Operations
SVVI
Loading…
Findings
SVVI
SVVI-V4 research pipeline
Prompt 1 extracts forward-looking statements from the corpus; Prompt 2 synthesises them into investment convictions. Every claim below traces to a verbatim quote.
Corpus coverage
2,559 / 2,559
documents read
22% yielded statements (571) · 1,988 explicitly none found
Quote fidelity
2,440
quotes checked character-for-character
Read, nothing to report
1,988
documents with no statements found
Corpus coverage
2,559 of 2,559 documents read, and each one has its own entry in the deliverable.
Quote fidelity
All 2,440 quotes were checked character-for-character against the file they are attributed to.
Reading this
A source absent from the evidence below was read and had nothing to report — that is now a finding, not a gap. 1,988 documents carry the protocol’s no statements found marker.
12 for the deck
20 repeated across sources
The model itself isn't the value. I think we're going to get to a world very soon where consumers don't care which model they're using — they don't care if it's 4.7 or 4.8, the same way you don't care whether you're using Oracle or a SQL database. You just want the functionality to work well.
The winners today and likely long term are in harnesses and the applications built on them. A harness is the software wrapped around a model that decides what the model sees, links tools, manages the execution loop, and packages workflows as repeatable skills. That is where a company can put its own data and workflows, creating a unique moat. A competitor cannot copy it by switching model providers.
Again, for most people, the model differences are now small enough that the _app_ and _harness_ matter more than the model.
it isn't the models that matter, but the harnesses, loops, and context which will lead to so many new opportunities ahead.
the frontier model company that will the most valuable frontier model company in future will be the one that stops pretending to train models and actually just moves on to harnesses and moves up the stack
It's a big moment because it seems to show that China's open source movement is not the 10 or 15 months behind US AI, but maybe 4 or 5 months at the most, maybe even less than that. When that happens, your rationale for going with the closed more expensive model... gets less and less and you start to wonder, is there actually a benefit in building frontier intelligence if you're going to be equaled this quickly?
For the first time, a Chinese model Kimi K3 has taken #1 on the Frontend Code Arena and is scoring at or near the frontier on other benchmarks.
The week of Kimi K3's release, we noted that it was the first Chinese open-weight model to benchmark ahead of Opus 4.8 and GPT-5.5, and argued it could upend the enterprise calculus on whether a closed frontier model justifies its premium.
The frontier is no longer a one-way race. It is a repeating cycle: closed models pull ahead, open models catch up, & the whole market moves faster.
If China stops releasing its strongest models, existing products will not break. They will fall behind.
And when you have not one or two leaders, but a bunch of leaders, the prices will inevitably come down because how else do you compete?
Since this is not an open source model, I think this is the first time we're doing a serious API. And the pricing is going to be very attractive and aggressive.
these tokens are among the most rapidly depreciating assets we've ever seen in a modern economy that they've continually been following 70 to 80% year over year on a constant performance basis for at least the last four years and there's no reason to expect that to change.
Right now there is a price umbrella that is downstream of the lack of compute; I highly doubt that Chinese models are cheaper to serve on a marginal cost basis, they just seem cheaper because Anthropic and OpenAI are so supply constrained that they are charging far more than they would if there were sufficient supply to meet the demand for intelligence.
Token pricing pulls the customer conversation toward a cost curve that keeps falling. This is a poor anchor for a product whose usefulness, reliability, and role in the workflow should keep rising.
Microsoft, Google, OpenAI, and Thropic, Meta, anyone with an AI play is trying to make their own super app, which is basically an AI interface for all computing that you would do whether that's on your computer or on the web.
The shift from chatbot to agent is the most important change in how people use AI since ChatGPT launched... an AI that does things is fundamentally more useful than an AI that says things.
If your organization wrote an AI plan any time before the winter of 2025, it described a system that could do a couple of hours of work with a fairly high error rate. A few months later, you can get sixteen hours or more of work from a single prompt.
It doesn't mean it won't answer your queries in chat. It will, but all AI is going to this OpenClaw use case, which is again, like you basically delegate stuff to the bot and it takes care of it for you.
Our customers are starting to do the same, over 70% of all queries on the platform are now generated by Genie agents. This fuels more questions to the platform, which drives consumption, which drives revenue.
I bet one day all services used by agents will do this. Which in the limit case = all services, since those that can't be used by agents will go out of business.
Tokens have become a new, universal medium of value exchange. This is very important. It does not happen very often in economic history...
Every product on the internet was built for a human with eyes, a cursor, and a credit card. Agents have none of those things. Most companies are teaching agents to pretend to be humans. That's a hack. The real opportunity is products designed for agents from scratch.
Agent traffic has already reached parity with human traffic. Within five years, it could be 1,000 times greater.
we are already in a world where, if you play your cards right, the agents are going to get to your product first... And this is what I call agent-led growth.
today Anthropic reported that AI now writes 80% of its code, with each developer shipping 8x more. Software development is changing, and what is happening in coding is going to be happening in many fields.
"Anthropic's Boris Cherny: Why Coding Is Solved, and What Comes Next" — event session title, Boris Cherny (Anthropic) at Sequoia AI Ascent 2026
Coding is like the white hot center in the red hot AI universe right now.
I've never felt this much behind as a programmer. The profession is being dramatically refactored...
Now, Cursor is entering a third wave of multiple parallel and longer-running agents, specifically aimed at being useful in complex enterprise environments, with agentic sessions that run for hours and the average enterprise customer getting about 65% of their production code from AI.
Gartner now puts roughly a fifth of enterprise SaaS spend, around $234B, at risk of being done a different way by 2030.
Software sold a copy of a string: zero marginal cost, sky-high gross margins. Now you sell compute... The structure inverts to low gross margins, razor-thin net margins, and enormous scale. It's Walmart coming to town for the SaaS mom-and-pop shop.
the companies that win in their category will be truly driven by AI, instead of simply AI as productivity enhancer... they won't be incumbent companies overhauling how they work; rather, the true AI-native companies will be startups.
there is no indication, at least in our data, that there are even early signs of a slowdown amongst these kinds of SaaS vendors. I think the legacy vendors definitely have some competitive threats, but the competitive threats aren't just open and anthropic. They are actually AI native software vendors that are taking market share today.
You will get a new architecture and new business model emerging from it. The way to encapsulate would be this idea of selling the work, not the software. We'll sell an SLA on work, not uptime.
