$415M Valuation for ERP Software While Everyone Else Chases AGI
How Smart Money Wins While Everyone Else Chases $1B Valuations for Calendar Apps
Everyone’s throwing billions at AI startups right now.
Most of them are going to fail spectacularly.
And if you’re not paying attention to why, you’re about to get left holding the most expensive bag since WeWork promised to revolutionize consciousness through overpriced desk space and kombucha on tap.
Let me walk you through what actually happened this week in AI land, what’s coming next week, and why 90% of the players are reading the game completely wrong.
What Just Went Down: The Reality Check
DualEntry just raised $90 million in a Series A.
Their valuation?
$415 million. They’re doing ERP and finance automation, which sounds about as sexy as watching Excel spreadsheets mate. But here’s the thing nobody’s saying out loud: they’re AI-native, which means they didn’t bolt ChatGPT onto some legacy monstrosity from 1997 and call it innovation.
They built the whole damn thing to think like an AI from day one.
Eve, a legal-AI startup, just crossed into unicorn territory with $103 million in funding. One billion dollar valuation. They’re targeting plaintiffs’ lawyers, which is either genius or insane, depending on whether you’ve ever tried to get a lawyer to change their workflow. (Spoiler: it’s like teaching a cat to fetch.)
Meanwhile, Mira Murati—you know, the former CTO of OpenAI who jumped ship—just launched Thinking Machines Lab’s first product. It’s an API for fine-tuning large language models. Translation: she’s selling picks and shovels while everyone else is panning for gold. Smart move. The people who got rich in the California Gold Rush weren’t the miners.
Anthropic is expanding globally like they’re trying to colonize every timezone before lunch. India, Europe, everywhere. The talent war is heating up, and if you’re a researcher with “alignment” or “multimodal” on your LinkedIn, your inbox is probably exploding with offers that include free lunch, unlimited PTO, and possibly a small island nation.
Oh, and there’s regulatory chaos brewing. The U.S. government might shut down, which could freeze SEC approvals and kill IPO activity for AI companies. Fermi, the AI energy startup, is watching this with the kind of sweaty anxiety usually reserved for founders who just realized their runway ends in 47 days.
So yeah. Lots of capital flowing. New products launching. Everyone scrambling for talent. And the regulatory environment about as stable as a three-legged chair on a trampoline.
The Forecast: What’s Actually About to Happen
Here’s what nobody wants to admit: we’re in the stupid phase of the AI cycle.
Not the end.
Not even close.
But the phase where valuations detach from reality, where every SaaS tool gets rebranded as an “agent,” and where founders wake up convinced that adding “AI-powered” to their pitch deck is worth an extra $50 million in valuation.
I’ve seen this movie before. Dot-com bubble. Blockchain. The metaverse that lasted about 14 minutes before everyone remembered they hate wearing headsets. The pattern is always the same: massive capital influx, frothy valuations, everyone convinced this time is different, and then a correction that leaves bodies scattered across the venture landscape.
But here’s the twist.
This time actually is different. Not in the way the hype merchants are selling it, but in a way that’s going to separate the players who understand operational architecture from the ones who think slapping GPT-4 on a broken process counts as innovation.
Funding & Valuations: The Pre-Pre-Seed+++ Insanity
More mid-stage AI startups are going to close rounds next week. Series A, Series B, all targeting verticals. Legal AI. Fintech AI. Healthcare AI. Dog grooming AI. (Okay, I made up that last one, but give it six months.)
The valuations are going to be completely bananas.
We’re talking about companies with $2 million in revenue getting valued at $200 million because they have “autonomous agent capabilities” and a roadmap that promises AGI by Q3.
Some founder somewhere is going to wake up, use ChatGPT to write their own pitch deck, auto-submit to every VC on Sand Hill Road, and actually close a round based on “promised tweet impressions” as a growth metric.
This sounds ridiculous.
It is ridiculous.
And it’s going to happen anyway, because when capital is abundant and FOMO is high, due diligence becomes “how fast can we wire the money before someone else does?”
The Talent Wars: LinkedIn on Steroids
The big players are going to keep poaching researchers like they’re collecting Pokémon cards. OpenAI, Anthropic, Google, Meta—they’re all fighting over the same 500 people who actually understand alignment, causal reasoning, or how to make multimodal models not hallucinate your grandmother into existence.
Top researchers are going to wake up to LinkedIn DMs that read: “Want to head my new AGI division in Antarctica? We have penguins and unlimited compute.” The compensation packages are approaching the absurd. We’re talking seven figures, equity that might actually be worth something, and benefits that include things like “personal AI alignment coach” and “existential dread counseling.”
If you’re running an AI company right now and you’re not actively recruiting, you’re already behind. The talent gap is going to define winners and losers more than the technology itself, because everyone has access to the same foundation models. The difference is whether you have people who know what to do with them.
Tools & Model Releases: DIY LLM Starter Kits
More tooling is coming for the “fine-tune your own model in 5 lines of code” crowd. Small teams, stealth startups, people building in their garage—they’re all releasing “DIY LLM starter kits” that promise to democratize AI development. Some of these will be legitimate. Most will be vaporware wrapped in impressive landing pages and Medium posts that start with “We’re excited to announce...”
The meme version? Someone’s going to announce “Model in a Box”—just plug in your dreams, spit out a startup. Comes with free existential dread API and a Notion template for your Series A deck. The scary part is this isn’t that far from reality. The barrier to entry for AI development is collapsing faster than most people realize, which means we’re about to see an explosion of terrible AI products built by people who have no business building AI products.
Quality control is about to become the defining competitive advantage.
Regulation: The Reckoning Nobody’s Prepared For
Policy interventions are coming. Data privacy audits. AI impact assessments. Mandates about transparency that nobody actually wants to implement because it would expose how much their “AI” is actually just decision trees from 2003 wrapped in a transformer API.
Congress is going to demand increasingly bizarre things. What’s your AI’s favorite color? If it refuses to answer, that’s non-compliance. GPU manufacturers will be required to publish their “climate guilt quotient” because someone’s going to notice that training these models burns enough electricity to power a small nation for a year.
The companies that are thinking about this now, building compliance into their architecture from the start, are going to have a massive advantage. Everyone else is going to scramble to bolt on governance frameworks after the fact, which is like trying to install airbags after you’ve already crashed the car.
Market Tactics: Verticalization or Death
The “horizontal AI platform” play is dead. Verticalization is the only path forward. AI for legal. AI for HR. AI for supply chain optimization. AI for predicting when your office plants need watering. (Okay, that one actually might be useful.)
Every SaaS tool you’re currently using is going to get rebranded as an “agent” whether it deserves the title or not. Your calendar app will become “The Calendar Agent™ that knows when you need tacos” because apparently we can’t just call things schedulers anymore. The pressure to move from experimentation to actual revenue is going to force a reckoning.
Companies that can demonstrate clear ROI—not “potential value” or “strategic positioning,” but actual dollars saved or earned—are going to win. Everyone else is going to discover that CFOs are surprisingly resistant to renewing six-figure contracts for tools that make pretty dashboards but don’t actually do anything.
The Week-by-Week Breakdown: A Satirical Field Guide
Monday arrives with scattered rounds of “raise or die” panic. Founders will flood Slack channels and email threads with pitch decks that all say the same thing in slightly different fonts.
Tuesday through Wednesday brings talent lightning strikes. Compute engineers get flash-poached mid-meeting. Some rogue startup leaks an “AGI roadmap” written in Comic Sans because apparently we’ve reached the timeline where even typography has given up. Investors will storm virtual meeting rooms to back any startup claiming “agentic autonomy,” regardless of whether they actually know what that means.
Thursday delivers regulatory thunderclaps. Someone in government tweets “Are AIs tax residents?” and triggers 47 think pieces. A startup pivots mid-week from fintech AI to predictive dog-walk scheduling AI because why not? The market’s clearly insane anyway.
Friday brings exit rumors and IPO speculation. Someone announces they’re “preparing for SPAC but calling it a SPAC-ing unicorn” which makes exactly zero sense but generates 10,000 retweets. Founders celebrate by posting cryptic product teaser GIFs that reveal nothing and generate maximum engagement.
The weekend hits and media outlets plaster “AI Is Eating the World” across every headline. Founders who haven’t slept all week write open letters about “aligning to humanity” while simultaneously trying to automate every human job they can think of. Random AI ethics controversies emerge, get amplified beyond reason, and disappear by Monday morning.
What This Actually Means (If You’re Smart)
The stupidity phase is opportunity.
When everyone’s throwing money at everything, the smart operators are building actual businesses with sustainable unit economics and clear value propositions. When everyone’s chasing hype, the ones who understand operational fundamentals are building competitive moats that can’t be copied by adding another API call.
The companies that will matter in five years aren’t the ones raising at insane valuations right now. They’re the ones solving real problems for customers who will actually pay them money. They’re the ones building teams that can execute, not just PowerPoint decks that can wow investors.
If you’re watching this circus and thinking “everyone’s lost their minds,” congratulations. You might be one of the few people who’s actually paying attention.
The gold rush is getting stupider.
And that’s exactly when smart money moves.
Sources & Links:
Reuters - DualEntry $90M Series A funding announcement and $415M valuation details for AI-native ERP/finance automation platform
Reuters - Eve Legal AI achieves unicorn status with $103M funding round reaching $1B valuation targeting plaintiffs’ lawyers segment
The Times of India - Thinking Machines Lab launch announcement for LLM fine-tuning API by former OpenAI CTO Mira Murati
The Times of India - Anthropic global expansion coverage including hiring initiatives across India and European markets
Chron - Fermi AI energy company IPO threat analysis related to potential U.S. government shutdown impact on SEC approval processes