The Hack Link

Tag: generative AI

  • AI Agents in 2026: The Shift From Chatbots to Digital Coworkers

    Something strange happened in enterprise software over the past twelve months. The conversation about AI agents stopped being theoretical. Nobody at industry conferences is asking “what is an AI agent?” anymore. The questions have gotten sharper, more specific, and far more interesting: How do you price an agent that replaces a $95,000-a-year workflow? What happens when one agent spawns another agent and nobody can trace the decision chain? Who’s liable when an autonomous system approves a transaction it shouldn’t have?

    That shift — from curiosity to operational reality — is the story of AI agents in 2026. And if you’re building software, running a business, or just trying to understand where technology is actually heading (as opposed to where LinkedIn influencers say it’s heading), this is the category worth paying attention to.

    What Changed, and Why It Matters Now

    A year ago, most AI agents were glorified chatbots with a few API connections bolted on. They could answer questions, maybe draft an email, occasionally pull data from a spreadsheet. Useful, sure. But nobody was restructuring their operations around them.

    That era is over. The agents being deployed today don’t just respond to prompts — they observe, plan, execute multi-step workflows, use external tools, and loop back to correct their own mistakes. Think of the difference between asking someone a question and hiring someone to manage a process. That’s the gap that just closed.

    The numbers tell part of the story. Over 80% of technical teams have moved past the planning stage into active testing or production deployment. Nearly six in ten organisations now have agents running in live environments. And the market itself is on a trajectory that analysts project will grow from under $10 billion today to over $250 billion within the next decade.

    But raw market projections don’t capture what’s actually happening on the ground. What’s happening on the ground is that companies are discovering agents can do things that traditional automation never could — because agents don’t need a rigid script. They adapt.

    Where Agents Are Actually Working (Not Just Demoing)

    The gap between demo and deployment has always been the graveyard of enterprise technology. Plenty of tools look brilliant in a sales presentation and collapse the moment they encounter messy, real-world data. So where are AI agents actually delivering?

    Operations and workflow orchestration is the biggest deployment category. An agent that reviews incoming requests, classifies urgency, identifies the right approver, checks for missing information, sends follow-ups, and escalates when deadlines slip — that’s not a hypothetical. That’s running in production at dozens of companies right now. The agent handles the process; humans handle the judgement calls.

    Customer service has moved well beyond scripted chatbots. Sierra, which builds AI agents for enterprise customer support, is serving more than 40% of the Fortune 50. Their agents don’t just answer FAQs — they access account data, process changes, and resolve issues end-to-end. The economics are compelling: companies paying $8 to $15 per human-handled support interaction are seeing agent-handled interactions cost a fraction of that, with comparable satisfaction scores.

    Software development is arguably the most visible category. Coding agents like Claude Code and Cursor don’t just autocomplete lines of code — they read entire repositories, understand project architecture, implement features across multiple files, run tests, and iterate on failures. Claude Code alone is now responsible for roughly 4% of all public commits on GitHub. That’s not a tool. That’s a team member.

    Healthcare administration is a quieter but potentially larger story. Mayo Clinic has piloted AI agents to automate scheduling, documentation, and back-office administrative work. Oxford University Hospitals built agents that summarise patient charts, determine cancer staging, and draft treatment plans for tumour boards. The clinical staff focus on patients; the agents handle the paperwork that was eating their days alive.

    Drug discovery is being reshaped at the research layer. Genentech built agent ecosystems on cloud infrastructure to automate complex research workflows, freeing scientists to concentrate on the creative and interpretive work that actually leads to breakthroughs.

    The Pricing Question Nobody Has Solved

    Here’s where things get genuinely interesting — and genuinely messy. Traditional SaaS charges per seat, per month. But an AI agent doesn’t occupy a seat. It might replace half a workflow that three people share, or it might handle a volume of work that fluctuates wildly from week to week. Per-seat pricing doesn’t map onto what agents actually do.

    The industry is experimenting with three models, and none of them have clearly won.

