Pour yourself a fresh cup. The startup funding world just had one hell of a week.

While you were grinding through your workweek, VCs deployed over $3 billion across deals ranging from AI chips that rewrite physics to robots learning in real factories, to autonomous cars that might actually work this time.

This isn't your typical "startup raised money" news dump. These are the deals that signal where the smart money is moving, and where the opportunities are hiding.

Grab your coffee. Let's break down the 5 biggest startup funding rounds from this week that actually matter.

Read time: 4 minutes

The deal: $1.2B Series D + $300M in milestones = $1.5B total
The backers: Mercedes-Benz, Stellantis, Nissan, Uber (yes, Uber)
Total raised: $2.8B
What they do: Autonomous driving AI, launching Uber robotaxis in multiple cities in 2026

Why this matters:

Most self-driving companies are dead or dying. Cruise shut down. Argo AI folded. Tesla's FSD is... let's call it "in progress."

But Wayve is different.

They raised $1.5 billion from actual car companies (Mercedes, Nissan, Stellantis) who are betting their futures on this tech.

And Uber — who tried and failed to build their own self-driving cars — is now betting on Wayve to power their robotaxi fleet.

Translation: This isn't a science project. It's a deployment bet.

The opportunity:

Self-driving is one of those categories where everyone thought it was "right around the corner" in 2018... and then reality hit.

Now, 8 years later, the timeline is actually realistic: 2026-2028 for limited robotaxi deployments.

If Wayve executes, this $1.5B round at private market pricing will look cheap in 3 years.

The deal: $500M Series B
The backers: Jane Street, Situational Awareness, Marvell (chip giant), Spark Capital, Stripe founders
What they do: AI chips that train models 10x faster than current GPUs

Why this matters:

Nvidia just reported $68.1 billion in Q4 revenue. They have 90%+ market share in AI chips.

Everyone wants to kill that monopoly.

Google has TPUs. Amazon has Trainium. Microsoft is working on custom silicon.

But MatX is the first startup to raise a $500 million Series B with the explicit goal of making Nvidia irrelevant for AI training.

The claim: 10x performance improvement over GPUs for training large language models.

If true, this changes everything.

The opportunity:

Nvidia's dominance creates a massive opportunity for challengers.

If MatX delivers even 3-5x performance (not 10x), they'll win market share from hyperscalers looking to reduce Nvidia dependence.

At $500M Series B valuation, you're betting on execution risk. But the TAM is $100B+ if they succeed.

Founderscrowd Premium gives you one fully analyzed investment opportunity every week, reviewed by our analysts and investment bankers. Get access to pre-IPO companies, private-market deals, and alternative investments that most investors never see.
$40/month (beta pricing, normally $120).

JOIN NOW.

See you on the other side.

The deal: ¥2 billion (~$290.5M)
Valuation: ¥10 billion (~$1.5B)
The backers: YF Capital, HongShan Capital, Chaos Investments, state-backed funds
What they do: "Embodied intelligence" — AI models that control physical robots for logistics/industrial use

Why this matters:

While U.S. investors obsess over AI chatbots, China is going all-in on physical AI.

Spirit AI builds:

The thesis: Software AI is commoditizing. Physical AI (robots) is the next frontier.

The opportunity:

U.S. investors are underweight on robotics. Most exposure is through Tesla (Optimus, still vaporware) or Boston Dynamics (not investable).

China is building the entire robotics stack — hardware, software, AI models, manufacturing — faster than the West.

If you believe robotics is a $1T+ TAM by 2035, you need exposure to Chinese robotics companies.

Spirit AI at $1.5B valuation is early enough to matter.

The deal: $100M Series B
Valuation: $1.15B (unicorn status)
The backers: Accel, GV (Google Ventures), Lloyd Blankfein (former Goldman CEO), Khosla Ventures
What they do: AI agents that automate multi-step accounting tasks

Why this matters:

Accounting is a $500 billion industry that still runs on spreadsheets, manual data entry, and humans double-checking other humans.

Basis is automating it.

Their AI agents handle:

The kicker: They're not replacing accountants. They're augmenting them, handling the grunt work so humans focus on judgment calls.

Why Lloyd Blankfein (Goldman's former CEO) invested:

He sees enterprise AI moving from "nice to have" to "table stakes."

If every company needs AI-powered accounting by 2028, Basis is positioned to be the Salesforce of finance.

The opportunity:

SaaS + AI is the new gold rush.

Accounting software (QuickBooks, Xero) is massive but stagnant. Basis is inserting AI into the workflow, creating 10x productivity gains.

At $1.15B valuation with $100M raise, they have 2-3 years of runway to hit $100M+ ARR and IPO.

This could be a 5-10x in 3 years if they execute.

The deal: $169M (largest AI chip financing of the year so far)
Total raised: $219M
The backers: Quiet Capital, Fidelity, Pierre Lamond (legendary chip investor)
What they do: Custom AI chips that "print parts of AI models onto silicon" for 10x faster inference

Why this matters:

Everyone's focused on AI training (building the models). But inference is where the money is.

The numbers:

Deloitte estimates inference will be 2/3 of AI compute spend by 2026.

Taalas is building chips specifically for inference — trading off generality for insane speed and cost savings.

The bet: You don't need a general-purpose GPU for inference. You need a chip optimized for your specific model.

The opportunity:

If Taalas can deliver 5-10x cost reduction for inference, every AI company will want their chips.

OpenAI spends billions on inference. Anthropic spends billions. Google, Meta, Microsoft — billions.

A 10x cost reduction = $10s of billions in savings.

At $219M total raised, Taalas is still early-stage. But this could be a $10-50B outcome if inference optimization becomes critical.

Look at the 5 deals:

1. Wayve: Physical AI (autonomous cars)
2. MatX: AI infrastructure (chips for training)
3. Spirit AI: Physical AI (humanoid robots)
4. Basis: AI applications (vertical SaaS)
5. Taalas: AI infrastructure (chips for inference)

Notice something?

3 out of 5 are infrastructure. 2 out of 5 are physical AI.

Zero are consumer AI chatbots.

The funding market is telling you where the money is:

Not:

If you're investing in private markets, this is your roadmap.

If you're a Founderscrowd member (free or Premium), here's how to think about these deals:

✅ Good bets:

❌ Bad bets:

The simple rule:

If Nvidia, OpenAI, or Anthropic could kill your business by adding one feature — don't invest.

If your business enables Nvidia, OpenAI, or Anthropic — that's the play.