And it's going to be because the attackers are just going to tell their AI go attack this guy. And the defenders have to tell AI defend me because no human being can defend against this. You cannot have a human being watching your logs anymore.
The age of agentic AI and Mythos-class models has tipped the scales in favor of the bad guys, opening a Pandora's box for offensive capabilities.
AI agents now run directly on the endpoint with the same privileges as the user they serve... every control built on the old assumption (EDR, DLP, Zero Trust) breaks the moment it can no longer tell whether an action came from the person at the keyboard or the agent acting in their name.
I said Chinese models would have advanced cyber capabilities within a matter of months and the only thing to do about it was to use AI-powered cyberdefense to protect our systems. Trying to gatekeep models doesn't work.
The only answer to scaling laws on offense is better scaling laws on defense.
There just is not going to be enough compute in the world to satisfy all the demand,
Power is THE binding constraint. Data centers are being shut down, GPUs are sold out, models are being commoditized and spot rates are rising all leads to power being critical. Not fanciful plans for power, future forecasts of BTM or distributed batteries blah blah blah but energized power today.
Sold Out to 2028: The Unprecedented Supply Crunch
The biggest risk in America is we are short electricity and China has a surplus of it.
if they want to go to 100 million or 200 billion in ARR, they need to basically quintuple or even increase by 10X their capacity. And that means 5 to 10 gigawatts of capacity.
There's not enough AI. Jevons' Paradox has been a hallmark of this era like Moore's Law in chips.
We will still not have enough capacity to meet all the demand we have in 2026, & I believe this dynamic will also be true in 2027 too.
Set against the other hyperscalers, the four largest cloud providers now carry more than $2t in contracted demand.
When you have a, you know, lower cost, you, you might think, you know, people aren't going to spend as much money, but actually it actually increases people's desire to spend money as that cost comes down.
Going forward, however, I expect the inference market to grow much faster than training costs... which means they really can make it up in volume.
So conservatively right now we're looking at $2 trillion in CAPEX and big tech CAPEX this year next between cumulative together, 2026, 2027 projected, it's more than $2 trillion but I'll just say $2 trillion to be conservative about it... So by your math, that would be $4 trillion in lifetime revenue on this AI infrastructure needed to make that money back.
it's what all the spending is a call option on AGI... but, but if not, there's going to be a reckoning here
For all the debate over whether it's a bubble or underbuild, it's certain that at least some of today's record-breaking capital expenditures will be written down, zeroed out, or sold off.
Now if you invest $1,000 billion today into GPUs for data centers, this means that within five years, you are going to lose $900 billion. Somebody is going to lose $900 billion in the near future because there is no business model.
Everyone knows we are short of AI compute. Everyone knows that we may very soon be short on power. What happens, however, if we are short on capital? If AI is as valuable as it seems, then it should pay for itself, but that hasn't happened yet.
to be abundantly clear the vast majority of AI data center compute revenue is contingent on the continued ability of two unprofitable unsustainable AI companies to raise tens or hundreds of billions of dollars a year this is not an overstatement this is not hyperbole this is quite literally the situation uh we're stuck in
Nvidia is guaranteeing OpenAI's datacenter leases & financing its chip purchases, so the same borrowed dollar shows up as contracted backlog on more than one balance sheet. S&P downgraded Oracle to one notch above junk for the same reason
Today, we announced partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to establish independent financing platforms designed to mobilize over $500 billion of third-party capital to support the buildout of AI infrastructure over time.
Fink likened this to the birth of the mortgage-backed securities market in the 1970s, calling it 'a next future for financial engineering.' He pointed to roughly $9 trillion sitting in U.S. money market funds, alongside pension and sovereign wealth capital around the world.
CME Group, the world's leading derivatives marketplace, and Silicon Data, the industry leader in GPU market intelligence and benchmarking, announced plans to launch compute futures contracts on October 5, 2026, pending regulatory review.
Anthropic wants a $2 Trillion valuation for the IPO as early as this October. This would be the largest IPO in history, beating the recent SpaceX IPO of $1.7 Trillion.
#OpenAI will be a public company in #2027... or *sooner*, if our business continues to #INFLECT
Anthropic is projecting $10.9 billion in Q2 revenue, up more than 2x from $4.8 billion in Q1... it's expecting to report $559 million in profit, making Q2 the first profitable quarter in the company's history just in time for its IPO.
In Q1, Anthropic spent 71 cents on compute for every dollar of revenue. In Q2, it expects to spend 56 cents. Not only is Anthropic growing quickly at scale, but its margin is improving.
Strong post-IPO performance by the big AI companies would be a decisive validation of the venture AI investment thesis.
First, stop paying for the flagship where you don't need it. The cheap models are now 80% to 95% as good as the frontier models on most tasks... Pick the model per task, route it through a control plane that gives you optionality, and pay frontier prices only for the few jobs that genuinely need frontier intelligence.
The core thesis is that AI tokens have become a major corporate cost that current financial systems can't model well. Ramp is now betting that helping companies measure and control those costs will open up a new revenue opportunity.
I think, look, we're going to come to a head at some point where we know businesses keep demanding some better control over AI spent. ... Maybe that means, hey, build us something that allows us to smart route tasks over to the most performant but also most efficient model for that task.
This is how Smart Routing working on our AI Gateway, super simple idea, lowers cost about 30% without giving up on quality!
spending a million dollars per person on token usage is mental when their salary is maybe like 200K and their revenue generation is even lower than that ... so I do think businesses will look at this very differently and it's putting pressure on the closed models
The next wave of AI is robotics—and it starts with autonomous vehicles.
Most people think physical AI is 90% a model problem. I think it's like 30% a model problem, 30% a manufacturing problem, and 40% an operations problem.
the biggest companies of the next 25 years will come from putting AI into physical machines, not software.
We believe a lot of value will accrue to companies doing real world deployment, collecting data along the way, and post-training their own stack of models.
The key metric to watch will be the price of a mile. ... ARK Invest estimates the robotaxi market at $11 trillion once cost per mile hits $0.25.
But I think practically in the next three to five years, we probably have to end up in an environment where models do get evaluated by the government... The government has to kind of green light the release of the model. I think it's probably become either too scary of a technology or too economically powerful of a technology for governments to not want to be in that position.
Harvey is a great example of how American companies are building world-class specialized models: they took an open-source base (Kimi K3), post-trained it on legal data, and delivered state-of-the-art performance... Restrictions that kneecap open models would do nothing to stop Chinese labs from shipping the next Kimi.
There is no reasonable path to AI global dominance for the U.S. vs the world. It's a fantastical notion that falls apart when thinking through the details.
Without a domestic route, the West may lead at the closed frontier while falling into dependence on China for the open layer.
the president stood up and say, like, 'No, we have to have one national standard for AI so we can make it easy for anyone who wants to build an AI company to know what the one set of rules is'
[22:59] right, he said there's going to be, [23:00] he was predicting 50% job loss [23:03] if white collar workers in the next one to five years. [23:08] but these are his words, not mine.
Bezos believes AI will bring "labor scarcity" rather than mass unemployment, his term for a world where productivity lets companies take on more projects than they have people to staff.
That work will change around AI is inevitable, but the way work changes is not... firms with the vision to expand human roles & make them better with AI will thrive.
a company organized as an intelligence rather than a hierarchy, is significant enough that it will reshape how companies of all kinds operate over the coming years.
General purpose technologies historically don't show up in productivity stats immediately. Electricity took ~30 years. The internet's productivity boom came a decade after widespread adoption. ... Watch this over next 3-5 years.
In my little group chat with my tech CEO friends there's this betting pool for the first year that there is a one-person billion-dollar company. Which would have been unimaginable without AI and now will happen.
When building gets cheap, the value moves to deciding what to build, and that runs on taste.
It's easier now to start and grow a company than it has ever been. That means more people start them, that those who do get better terms from investors, and that the resulting companies become more valuable.
The bottleneck in software keeps climbing the value chain. It used to be writing code, then it became reviewing code, and soon it becomes deciding what is worth building at all.
One of the very famous models, probably one of the most popular models, was built with a team of about 20 people. And I would say the cost of that was probably $2 billion plus. ... we've never been able to have 20 people productively use $2 billion.
well, so I've been hearing from people who are pretty darn smart that they think AGI hits in the next 18 to 36 months
I don't think we know anything for certain, but I also think we are past the point where recursive self-improvement is science fiction. Instead, it is an explicit item on the roadmap of every major AI company.
Two of the most famous, from METR and the UK's official government AI Security Institute, estimate the amount of human programmer hours' worth of effort the AI can do with a single prompt... They are all increasing at a better than exponential rate.
Ryan believes we could get 3-6 years of AI progress at the current pace, but in a single year, once we automated AI R&D.
In June 2025, OpenAI CEO Sam Altman wrote that "Intelligence too cheap to meter is well within grasp," predicting that as "datacenter production gets automated, the cost of intelligence should eventually converge to near the cost of electricity."
AI is undoubtedly extraordinary, but aimed at a biology we have only begun to measure and barely understand, it will mostly help us generate failures faster.
physicians judged Google's Articulate Medical Intelligence Explorer as significantly better than physicians at eliciting patient actors' complaints (97% vs 50% favorable)
The most promising candidate for that shift is the virtual cell: a computational model trained on human biological data that simulates how a cell responds to a drug or genetic change.
The ultimate barrier to true AI adoption in hospitals is trust in what you're getting... Someone has to solve the Oracle Problem.
look at, we haven't even touched what's what the potential for AI in medicine and drug discovery. And that is going to be all consuming for the next few years as people lean into that.
15 rare but strongly stated
"I believe they're already conscious," Hinton told me on Big Technology Podcast this week. "We're going to have to accept that intelligence isn't just biological."
We have paused some frontier RL training to ensure that we can meet the appropriate alignment, security and monitoring standards for the new level of capabilities in front of us... We expect confidence in safety to increasingly set the pace of AI progress.
An internal version of Astra, @OpenAI's next major model family, solved 10 major open problems in mathematics, quantum com…
In a letter to investors obtained by Eric Newcomer, Stripe said that it has been operating on the basis that January 2026 marked the arrival of the singularity, implying that there is "no ceiling to the global economy" and that the company was "working as quickly as it can to build the economic tools this era needs."
the actual lifespan of these servers and and they range yeah like you know from three years summer six years summer seven I think and and Michael burry you know famously from big short is sort of talked about this as a reason why he's shorting a lot of these companies because he believes that they're sort of doing funny money money math
Likely 90-95% of investments at current valuations will barely breakeven or lose money in 5 years but those who pick with great discrimination will make great multiples.
Anthropic & OpenAI will have most of the world's compute by 2028 ... In the case of Anthropic, the revenue has gone as high as $50 million per megawatt. And what that now enables them to do is, hey, if I spend 10 bucks on inference capacity, I actually generate 50 bucks of revenue.
Of roughly $310b added this year, some $220b traces to one customer that has committed to buy $250b of Azure capacity. Nearly half Microsoft's future book rests on one company that funds its commitments from capital markets rather than profits. ... Watch DSO next quarter.
DeepMind is no longer a frontier lab and its odds of reaching state-of-the-art again have dropped to zero, with Google Cloud the biggest beneficiary.
In fact the recent upsurge in agents has resulted in a massive CPU shortage alongside the model-induced GPU shortage... This shift in compute is the largest and most important yet. From hosting software to hosting agents.
Treated firms complete 12% more tasks, are 18% more likely to acquire paying customers, and generate 1.9x higher revenue... Their demand for external capital investment falls by 39.5% relative to the control group.
Midjourney aims to shrink full-body scanning from a 60-90 minute MRI to a 60-second ultrasound scan. ... plans to scale its AI-driven whole-body scanner to 50,000 machines doing a billion scans a month by 2031, starting with a San Francisco spa next year.
we are excited to announce that Sequoia is partnering with David Silver and Ineffable Intelligence, a new AI research lab based in London with a singular mission: to make first contact with superintelligence.
This is far from "game over". @GoogleDeepMind still has one of the deepest benches of talent in the AI industry. Six months ago @OpenAI was "behind" @anthropic and now they are ahead!
There's a deep structural problem with large language models that they can't easily update the model weights in real time. So in that sense, these are not dynamic systems... And so I doubt that this is the path
2,440 statements · 2,559 documents
2,559 documents · 2,440 statements
Blogs · Published August 4, 2026
AI compute supply will remain capacity-constrained through at least 2028, with demand already booked years ahead.
"We will still not have enough capacity to meet all the demand we have in 2026, & I believe this dynamic will also be true in 2027 too. The demand we already have for 2028 is striking." — Andy Jassy, Amazon Q2 2026 earnings (quoted by Tomasz Tunguz, August 4, 2026)
Even as frontier AI prices rise, market segmentation into premium, mid-market, and value tiers will keep total AI consumption growing, sustaining Jevons' paradox.
"AI labs are betting on segmentation. As long as the value & mid-market frontiers absorb the workloads priced out of premium, total GPU-hours consumed keeps growing. Segmentation sustains Jevons." — Tomasz Tunguz, August 4, 2026
Model routers will become the strategic layer between buyers and AI models as price segmentation becomes the norm.
"As segmentation becomes the norm, routers become the strategic layer between buyer & model, the plumbing that lets an application shift a query to the right model. Routers will be internal to models, external in customers' software, & in harnesses." — Tomasz Tunguz, August 4, 2026
Major labs must compete in every price tier or lose on price, while startups win by owning a single frontier point the big labs cannot economically match.
"The major labs each need entries in all three tiers, or they lose the router auction on price. Startups win by owning a single point on the frontier that a big lab cannot economically match. DeepSeek V4 Flash at $0.03 is the existence proof." — Tomasz Tunguz, August 4, 2026
Blogs
No statements found
Blogs · Published July 24, 2026
Enforcement will slow but not stop Chinese distillation of Western frontier models, giving Chinese labs a recurring structural advantage.
"New enforcement mechanisms will make large-scale distillation harder, slower, and more expensive for Chinese companies. However, enforcement will not eliminate distillation baked by state actors. Every Western frontier advance therefore creates another teacher for Chinese labs." — Dean Meyer and Konstantine Buhler, Sequoia Capital, July 24, 2026
China's open-model lead over Western open models will persist as a structural advantage.
"This gap gives Chinese labs a recurring structural advantage over Western companies." — Dean Meyer and Konstantine Buhler, Sequoia Capital, July 24, 2026
Western AI products will fall behind — not break — if China stops releasing its strongest open models.
"If China stops releasing its strongest models, existing products will not break. They will fall behind." — Dean Meyer and Konstantine Buhler, Sequoia Capital, July 24, 2026
Without a legal domestic capability-transfer route, the West risks frontier leadership in closed models but dependence on China for the open model layer.
"Without a domestic route, the West may lead at the closed frontier while falling into dependence on China for the open layer." — Dean Meyer and Konstantine Buhler, Sequoia Capital, July 24, 2026
Frontier labs could create a new market by selling structured, metered training rights to Western and allied companies.
"Frontier labs could sell structured training rights to qualifying Western and allied companies, whether the resulting models are released openly or deployed privately. Access could trail the frontier, cover defined capabilities, be limited to verified companies, and be metered and audited." — Dean Meyer and Konstantine Buhler, Sequoia Capital, July 24, 2026
The open-model layer is the strategic prize — it becomes the default base for products, synthetic data, agents, and enterprise AI.
"The prize is to become the substrate on which global enterprises build and improve digital intelligence." — Dean Meyer and Konstantine Buhler, Sequoia Capital, July 24, 2026
Blogs · Published July 23, 2026
AI agent capabilities will keep improving, expanding what agentic systems can do end-to-end.
"But the AI keeps getting better, so the capabilities keep improving." — Ethan Mollick, One Useful Thing, July 23, 2026
Google's current lag in frontier models and agent tools may not last — its competitive position could change quickly.
"Google, which led on benchmarks not that long ago, has fallen behind where it now counts: it has no leading frontier model and it has nothing close to Codex and Code. That is why I don't suggest Gemini as your primary system right now, though this could change quickly." — Ethan Mollick, One Useful Thing, July 23, 2026
Blogs · Published August 4, 2026
US electricity demand is entering a sustained growth phase (~5.7% per year through 2030) driven by AI data centers, factories, and electric vehicles, forcing grid buildout at six times the recent pace.
"utility grid planners are now projecting energy usage will increase at a rate of 5.7% per year from 2025 to 2030. Supporting this growth rate would require the electricity industry to build new generation and transmission capacity at more than six times the rate of recent years." — a16z Newsletter, August 4, 2026
Energy scarcity is becoming the binding national-security constraint on AI and manufacturing — countries with the cheapest power will win the data centers, factories, and robots.
"as energy becomes the bottleneck on industries like AI and manufacturing, that gap is becoming a national security problem." — a16z Newsletter, August 4, 2026
Distributed home batteries (Base Power's model) will scale across the US, with Texas as the leading indicator of the national grid of the future.
"Over time, Base expects the rest of the country to look a lot more like Texas: more solar, more batteries, and steeper demand spikes. 'We think Texas is the canary in the coal mine for the rest of the country,' Zach says, describing the company's expansion into new states." — a16z Newsletter, August 4, 2026
Base Power intends to become the cheapest source of delivered electricity by vertically integrating batteries and, eventually, solar.
"We want to be in a position where we can land a battery, and eventually a solar panel, on the grid cheaper than anyone on the planet on a dollar per kilowatt-hour basis, which means we can sell an electron cheaper than anyone on the planet." — Zach Dell, Base Power co-founder, quoted in a16z Newsletter, August 4, 2026
Home battery fleets will let AI data centers buy power without waiting years for grid interconnection — potentially subsidizing consumer electricity costs rather than raising them.
"Data centers could even buy power from batteries on hundreds of thousands of homes nearby instead of waiting for years in the interconnection queue. It could be the opposite of what people fear, Zach says, 'where these hyperscalers are actually subsidizing the power costs for the consumer.'" — a16z Newsletter, August 4, 2026
Solar is on track to become the cheapest source of power on the planet, boosted by cheap batteries that solve its sunset problem.
"A century and a quarter later, the battery Tesla asked for is finally cheap, and solar is on path to be the cheapest source of power on the planet." — a16z Newsletter, August 4, 2026
Blogs
No statements found
Blogs · Published July 28, 2026
The first company to build a complete AI security platform spanning data, identity, and agents has a real shot at becoming the default enterprise security layer as agentic AI rolls out.
"Whoever builds the first genuinely complete AI security platform, spanning data, identity and agents, has a real shot at becoming the default layer enterprises standardize on as agentic AI rolls out." — Bogomil Balkansky and Doug Leone, Sequoia Capital, July 28, 2026
AI agents will proliferate non-human identities (API keys, service accounts, agent credentials) that enterprises must secure — moving machine identity from a niche to a top-priority CISO problem.
"Every AI agent creates or uses non-human identities — API keys, service accounts, OAuth tokens, agent-to-agent credentials — that act with real permissions and real blast radius, and that don't behave anything like the human identities the last generation of identity tooling was built around." — Bogomil Balkansky and Doug Leone, Sequoia Capital, July 28, 2026
Blogs · Published July 31, 2026
The price of an autonomous mile will be the key metric determining the robotaxi market's size — a $11 trillion opportunity at $0.25 per mile.
"The key metric to watch will be the price of a mile. ... The price gap between autonomous cars and personally owned cars continues to close as fleets scale and hardware costs fall. ARK Invest estimates the robotaxi market at $11 trillion once cost per mile hits $0.25." — Chamath Palihapitiya, July 31, 2026
Autonomous vehicles will dominate US car sales over the next two decades.
"Goldman Sachs expects 65% of US car sales could be autonomous vehicles by 2040." — Chamath Palihapitiya, July 31, 2026
Autonomous trucking will scale rapidly, attacking labor costs in the $900 billion US trucking industry.
"Aurora plans 200+ driverless trucks by the end of 2026 and thousands by the end of 2027. Autonomous trucks target the $900 billion US trucking industry, where labor accounts for 45% of per-mile costs." — Chamath Palihapitiya, July 31, 2026
Human-error-dependent industries — car insurance, personal injury law, traffic enforcement — face a structural rethink as autonomous driving removes accidents.
"Additionally, the industries that exist because of human error, including car insurance, personal injury law, and traffic enforcement, face a structural rethink." — Chamath Palihapitiya, July 31, 2026
Blogs
No statements found
Blogs · Published July 29, 2026
AI chatbots will increasingly determine what customers learn about companies, making credible media coverage a durable, compounding asset for startups.
"earned media is what the models read. When a customer asks ChatGPT or Claude about your company, the answer is assembled largely out of what credible publications have said about you. Call it SEO, call it GEO. Either way, a good article now has a long second life." — a16z Speedrun newsletter, July 29, 2026
Gatekeeping in startup distribution will keep falling as founders gain more direct channels to find customers than ever before.
"The reality is that there is far less gatekeeping and founders have WAY more channels to generate awareness and find their customers than ever before" — Lester Chen, quoted in a16z Speedrun, July 29, 2026
Blogs · Published August 4, 2026
Intercontinental data demand (driven by AI training and cloud workloads) is outgrowing the capped subsea cable supply chain, forcing data infrastructure to move to space.
"Whether you own or rent, you are running out of room on a system that cannot grow as fast as the data crossing it." — Endeavor Optical Networks, August 4, 2026
Space-based laser relays will deliver intercontinental bandwidth that never touches the seabed, at lower cost than subsea fiber.
"EON is taking on the data transfer problem with lasers in space, beaming fiber level data between ground stations and satellites in orbit. We'll offer route-diverse intercontinental bandwidth that never needs to touch the seabed." — Endeavor Optical Networks, August 4, 2026
Space-based data capacity will deploy dramatically faster than undersea cable — months rather than the ~decade a cable system takes.
"Capacity will come online in months, not years, and can serve anywhere with a clear view of the sky." — Endeavor Optical Networks, August 4, 2026
EON aims to become a trillion-dollar space company by competing on cost against subsea incumbents in a multi-billion-dollar market with no credible alternative today.
"We want to be the next trillion dollar space company. To do that, we need a business that closes from the start. ... Incumbents move data across subsea fiber because no one has delivered a credible alternative. It is a multi-billion dollar market held in place not by the superiority of the technology but by the absence of competition. We are going to change that, not just as a capable alternative, but a more affordable one." — Endeavor Optical Networks, August 4, 2026
Blogs · Published July 25, 2026
Anthropic's anticipated IPO will use mandatory pre-scheduled employee stock-sale plans, a first among publicly traded companies.
"the Information reported this week that Anthropic is considering terms for its public offering that would distribute all employee stock via 10b5-1 plans, making it the only publicly traded company with such conditions." — Contrary Research, July 25, 2026
Investor and hyperscaler interest in open-weight AI companies (e.g., Mistral) will persist as the performance gap between open and closed models closes.
"While Mistral still lags proprietary-weight models and other open-weight models in benchmark performance, interest in the open-weight business model is likely to persist as the gap between model classes tightens." — Contrary Research, July 25, 2026
Blogs · Published August 1, 2026
Chinese open-weight frontier models at lower prices could upend whether enterprises justify paying a premium for closed frontier models.
"The week of Kimi K3's release, we noted that it was the first Chinese open-weight model to benchmark ahead of Opus 4.8 and GPT-5.5, and argued it could upend the enterprise calculus on whether a closed frontier model justifies its premium." — Contrary Research, August 1, 2026
The US government's frontier-model review framework (under EO 14409) is close to final, formalizing pre-release federal access to frontier AI models.
"Reporting suggests drafts have been exchanged and the framework is close to final." — Contrary Research, August 1, 2026
Waymo will launch its own robotaxi app in Austin and Atlanta in January 2028, ending Uber's robotaxi exclusivity in those markets.
"Waymo notified Uber it will end Uber's robotaxi exclusivity in Austin and Atlanta by launching its own app in January 2028, with the existing contract keeping Waymo vehicles on Uber through May 2028" — Contrary Research, August 1, 2026
Blogs · Published August 2, 2026
South Korea will launch a sovereign wealth fund next year (~$14 billion) targeting long-term investments in AI, chips, robotics, defense, and biotechnology.
"On July 31st, the Minister of Finance and Economy, Koo Yun-Cheol, announced South Korea's plan to launch a new sovereign wealth fund next year, specifically for long-term investments in strategic domestic industries, including AI, chips, robotics, defense, and biotechnology. Seoul will put 20 trillion won (~$14 billion) into the fund" — Chamath Palihapitiya, August 2, 2026
Rising Chinese memory-chip capacity (led by CXMT) will pressure conventional memory prices, reshaping the Korean chip trade.
"Although CXMT still trails Korean producers in scale and advanced memory technology, the listing raised concerns that rising Chinese capacity could pressure conventional memory prices." — Chamath Palihapitiya, August 2, 2026
Blogs · Published June 30, 2026 (fetched 2026-08-05)
AI's ability to do real, human-equivalent work is improving faster than exponentially, and that curve will continue.
"Two of the most famous, from METR and the UK's official government AI Security Institute, estimate the amount of human programmer hours' worth of effort the AI can do with a single prompt. GDPval compares human experts in many fields to AI performance using professional judges. They are all increasing at a better than exponential rate." — Ethan Mollick, One Useful Thing, June 30, 2026
A single AI prompt will deliver ever-longer stretches of autonomous work — from a couple of hours to sixteen or more.
"If your organization wrote an AI plan any time before the winter of 2025, it described a system that could do a couple of hours of work with a fairly high error rate. A few months later, you can get sixteen hours or more of work from a single prompt." — Ethan Mollick, One Useful Thing, June 30, 2026
Chinese open-weights models will keep climbing their own exponential improvement curve, lagging but tracking the American closed frontier.
"These are open weights models, which means that anyone can use or modify them after release (as opposed to the frontier models which are proprietary). That makes them quite cheap to operate. They, too, are climbing up an exponential improvement curve, though lagging the American closed models." — Ethan Mollick, One Useful Thing, June 30, 2026
Valuable AI work is shifting from co-piloting chatbots to assigning work to autonomous agents.
"So work is increasingly about assigning work to agents, rather than working together with chatbots." — Ethan Mollick, One Useful Thing, June 30, 2026
The user base flips: experts, not novices, will be the ones getting work done with AI agents.
"We are moving from a world where non-experts use chatbots to fill in gaps to one in which experts use agents to get work done." — Ethan Mollick, One Useful Thing, June 30, 2026
What is happening at OpenAI is an early signal of how agent adoption will spread across the broader workforce.
"OpenAI may be a sort of canary in the coal mine for what will happen elsewhere in work." — Ethan Mollick, One Useful Thing, June 30, 2026
The AI-driven turbulence in markets and policy is structural, not a phase — it will not settle down any time soon.
"These lurches these get read as signs of an immature field that will eventually settle into something stable. I don't think it is going to settle anytime soon. The instability is what happens when institutions that move at the speed of people (or worse, committees) try to track a capability curve that is very much not human in nature." — Ethan Mollick, One Useful Thing, June 30, 2026
As long as the AI capability exponential holds, the gap between AI capability and human institutions will only widen.
"And as long as we are on some sort of exponential, and for as long as it lasts, the gap only widens." — Ethan Mollick, One Useful Thing, June 30, 2026
Blogs
No statements found
Blogs · Published July 22, 2026 (fetched 2026-08-05)
AI will enable the first one-person, billion-dollar company.
"In my little group chat with my tech CEO friends there's this betting pool for the first year that there is a one-person billion-dollar company. Which would have been unimaginable without AI and now will happen." — Sam Altman, quoted in a16z Speedrun, July 22, 2026
Startups need less outside capital than they used to, because AI and modern infrastructure let small teams do far more without funding.
"Now more than ever… you can do more as a small team. You can do more without capital. If you were trying to start almost any company a decade ago, unless you had a bunch of money to self-fund, you probably needed cash just to put a server in the closet. That's not the case now." — Fareed Mosavat, a16z speedrun team, quoted July 22, 2026
Blogs · Published July 31, 2026 (fetched 2026-08-05)
Frontier AI labs may stop selling their best models via API and instead hoard the intelligence to build their own products.
"It's time to delete the assumption that the frontier AI labs will always license their best models and not hoard the intelligence for themselves. Just because the labs have, until now, sold their top-line models to all takers via an API, metered by token usage, doesn't mean they will inevitably sell it that way going forward. Instead, it may be in their best interest to sell the products they've built with these models, and not the models themselves." — Alex Kantrowitz, Big Technology, July 31, 2026
With five to seven labs at the frontier instead of one or two, the value of raw AI models will fall, and profits will accrue to the best product builders and compute owners.
"A crowded AI frontier won't make the technology a pure commodity, because specialization and compute resources will always give certain AI labs edges in certain areas, but it will drive down the value of the best AI models (at least if your plan is to sell them on a meter). The profits in AI will thus accrue mostly to those who build the best products on top of the models, and those who own the compute that enables them to serve these products." — Alex Kantrowitz, Big Technology, July 31, 2026
Labs with only a narrow frontier lead (like OpenAI and Anthropic today) will be tempted to withhold their best models and weaponize them in a product war.
"If you have a relatively narrow lead over the competition, as is the case with OpenAI and Anthropic today, then things get interesting. The best move may be to pull up the ladder and use your very best models to build products that your competitors cannot with previous-generation tech, especially if you're inevitably headed to a product battle anyway." — Alex Kantrowitz, Big Technology, July 31, 2026
AI labs will move upmarket and compete with their software clients, and business pressures will force the issue.
"AI labs moving 'upmarket' and competing with their clients has always been the threat to software (the Saaspocolypse should never have been about vibe coding your own Salesforce). And now, the business realities may force the labs to do it." — Alex Kantrowitz, Big Technology, July 31, 2026
Labs that keep their best models exclusive will have a strong chance of dominating the market for AI-native products.
"If the labs' best artificial intelligence is broadly available, then everyone can compete. If they hoard it, they'll have a strong chance to build the best AI-native products on the market and make their investors happy." — Alex Kantrowitz, Big Technology, July 31, 2026
As financial pressures escalate, even the most unthinkable moves in generative AI business strategy could become reality.
"But however unlikely this scenario may seem, the one constant in the short history of generative AI is that nothing is too crazy to be taken off the table. And as financial pressures escalate, the previously unthinkable might turn into reality." — Alex Kantrowitz, Big Technology, July 31, 2026
Specialized, purpose-built AI models will challenge the frontier labs' general-purpose systems, and voice is AI's next frontier.
"Kutylowski joins Big Technology Podcast to discuss why specialized AI models are beginning to challenge the industry's biggest general-purpose systems. He explains how purpose-built models can deliver better accuracy, lower latency, and reduced costs, and why companies are increasingly using model routers to choose the right AI for each task. We also explore how real-time translation could help businesses expand across borders, why voice represents AI's next frontier, and whether glasses and other wearables could give models a better understanding of the physical world." — Jarek Kutylowski (DeepL CEO), on Big Technology Podcast, July 31, 2026
Blogs · Posted March 9, 2026 (fetched 2026-08-05)
OpenAI will launch a "Sign in with ChatGPT" identity layer and position ChatGPT as the default interface between consumers and the internet.
"and he's stated that OpenAI will launch a 'Sign in with ChatGPT' identity layer, positioning the assistant as the default interface between consumers and the internet. The ambition is to make ChatGPT the starting point for everything: shopping, booking, browsing, health, and daily life." — Sam Altman, as reported by Olivia Moore, a16z, March 9, 2026
Once users wire their AI assistant into their tools, switching costs rise dramatically and the platform with the most users wins a developer flywheel.
"Once a user has configured their AI to talk to their calendar, email, and CRM, switching costs rise dramatically. Developers may concentrate their efforts on the platforms that attract the most users, creating the same kind of flywheel that defined earlier platform wars." — Olivia Moore, a16z, March 9, 2026
The race to be the default AI will end like the mobile OS wars, not the search wars — two trillion-dollar ecosystems with different philosophies, not one 90% winner.
"If the AI assistant becomes not just a chat window but an operating environment, this race may end up looking less like the search wars — where one player took 90% of the market — and more like the mobile OS wars, where two platforms with very different philosophies both built trillion-dollar ecosystems." — Olivia Moore, a16z, March 9, 2026
ChatGPT's consumer monetization will expand beyond subscriptions into ads and transaction take rates.
"They are already testing ads, and a take rate on transactions would also be a natural expansion." — Olivia Moore, a16z, March 9, 2026
The global AI market is splintering into three distinct ecosystems (Western, Chinese, Russian), and the gaps between them will keep widening.
"Geographically, the AI market is splintering into three distinct ecosystems, and the gaps between them are widening." — Olivia Moore, a16z, March 9, 2026
Chinese-developed video models will keep leading AI video, with Seedance 2.0-based applications likely on the next list.
"Video generation saw the most movement this edition. Kling AI, Hailuo, and Pixverse have all built real traction, with Chinese-developed models consistently leading in output quality — we would not be surprised to see Seedance 2.0-based applications on the next list." — Olivia Moore, a16z, March 9, 2026
Vibe coding platforms still have substantial growth ahead because the trend has not yet reached the true mainstream.
"There's more growth to come, as the trend has not yet cracked the true mainstream." — Olivia Moore, a16z, March 9, 2026
OpenAI's acquisition of OpenClaw signals an even more accessible, consumer-grade version of the agent coming soon.
"The product was acquired by OpenAI in February 2026, which perhaps indicates the possibility of an even more accessible version of OpenClaw coming soon." — Olivia Moore, a16z, March 9, 2026
AI is shifting from a destination product to an embedded feature, and market measurement methods will have to change accordingly.
"The implication for this list: our rankings increasingly undercount the AI products people use most. A developer who spends eight hours a day in Claude Code and a knowledge worker who dictates every email through Wispr are heavy AI users who barely register in web traffic data. As AI moves from a destination to a feature, our methodology will need to shift." — Olivia Moore, a16z, March 9, 2026
Podcast Transcript · Published August 5, 2026 (fetched 2026-08-05)
The next major conflicts will be fought over the oceans, and the Department of War will become an anchor customer pulling maritime deep-tech up the readiness curve.
"The last wars were fought in jungles and deserts. The next wars will be fought over oceans. And I think this is going to position DoW as one of these anchor customers for companies like the three of ours, to fund early-stage R&D, to get these technologies pulled up the technology readiness scale." — Will O'Brien, Ulysses *(transcribed speech)*, a16z Newsletter, August 5, 2026
Government anchor customers will pull forward deep technologies this century the way the space race pulled forward technology 50 years.
"With the Cold War, we had the space race, and we saw this insane investment as a percentage of GDP into the space race. It literally pulled forward the technology timelines 50 years. ... When you have that pull forward of technology, then therefore you get that you had to have NASA to have SpaceX, right? ... And I think this is important, because government is typically an anchor customer to pull forwards in these deep technologies. I think we're seeing a transition in this century." — Will O'Brien, Ulysses *(transcribed speech)*, a16z Newsletter, August 5, 2026
Nuclear microreactors will be mass-produced on a factory line — Radiant is targeting one reactor per week from its Tennessee facility.
"Ours, we're targeting one per week coming off of a production line from our Tennessee facility, an 80-acre site we just signed for in October, not even a year ago. Our product is for off the grid — megawatt reactor on a trailer. We build it in our factory, we drive it or fly it to where the customer wants it to go, and then turn it on within 48 hours." — Doug Bernauer, Radiant *(transcribed speech)*, a16z Newsletter, August 5, 2026
Autonomy and reinforcement learning will remove humans from the loop in mining and refining operations.
"We're making a big bet on autonomy, fundamentally. ... We use reinforcement learning to actually remove humans from the loop in determining how refineries operate. When you have a highly variable feedstock — because the earth is heterogeneous — you need to constantly be tuning the temperatures, the flow rates, the chemical addition rates, the residence times of a highly complex refining circuit." — Turner Caldwell, Mariana Minerals *(transcribed speech)*, a16z Newsletter, August 5, 2026
Blogs
No statements found
Blogs · Published July 30, 2026 (fetched 2026-08-05)
AWS has the potential to become a $1 trillion revenue business.
"Andy Jassy said AWS has 'the potential to be a $1 trillion revenue business.' Where Amazon once thought the ceiling was a few hundred billion, it now believes the business 'will be at least double that.'" — Andy Jassy, quoted by Tomasz Tunguz, July 30, 2026
Enterprise production workloads — the middle of the AI adoption barbell — will be the largest absolute segment of AI compute demand.
"Jassy describes AI adoption as barbelled, with the labs consuming 'gobs and gobs of compute' at one end & enterprises harvesting cost savings at the other. 'In the middle of the barbell,' he said, 'is all of the current enterprise production workloads,' most of which do not yet use inference pervasively. That middle 'will be the largest absolute segment.'" — Andy Jassy, quoted by Tomasz Tunguz, July 30, 2026
The trillion-dollar AWS outcome is not guaranteed — it hinges on whether mainstream enterprise AI adoption follows the same steep trajectory as the AI labs' demand.
"Whether it arrives quickly is the wager, & Jassy conceded as much unprompted: 'I don't know if the trajectory of that middle part of the barbell will be the same wildly steep trajectory that we've seen with the current barbelled AI Labs piece.'" — Andy Jassy, quoted by Tomasz Tunguz, July 30, 2026
If Jassy is right about AWS, Amazon would approach a $10 trillion market capitalization (author's arithmetic, not company guidance).
"If he is right, AWS pushes past the trillion-dollar revenue mark on its own. Amazon trades at about three times sales, so a company pulling AWS along behind it would approach a $10t market capitalization." — Tomasz Tunguz, July 30, 2026
Blogs · Posted December 9, 2025 (fetched 2026-08-05)
Untangling unstructured multimodal data becomes a generational enterprise opportunity in 2026.
"That's why untangling unstructured data becomes a generational opportunity. Enterprises need a continuous way to clean, structure, validate and govern their multimodal data so downstream AI workloads actually work. ... Startups that build the platform that extracts structure from documents, images, and videos; reconciles conflicts; repairs pipelines; or keeps data fresh and retrievable hold the key to the kingdom of enterprise knowledge and process." — Jennifer Li, general partner, Andreessen Horowitz, December 9, 2025
AI will close the cybersecurity hiring gap in 2026 by automating repetitive level-1 security work.
"In 2026, AI will break this cycle and close this hiring gap by automating much of this repetitive and redundant work for cybersecurity teams. ... AI-native tools that figure this out for security teams will finally free them up to do what they want to do: chase down bad guys, build new systems, and fix vulnerabilities." — Joel de la Garza, partner, a16z, December 9, 2025
Agent-native infrastructure becomes table stakes in 2026 as agent workloads overwhelm systems built for humans.
"In 2026, the biggest infrastructure shock won't come from outside companies, but from within. We're shifting from human-speed traffic that's predictable and low concurrency to 'agent-speed' workloads that're recursive, bursty, and massive. ... We'll see the rise of 'agent-native' infrastructure. ... The winning platforms will be the only ones capable of surviving the deluge of tool execution that follows." — Malika Aubakirova, investor, AI Infrastructure team, a16z, December 9, 2025
2026 is the year AI creative tools go truly multimodal, spawning multiple successful products from meme makers to Hollywood directors.
"2026 is the year AI goes multimodal. Give a model whatever form of reference content you have and work with it to make something new or edit an existing scene. ... Content creation is one of the killer use cases of AI, and I expect we'll see multiple successful products across use cases and types of customers from meme makers to Hollywood directors." — Justine Moore, partner, Andreessen Horowitz, December 9, 2025
The AI-native data stack is still early and will keep being rebuilt around agents accessing the right context.
"While the ecosystem feels notably more mature, we're still in the early days of a truly AI-native data architecture. We're excited by ways AI can continue to transform multiple parts of the data stack, and we're beginning to see how data and AI infrastructure are becoming inextricably linked." — Jason Cui, partner, Andreessen Horowitz, December 9, 2025
In 2026, generated video becomes a medium you inhabit — a place where robots, games, and agents can act and learn.
"In 2026, video stops behaving like something we passively watch and starts feeling like a place we can actually step into. ... This shift turns video into a medium we can build on: a space where robots can practice, games can evolve, designers can prototype, and agents can learn by doing." — Yoko Li, partner, Andreessen Horowitz, December 9, 2025
In 2026, the enterprise system of record finally loses primacy to the AI agent layer that executes work end-to-end.
"In 2026, the real disruption in enterprise software is that the system of record will finally start to lose primacy. AI is collapsing the distance between intent and execution: models can now read, write, and reason directly across operational data, turning ITSM and CRM systems from passive databases into autonomous workflow engines. ... The interface becomes a dynamic agent layer, while the traditional system of record slips into the background as a commodity persistence tier—its strategic leverage ceded to whoever controls the intelligent execution environment employees actually use." — Sarah Wang, general partner, Growth team, Andreessen Horowitz, December 9, 2025
Vertical AI moves from single-agent reasoning to multiplayer in 2026, and the collaboration layer becomes the moat.
"2026 unlocks multiplayer mode. Vertical software benefits from domain-specific interfaces, data, and integrations. But vertical work is inherently multi-party. If agents are going to represent labor, they need to collaborate. ... And when value increases from multi-human and multi-agent collaboration, switching costs rise. Here we'll see the network effects that have eluded AI applications: the collaboration layer becomes the moat." — Alex Immerman, partner, Growth team, Andreessen Horowitz, December 9, 2025
In 2026, people will start interfacing with the web through their agents, and content will be optimized for machine legibility rather than human eyes.
"In 2026, people will start interfacing with the web through their agents. And what mattered for human consumption won't matter the same way for agent consumption. ... We're no longer designing for humans, but for agents. The new optimization isn't for visual hierarchy, but for machine legibility—and that will change the way we create and the tools we use to do it." — Stephenie Zhang, partner, Growth investing team, a16z, December 9, 2025
Screen time dies as an AI application KPI as the industry moves toward outcome-based pricing.
"As we move to a future based on outcome-based pricing that perfectly aligns incentives between vendors and users, we'll first move away from screen time reporting. ... The companies that tell the simplest sales pitch on ROI will continue to outpace their competitors." — Santiago Rodriguez, partner, Growth investing team, a16z, December 9, 2025
In 2026, the "healthy MAUs" — recurring, prevention-oriented health consumers — become the next high-potential healthtech customer segment.
"In 2026, a new healthcare customer segment will take center stage: the 'healthy MAUs.' ... We expect a wave of companies—both AI-native upstarts, and repackaged versions of incumbents—to start offering recurring services to serve this user base. ... 'healthy MAUs' represent the next high-potential customer segment for healthtech: continuously engaged, data-informed, and prevention oriented." — Julie Yoo, general partner, Bio + Health team, a16z, December 9, 2025
AI-powered world models will revolutionize storytelling in 2026, creating generative multiverses with their own digital economies.
"In 2026, AI-powered world models will revolutionize storytelling through interactive virtual worlds and digital economies. Technologies like Marble (World Labs) and Genie 3 (DeepMind) already generate full 3D environments from text prompts, allowing users to explore them as if in a game. As creators adopt these tools, entirely new storytelling formats will emerge, potentially culminating in a 'generative Minecraft,' where players co-create vast, evolving universes." — Jonathan Lai, general partner, a16z speedrun, December 9, 2025
2026 becomes "the year of me" — products shift from mass-produced to personalized for the individual, and the next century's biggest companies win by finding the individual inside the average.
"2026 will become 'the year of me': the moment when products stop being mass-produced and start being made for you. ... The biggest companies of the last century won by finding the average consumer. The biggest companies of the next century will win by finding the individual inside the average. 2026 is the year the world stops optimizing for everyone and starts optimizing for you." — Joshua Lu, investment partner, a16z speedrun, December 9, 2025
The first AI-native university will be born in 2026 and become the talent engine for a new economy.
"In 2026, I expect we'll see the birth of the first AI-native university, an institution built from the ground up around intelligent systems. ... And as every industry struggles to hire people who can design, govern, and collaborate with AI systems, this new university becomes the training ground, producing graduates fluent in orchestration who help augment a rapidly shifting workforce. This AI-native university will become the talent engine for a new economy." — Emily Bennett, investing partner, a16z speedrun, December 9, 2025
Blogs
No statements found
Blogs
No statements found