    The first is subscription with usage caps — a flat monthly fee that includes a certain volume of agent actions, with overages billed on top. This is familiar to buyers and easy to budget for, but it creates awkward incentives. If the agent gets better and handles more volume, the customer pays more for the same outcome.

    The second is outcome-based pricing — charging per resolved ticket, per processed application, per completed workflow. This aligns the vendor’s incentive with the customer’s value, which sounds elegant in theory. In practice, it requires airtight definitions of what counts as a “resolution” and creates unpredictable revenue for the vendor.

    The third, and the one gaining the most traction in 2026, is a hybrid model — a base subscription that provides a revenue floor, plus per-outcome fees above a certain threshold. This gives vendors predictable income and gives buyers a sense that they’re paying for results rather than idle software.

    The companies that figure out pricing first will have a meaningful advantage, because the current confusion is slowing enterprise adoption. Procurement teams know how to approve a $50,000 annual software license. They don’t know how to approve an open-ended commitment that might cost $20,000 one month and $120,000 the next.

    The Security Problem That Keeps CISOs Awake

    If pricing is the unsolved business problem, security is the unsolved technical one — and it’s arguably more urgent.

    Traditional software security is built around a simple model: humans authenticate, software executes within defined permissions, and audit logs track who did what. AI agents break every part of that model. An agent isn’t a human, but it needs access to systems that were designed for human users. It makes decisions, but those decisions emerge from probabilistic models rather than deterministic code. It can be manipulated through prompt injection — instructions hidden in data that trick the agent into doing something its operators never intended.

    The data from 2026 is sobering. Only about 14% of organisations report that all their AI agents went into production with full security and IT approval. That means the vast majority of deployed agents are operating with incomplete oversight. A quarter of deployed agents can create and task other agents, which means the chain of accountability becomes nearly impossible to trace once you’re more than one layer deep.

    The U.S. federal government has taken notice. The National Institute of Standards and Technology issued a formal request for information on AI agent security earlier this year, specifically flagging the risks of agents that operate with little to no human oversight and interact with critical infrastructure.

    What does responsible agent security actually look like? The emerging consensus centres on three principles. First, treat every agent as an identity — the same way you’d onboard an employee, with specific permissions, access controls, and audit trails. Second, enforce minimum necessary scope: the agent should only access the systems and data it needs for its assigned workflow, nothing more. Third, build kill switches and human approval gates into any workflow where the stakes are high enough that a mistake would cause real damage.

    Companies that treat agent security as an afterthought are building on sand. The ones that build governance into the architecture from day one are the ones that enterprise buyers will trust enough to hand over their critical workflows.

    Multi-Agent Systems: When One Agent Isn’t Enough

    The next frontier — already in early production at some organisations — is multi-agent architectures, where specialised agents collaborate to complete workflows that would be too complex for any single agent.

    Picture a lead qualification pipeline. A research agent gathers company and contact data from public sources. A scoring agent evaluates the lead against ideal customer profile criteria. A writing agent drafts personalised outreach. An orchestration agent coordinates the sequence, handles exceptions, and routes the final output to the right salesperson. Each agent is focused and specialised. Together, they run a process that used to require a team of SDRs and hours of manual work.

    This is not science fiction. Tools like n8n, LangChain, AutoGen, and CrewAI are enabling these multi-agent workflows today, and the patterns are becoming repeatable. The sophistication is growing quickly — but so is the complexity of managing, debugging, and securing these systems when something goes sideways.

    The practical advice from teams already running multi-agent systems is consistent: start with a single-agent workflow that handles one task extremely well. Prove reliability. Then add a second agent when specialisation clearly improves the outcome. Don’t design a multi-agent orchestra before you’ve built a single instrument that plays in tune.

    What This Means If You’re Building (or Buying)

    For founders and builders, the opportunity is in vertical agents — systems designed for a specific industry with deep domain knowledge, proprietary data, and tight integration into existing workflows. Generic agent platforms will struggle against the foundation model providers (OpenAI, Anthropic, Google) who can ship similar capabilities for free. But an agent that understands the specific compliance requirements of community banking, or the documentation standards of behavioural health, or the inspection workflows of commercial real estate — that’s defensible. The big players won’t bother building it, and the generic tools can’t match the depth.

    For enterprise buyers, the most important thing you can do right now is pick one high-volume, structured workflow and deploy an agent against it. Not a flashy demo. Not a company-wide transformation initiative. One workflow. Measure the outcome. Learn what breaks. Then expand. The organisations getting the most value from agents in 2026 are the ones that started small, proved ROI on a single process, and scaled from evidence rather than ambition.

    For everyone else — and honestly, this includes most of us — the practical takeaway is that AI agents are about to become as routine as email. Not because the technology is mature (it isn’t), and not because every deployment succeeds (they don’t). But because the gap between what agents can do and what businesses need done is narrowing fast enough that ignoring the category is no longer a viable strategy.

    The era of simple prompts is ending. The era of AI that actually does things — plans, executes, adjusts, and delivers outcomes — is just getting started. The companies and individuals who figure out how to work with these systems, rather than just talk about them, will have an edge that compounds every quarter.

    And that edge is already showing up in the numbers.

  • The AI Funding Machine: Why $255 Billion in a Quarter Should Make You Nervous, Not Excited

    Here’s a number that stopped me cold.

    In Q1 of 2026, AI startups raised $255.5 billion globally. That’s not a typo. A quarter trillion dollars. In three months. And here’s the thing — that single quarter blew past the entire 2025 full-year total for AI venture funding.

    I’ve been tracking startup funding for years and I’ve never seen anything like it. Not during the dot-com era. Not during the mobile app gold rush. Not during the crypto boom. The scale of capital moving into AI right now is genuinely unprecedented, and if you’re not paying close attention to where it’s going — and more importantly, where it isn’t — you’re missing the actual story.

    So let’s talk about it. Not the press release version. The uncomfortable one.

    Three Companies, Two-Thirds of Everything

    PitchBook dropped a report a few days ago that should honestly be required reading for anyone in this industry. Of that $255.5 billion in Q1, three companies — OpenAI, Anthropic, and xAI — accounted for $172 billion. That’s 67.3% of all AI venture capital.

    Let that sit for a second.

    The remaining $83.5 billion got split across 1,543 other companies. Quick math: that’s about $54 million per company on average. But averages lie, and they lie hard here. The median AI VC deal size in 2025 was $5 million. The mean was $35.8 million. The gap between those two numbers tells you everything about how concentrated this market has become. A tiny handful of deals at the top are yanking the averages up while most AI startups are raising perfectly normal, unremarkable rounds.

    And it’s getting worse, not better. Foundational AI companies — the ones building the actual models — raised $178 billion in Q1 2026 across just 24 deals. Twenty-four. In all of 2025, that category did $88.9 billion across 66 deals. The number of companies getting funded at this level is shrinking while the check sizes are exploding.

    There’s a word for this dynamic and it’s not “healthy ecosystem.”

    Here it is, in case you want to stare at the numbers directly:

    SegmentCompaniesFunding (Q1 2026)Share
    The Big Three — OpenAI, Anthropic, xAI3$172 billion67.3%
    Everyone else1,543$83.5 billion32.7%
    Total1,546$255.5 billion100%

    Three companies. Two-thirds of the money. The “everyone else” bucket works out to roughly $54 million per company on average — but remember, the median deal was $5 million. Most of those 1,543 companies raised far less than the average suggests.

    Follow the Money (It’s Not VC Anymore)

    One of the weirder subplots that doesn’t get enough attention: venture capital isn’t really driving this anymore.

    Corporate venture capital now represents 43% of AI startup funding. Sovereign wealth funds are piling in. Microsoft, Amazon, Google, NVIDIA — they’re not writing checks because they want 10x returns on a Series B. They’re writing checks because they need to lock down compute agreements, secure cloud revenue, and make sure they don’t get left behind in a platform shift that could eat their existing businesses.

