
Qualcomm's acquisition of software startup Modular reflects a fundamental shift in AI infrastructure: as hardware becomes heterogeneous and scarce, the software layers connecting different chip types are gaining strategic value alongside the silicon itself. This trend is accelerating consolidation, with major tech companies and chip makers racing to acquire infrastructure startups to manage rising AI computing costs and complexity.
Summaries like this, in your inbox every morning.
Sign up free →What happened
Qualcomm acquired Modular, a Palo Alto software startup that helps developers run AI models across different computer chips. Separately, SambaNova is finalizing an $800 million(約1300億円) funding round led by General Atlantic, valuing the company at $10 billion(約1.6兆円). The deals highlight how software layers connecting chips are becoming as valuable as the hardware itself.
Why it matters
As AI hardware remains scarce and expensive, companies are shifting toward using multiple chip types — CPUs, GPUs, and AI-specific chips — to lower costs and improve efficiency. Software that works seamlessly across these different chips is critical for this shift. For enterprises running inference (the step where AI produces answers), this means the ability to manage heterogeneous hardware rather than being locked into one vendor's ecosystem.
What to watch
Dave Munichiello, a managing partner at GV (Google Ventures) who holds board seats at both Modular and SambaNova, notes that the demand for inference is soaring across medicine, law, coding, customer support and finance. He also signals that 15 to 20 companies are planning to go public in the coming six months, suggesting continued consolidation activity ahead.
No discussion yet for this article
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack