much faster than in the past (the speaker wrote a piece called 'The Great Expansion' about this). Work-adjacent AI tools grow via product-led growth and then get pulled into enterprises, requiring founders to add privacy, security, and team plans. PLG is a very effective, very low-cost distribution channel for these tools. By contrast, Canva took six-plus years to build real team and enterprise revenue, while AI consumer products like Crea, Replit, ElevenLabs, and Gamma reach enterprise revenue much faster.

The a16z Show · The Top 100 Consumer AI Apps: Who’s Actually Paying?

Claim from The a16z Show

Comparative claims from the speaker's analysis.

View full analysis →Listen to the original ↗