    When Microsoft disclosed that 45% of its $625 billion cloud backlog was tied to OpenAI, the stock dropped 12% in a single day. That’s $440 billion in market cap. Evaporated. Because investors suddenly realized that a huge chunk of Microsoft’s future revenue depends on a company that burns $17 billion a year and doesn’t expect to turn a profit until 2030.

    That’s not a venture capital relationship. That’s a dependency. And it’s not limited to Microsoft.

    The “investment” flowing into these companies is increasingly circular. Microsoft invests in OpenAI, which spends the money on Azure compute, which shows up as Microsoft cloud revenue, which justifies the investment. Amazon invests in Anthropic, Anthropic buys AWS compute, Amazon reports cloud growth. Around and around it goes. Everyone’s revenue is someone else’s cost. It works brilliantly until someone decides the music stopped.

    The Unit Economics Are Brutal

    Let’s look at OpenAI, since they’re the biggest and most transparently messy.

    They’re targeting about $30 billion in revenue for 2026. That sounds impressive — and it is, in absolute terms. But they’re burning roughly $17 billion to get there. Their own internal projections show losses tripling to $14 billion this year against roughly $13 billion in sales, with total spending around $22 billion. They lost $13.5 billion in just the first half of 2025.

    Every dollar of revenue costs them more than a dollar to generate. The cumulative burn from 2025 through 2029? An estimated $115 billion.

    Let me put OpenAI’s numbers in one place, because seeing them side by side makes the absurdity harder to ignore:

    Metric20252026 (Projected)
    RevenueNot disclosed~$30B
    Net Loss$13.5B (H1 only)~$14B
    Cash Burn—~$17B
    Total Spending—~$22B
    Valuation$300B (Mar round)~$850B (Mar round)
    Forward Revenue Multiple—~28x

    They’re valued at roughly $850 billion as of their March 2026 round. At $30 billion in projected revenue, that’s somewhere around 28x forward revenue. For a company with negative gross margins. In what universe is that sustainable?

    Anthropic is doing better on the surface — they surprised everyone by hitting a $30 billion annualized run-rate earlier this year. But they’re also burning cash at a pace that makes traditional SaaS investors queasy, and their valuation sits around $380 billion. The model providers are in an arms race where the cost of competing goes up faster than the revenue does. Every new model generation requires more compute, more data, more infrastructure, and the pricing pressure from open-source alternatives and Chinese competitors means you can’t just pass those costs to customers.

    A quick side-by-side puts the two Western leaders in perspective:

    MetricOpenAIAnthropic
    Valuation~$850B~$380B
    2026 Revenue (projected)~$30B$30B annualized run-rate
    2026 Cash Burn~$17BNot disclosed
    Path to Profitability2029 targetNot disclosed
    Key Strategic BackerMicrosoftAmazon
    Revenue Multiple~28x forward~12.7x (on run-rate)

    Anthropic’s multiple looks almost reasonable by comparison. Almost. Neither company has demonstrated that selling intelligence at scale produces software-like margins rather than infrastructure-like ones.

    And then there’s DeepSeek. In February 2026, they launched V4 — a trillion-parameter coding model with a million-token context window. Built at a fraction of the cost of Western frontier models. If Chinese labs can produce competitive AI without burning through $17 billion a year, the entire “scale is the moat” thesis falls apart. You don’t need a fancy financial model to see the problem there.

    What About Everyone Else?

    The non-foundational-model startups — the application layer, the vertical SaaS plays, the tooling companies — are living in a completely different reality from the handful of giants at the top.

    They’re raising at perfectly normal multiples. The AI startup valuation range in 2026 is wide — 10x to 50x revenue, with the median landing somewhere around 20x to 30x. But those multiples come with strings attached. Investors are asking harder questions than they were two years ago. What’s your moat, really? What happens when the model providers ship a feature that does 80% of what your product does? How defensible is your data advantage?

