VC
Value Add VC
⚡HomePulse⚡Helpful Apps📝Blog
← Value Add PulseFUNDING$190M

Upscale AI Hits a $2B Valuation With a $190M Series A Extension for AI Networking

Upscale AI raised a $190 million Series A extension led by Premji Invest, taking its total Series A to roughly $500 million and its valuation to $2 billion. The company builds AI networking infrastructure -- the high-speed fabric that connects GPUs inside data centers -- positioning it in one of the hottest and most capital-intensive corners of the AI buildout.

$190M Series A
Extension
~$500M
Total Series A
$2B
Valuation
Premji Invest
Lead
AI networking infrastructure
Product
TC
Trace Cohen
Early-stage VC & angel · Founder, New York Venture Partners
June 26, 2026
2 min read
KEY TAKEAWAYS FOR VCs & FOUNDERS
1

Networking is the under-appreciated bottleneck of AI data centers, behind only chips and power

2

A $2B valuation on an extended Series A shows how fast AI-infrastructure marks are climbing

3

It pits a startup against incumbents like Nvidia, Broadcom and Arista in the data-center fabric

4

Premji Invest leading signals deep-pocketed crossover demand for the AI physical layer

TC
The VC Read · Trace's TakeTrace Cohen

Everyone funds chips and worries about power, but the third leg of the AI data center -- the network between the GPUs -- is where a $500M Series A just got built. Idle accelerators waiting on a slow fabric are the most expensive thing in tech, which is why interconnect is suddenly venture-hot. The hard truth is that Upscale is charging straight at Nvidia, Broadcom and Arista on their home turf, and networking is a long-cycle, incumbent-heavy game. At a $2B mark, the benchmarks against NVLink and InfiniBand are the only thing that matters -- spec wins, not press releases.

💰 Funding Tracker →⚡ AI Chip Wars →

Upscale AI has raised a $190 million Series A extension led by Premji Invest, bringing its cumulative Series A funding to roughly $500 million and lifting its valuation to $2 billion, according to Crunchbase News. The company builds AI networking infrastructure -- the high-bandwidth fabric and switching that move data between the tens of thousands of GPUs packed into modern AI data centers.

Networking has emerged as one of the critical, under-discussed bottlenecks of the AI buildout. Training and serving large models requires GPUs to communicate at staggering speeds; if the network between chips can't keep up, expensive accelerators sit idle. As clusters scale into the tens and hundreds of thousands of GPUs, the interconnect fabric becomes as decisive to performance as the chips themselves -- the same logic that has Elon Musk's world buying optical-interconnect startups and hyperscalers designing custom silicon.

“Networking has emerged as one of the critical, under-discussed bottlenecks of the AI buildout.”

The valuation trajectory tells the story of the moment. A startup reaching a $2 billion valuation on an extended Series A -- with half a billion dollars raised before a Series B -- reflects how aggressively capital is flowing into the AI physical layer, and how investors are willing to pay up for companies positioned at the data-center chokepoints. Premji Invest, the investment office tied to Wipro's Azim Premji, leading the round adds a deep-pocketed crossover backer to the cap table.

Upscale AI is taking on formidable incumbents. The data-center networking market is dominated by Nvidia (through its acquisition of Mellanox and its NVLink and InfiniBand franchises), Broadcom, Arista Networks and Cisco -- giants with entrenched ecosystems and enormous R&D budgets. The startup's pitch has to be that purpose-built, AI-native networking can deliver better performance or economics than retrofitted general-purpose gear.

The bear case is the competitive moat: networking is a hardware-intensive business with long design cycles, deep incumbent relationships and brutal price pressure, and a $2 billion valuation leaves little room for stumbles. What to watch: customer wins and deployed capacity, independent benchmarks against Nvidia's interconnect stack, and whether Upscale AI can convert its funding into a defensible position before the giants close the gap.

ShareXLinkedInEmail

Originally reported by Crunchbase News. Analysis and editorial commentary by Value Add Pulse.

← Back to Pulse

Markets Now

live
SPCX▲+0.75%
$234.85
CBRS▼-0.92%
$257.40
SPY▲+0.16%
5,961.80
QQQ▲+0.19%
20,098.50
NVDA▼-0.99%
$150.60
MSFT▼-0.52%
$480.10
GOOGL▲+0.57%
$210.30
META▲+0.38%
$657.90

Read Next

FUNDING$1.5B

Baseten Raises $1.5B Series F at a $13B Valuation as AI Inference Demand Explodes

AI inference provider Baseten closed a $1.5 billion Series F at a $13 billion valuation, the largest funding round of the week, led by Altimeter Capital and Conviction Partners with Spark, Sands and Wellington participating. The raise underscores how serving AI models -- not just training them -- has become one of the most valuable and capital-intensive layers of the stack.

FUNDING$200M

Stealthy AI-R&D Lab Mirendil Raises a $200M Seed From a16z and Kleiner Perkins

Mirendil, a new and largely stealth frontier AI lab focused on building systems for AI research and development itself, raised a $200 million seed round co-led by Andreessen Horowitz and Kleiner Perkins, with Nvidia among its backers. A nine-figure seed for a pre-product lab is one of the year's clearest signals that capital is racing to fund the next generation of foundation-model challengers.

FUNDING$200M

Quantifind Raises $200M to Scale AI Risk Intelligence for Banks and National Security

Quantifind raised $200 million in a growth round led by Summit Partners to expand its AI-driven risk-intelligence platform, which helps financial institutions and government agencies detect financial crime, fraud and national-security threats. The round places another applied-AI company with clear, regulated demand among the week's largest financings.

@Trace_Cohen·t@nyvp.com