AI Startups

$1.1B Lab Raise, Orbital Data Center $2.3B: AI Startups to Watch for August 24, 2026

A 2-month-old lab raised $1.1B, an orbital data center startup hit $2.3B, and accounting AI reached unicorn status—the startups moving billions.

Daniyal

· 3 min read
Stylized rockets and server racks orbiting a glowing Earth, representing the idea of computing in space

Nearly $1.5 billion moved into AI startups in the span of a few days, and the pattern is unmistakable: money is chasing personal agents, orbital compute, and software that quietly displaces the incumbents in boring industries. These are the AI startups to watch this week.

River AI: a 2-month-old lab pulls in $1.1 billion

River AI, founded by xAI co-founder Igor Babuschkin, closed a $1.1 billion round led by General Catalyst and AMP PBC, with Nvidia, AMD Ventures, Y Combinator, and Temasek participating. The company came out of stealth in June with an unusual thesis from Babuschkin, who previously worked at DeepMind, OpenAI, and xAI.

His argument: the industry is building models to replace human workers, and it should instead build agents that are personally trainable assistants. "The stack has to be rebuilt end to end: training, models, the product layer, and new hardware that lets personal AI live close to you," he wrote at launch.

Why it's disruptive: if personal-assistant AI works, it attacks the very business model of the worker-replacement AI wave, a direct challenge to every lab currently selling automation for jobs.

Starcloud: data centers in low Earth orbit

Starcloud, the Redmond startup building AI data centers in orbit, raised a $250 million Series A extension at a $2.3 billion post-money valuation, announced August 21. The premise is that space-based compute with solar power escapes the grid and land constraints strangling terrestrial data centers on Earth, a painful bottleneck we covered in yesterday's General AI News.

Why it's disruptive: terrestrial data centers face a physical and political wall. Any player that can credibly move compute off-grid is attacking the single biggest constraint on the industry.

Velaura AI: the power problem as a business

Velaura AI raised $110 million in Series A at a valuation north of $1 billion, vaulting it to unicorn status. The play is making AI infrastructure that does more per watt, aimed at data center operators desperate to squeeze compute from limited energy.

Why it's disruptive: power, not capability, is now the binding constraint on AI. Companies that sell efficiency sit at the intersection of two giant spend curves.

Rillet: AI accounting, a unicorn in 48 hours

AI accounting startup Rillet raised $100 million and reached unicorn status within 48 hours of launching its raise, per TechCrunch. The speed itself is the signal: investors are confident enough in AI agents replacing back-office accounting that they closed a nine-figure round in two days.

Why it's disruptive: accounting is a massive, entrenched, heavily regulated market. If Rillet's agents genuinely handle books end to end, it threatens the labor model of the entire finance function, not just a single product category.

Openchip: Europe's counterpunch in AI silicon

European chip startup Openchip closed around €115 million to build AI silicon, a reminder that the chip war is not only US versus China. The move lands as Washington tightens export controls and applies fresh tariffs on Chinese AI chips, so the pressure to build regional supply chains is mounting on every continent. The EU has been spending heavily to avoid depending on Nvidia and on downstream US controls, and Openchip is one of the clearer bets in that push.

Why it's disruptive: every regional player that builds its own AI chips chips away at the near-monopoly on training and inference silicon, which is exactly why Nvidia's price-hike warnings matter for startups everywhere, from Europe to Pakistan and Southeast Asia.

What the pattern says

Across all five, one theme holds: money is flowing to startups that either reduce AI's dependence on physical constraints (power, grid, silicon, or labor) or build compute in wholly new places. That is a healthier allocation than the pure model-lab funding of the past couple of years, and founders should read it as a signal that the winning deals are now about infrastructure and incumbents' soft underbellies, not another GPT clone.

For the bigger picture behind these moves, see today's General AI News digest, and compare with who was raising yesterday in AI Startups to Watch from August 23. Browse the full AI Startups tag for the archive.

Related reading from earlier coverage, with the numbers behind each one:

More on AI Startups

Browse every AI Startups story →