    The application-layer AI companies that are actually doing well tend to have one thing in common: they’re not just thin wrappers around someone else’s API. They own a workflow. They have proprietary data. They solve a problem that’s too specific, too regulated, or too operationally complex for a horizontal model provider to bother with. Everything else is walking dead — it just doesn’t know it yet.

    And early-stage funding? It’s getting squeezed. Only 14% of AI mega-deals in 2025 were early-stage. The money is flowing to the companies that already raised billions, not to the ones trying to raise their first million. That’s a problem for the ecosystem long-term, even if nobody wants to talk about it while the party’s still going.

    Is This a Bubble?

    I get asked this constantly. And the honest answer is: it depends on what you mean by “bubble.”

    If you mean “are AI company valuations completely unhinged from economic reality?” — yes, absolutely. The S&P 500 is trading at 23x forward earnings, the most stretched it’s been since the dot-com era. The Bank of England formally warned about overvaluation risks in the AI sector. A National Bureau of Economic Research study found that 90% of firms report no measurable productivity impact from AI yet — but executives project it will increase productivity by 1.4%. That gap between expectation and reality is exactly the kind of thing that precedes a correction.

    US mega-caps are expected to spend $1.1 trillion on AI between 2026 and 2029. Total AI spending is expected to surpass $1.6 trillion. Those are staggering numbers, and they assume a level of return that nobody has demonstrated yet.

    But here’s the counterargument, and it’s not nothing: the dot-com comparison misses something important. In 2000, most of the money went into companies with no revenue, no infrastructure, and no real technology moat. Pets.com sold dog food online. Today’s AI giants are building actual infrastructure — data centers, custom silicon, model architectures that take years to develop. You can correct the valuation of a software company overnight. You can’t wish a $15 billion data center out of existence. The physical assets create a floor that the dot-com era never had.

    What we’re probably heading for isn’t a pop. It’s a slow, grinding correction. The absurd valuations will come down. The “we put a chatbot on it” companies will disappear. The conversation will get more boring and more useful. That’s actually healthy.

    The risk isn’t that AI fails. The risk is that the returns don’t justify the $1.6 trillion price tag. And that’s a much more interesting problem to think about.

    What This Means If You’re Building

    I’ll keep this practical because that’s what actually matters.

    If you’re raising at the application layer, you need to be able to answer one question cold: what happens when OpenAI or Anthropic ships your feature? If your answer is “they won’t” or “our data moat is too deep,” you’d better have receipts. The model providers are not sitting still. They’re vacuuming up talent, acquiring companies, and expanding into enterprise workflows. Complacency is a death sentence.

    If you’re an early-stage founder, the funding environment is probably tighter than the headline numbers suggest. Yes, $255 billion moved in Q1. Almost none of it went to pre-seed or seed-stage AI companies. You’re competing for attention in a market where the mega-rounds suck up all the oxygen. That means you need to be capital-efficient by default, not by aspiration. The burn multiple — how much you spend for every dollar of new recurring revenue — has become the investor metric of choice in 2026. AI-native startups that use automation to keep burn multiples below 1.0x are getting funded. Everyone else is fighting uphill.

    And here’s something nobody in Silicon Valley wants to admit out loud: 29% of startups fail because they run out of money. Not because the product was bad. Not because the market wasn’t there. They just ran out of cash. In an environment where capital is concentrating at the top, that number goes up. If you’re building, know your runway down to the week. Twelve to eighteen months is comfortable. Anything under six is emergency mode.

    The brutal truth about AI funding in 2026: there’s more money in the system than ever before, and it’s never been harder to get your hands on it unless you’re already one of the chosen few. The concentration at the top masks how competitive things are for everyone else. And the unit economics at the top suggest that even the chosen few haven’t figured out how to make the math work.

    That’s not a reason to be bearish on AI. The technology is real, the adoption is real, and the long-term trajectory is up and to the right. But the funding environment has gotten weird in ways that reward skepticism. The people who win in markets like this are the ones who can tell the difference between a genuine growth story and a money-go-round — and who have the discipline to build something that works even when the easy money dries up.